diff --git a/.github/workflows/docs.yml b/.github/workflows/docs.yml index 12e45b8b..17e8b500 100644 --- a/.github/workflows/docs.yml +++ b/.github/workflows/docs.yml @@ -1,4 +1,3 @@ ---- name: Documentation on: push: diff --git a/.gitignore b/.gitignore index bd76ada1..780d95ae 100644 --- a/.gitignore +++ b/.gitignore @@ -11,12 +11,15 @@ dataframe/ static/generated runs +artifacts_five chroma Unit Testing Agent_state.json Devin_state.json json_logs Medical Image Diagnostic Agent_state.json D_state.json +artifacts_six +artifacts_seven swarms/__pycache__ artifacts transcript_generator.json diff --git a/README.md b/README.md index 42274f59..baa458ac 100644 --- a/README.md +++ b/README.md @@ -71,7 +71,8 @@ agent.run("Generate a 10,000 word blog on health and wellness.") `Agent` equipped with quasi-infinite long term memory. Great for long document understanding, analysis, and retrieval. ```python -from swarms import Agent, ChromaDB, OpenAIChat +from swarms import Agent, OpenAIChat +from playground.memory.chromadb_example import ChromaDB # Copy and paste the code and put it in your own local directory. # Making an instance of the ChromaDB class memory = ChromaDB( @@ -180,7 +181,7 @@ agent("Generate a transcript for a youtube video on what swarms are!") ``` ## Devin -Implementation of Devil in less than 90 lines of code with several tools: +Implementation of Devin in less than 90 lines of code with several tools: terminal, browser, and edit files! ```python @@ -406,56 +407,6 @@ print(f"Generated data: {generated_data}") ``` -### `Worker` -The `Worker` is a simple all-in-one agent equipped with an LLM, tools, and RAG for low level tasks. - -✅ Plug in and Play LLM. Utilize any LLM from anywhere and any framework - -✅ Reliable RAG: Utilizes FAISS for efficient RAG but it's modular so you can use any DB. - -✅ Multi-Step Parallel Function Calling: Use any tool - -```python -# Importing necessary modules -import os - -from dotenv import load_dotenv - -from swarms import OpenAIChat, Worker, tool - -# Loading environment variables from .env file -load_dotenv() - -# Retrieving the OpenAI API key from environment variables -api_key = os.getenv("OPENAI_API_KEY") - - -# Create a tool -@tool -def search_api(query: str): - pass - - -# Creating a Worker instance -worker = Worker( - name="My Worker", - role="Worker", - human_in_the_loop=False, - tools=[search_api], - temperature=0.5, - llm=OpenAIChat(openai_api_key=api_key), -) - -# Running the worker with a prompt -out = worker.run("Hello, how are you? Create an image of how your are doing!") - -# Printing the output -print(out) -``` - ------- - - @@ -1188,6 +1139,7 @@ Apache License # Citation Please cite Swarms in your paper or your project if you found it beneficial in any way! Appreciate you. + ```bibtex @misc{swarms, author = {Gomez, Kye}, diff --git a/docs/ReleaseNotes.md b/docs/ReleaseNotes.md deleted file mode 100644 index f8859fa0..00000000 --- a/docs/ReleaseNotes.md +++ /dev/null @@ -1,49 +0,0 @@ -# Release Notes - -## 3.7.5 - -2024-01-21 - -### Bug Fixes - -Fix imports of Agent, SequentialWorkflow, ModelParallelizer, Task, OpenAIChat, Gemini, GPT4VisionAPI - -### New Features - -New model: Odin for Object Detection and tracking -New mode: Ultralytics Object recognition YOLO - -New Tokenizers - -Schema generator for prompts. -New prompt for worker agent. - -New structure: plan, step - -New tool: execute tool - -New logger: get_logger - -Example for worker_agent - - - -## 3.6.8 - -2024-01-19 - -### Bug Fixes - -Removed ModelScope - -Removed CogAgent - -### New Features - -Added ultralytics vision models - -Added TimmModel to wrap timm models - -### Other - -Loosened version of timm diff --git a/docs/applications/compliance_swarm.md b/docs/applications/compliance_swarm.md new file mode 100644 index 00000000..e69de29b diff --git a/docs/ecosystem.md b/docs/community/ecosystem.md similarity index 100% rename from docs/ecosystem.md rename to docs/community/ecosystem.md diff --git a/docs/examples/bingchat.md b/docs/examples/bingchat.md deleted file mode 100644 index 3dabbf2e..00000000 --- a/docs/examples/bingchat.md +++ /dev/null @@ -1,71 +0,0 @@ -## BingChat User Guide - -Welcome to the BingChat user guide! This document provides a step-by-step tutorial on how to leverage the BingChat class, an interface to the EdgeGPT model by OpenAI. - -### Table of Contents -1. [Installation & Prerequisites](#installation) -2. [Setting Up BingChat](#setup) -3. [Interacting with BingChat](#interacting) -4. [Generating Images](#images) -5. [Managing Cookies](#cookies) - -### Installation & Prerequisites - -Before initializing the BingChat model, ensure you have the necessary dependencies installed: - -```shell -pip install EdgeGPT -``` - -Additionally, you must have a `cookies.json` file which is necessary for authenticating with EdgeGPT. - -### Setting Up BingChat - -To start, import the BingChat class: - -```python -from bing_chat import BingChat -``` - -Initialize BingChat with the path to your `cookies.json`: - -```python -chat = BingChat(cookies_path="./path/to/cookies.json") -``` - -### Interacting with BingChat - -You can obtain text responses from the EdgeGPT model by simply calling the instantiated object: - -```python -response = chat("Hello, my name is ChatGPT") -print(response) -``` - -You can also specify the conversation style: - -```python -from bing_chat import ConversationStyle - -response = chat("Tell me a joke", style=ConversationStyle.creative) -print(response) -``` - -### Generating Images - -BingChat allows you to generate images based on text prompts: - -```python -image_path = chat.create_img("Sunset over mountains", auth_cookie="YOUR_AUTH_COOKIE") -print(f"Image saved at: {image_path}") -``` - -Ensure you provide the required `auth_cookie` for image generation. - -### Managing Cookies - -You can set a directory path for managing cookies using the `set_cookie_dir_path` method: - -BingChat.set_cookie_dir_path("./path/to/cookies_directory") - - diff --git a/docs/examples/flow.md b/docs/examples/flow.md deleted file mode 100644 index 2e635111..00000000 --- a/docs/examples/flow.md +++ /dev/null @@ -1,382 +0,0 @@ -# Reliable Enterprise-Grade Autonomous Agents in Less Than 5 lines of Code -======================================================================== - -Welcome to the walkthrough guide for beginners on using the "Agent" feature within the Swarms framework. This guide is designed to help you understand and utilize the capabilities of the Agent class for seamless and reliable interactions with autonomous agents. - -## Official Swarms Links -===================== - -[Swarms website:](https://www.swarms.world/) - -[Swarms Github:](https://github.com/kyegomez/swarms) - -[Swarms docs:](https://swarms.apac.ai/en/latest/) - -[Swarm Community!](https://discord.gg/39j5kwTuW4)! - -[Book a call with The Swarm Corporation here if you're interested in high performance custom swarms!](https://calendly.com/swarm-corp/30min) - -Now let's begin... - -## [Table of Contents](https://github.com/kyegomez/swarms) -=========================================================================================================== - -1. Introduction to Swarms Agent Module - -- 1.1 What is Swarms? -- 1.2 Understanding the Agent Module - -2. Setting Up Your Development Environment - -- 2.1 Installing Required Dependencies -- 2.2 API Key Setup -- 2.3 Creating Your First Agent - -3. Creating Your First Agent - -- 3.1 Importing Necessary Libraries -- 3.2 Defining Constants -- 3.3 Initializing the Agent Object -- 3.4 Initializing the Language Model -- 3.5 Running Your Agent -- 3.6 Understanding Agent Options - -4. Advanced Agent Concepts - -- 4.1 Custom Stopping Conditions -- 4.2 Dynamic Temperature Handling -- 4.3 Providing Feedback on Responses -- 4.4 Retry Mechanism -- 4.5 Response Filtering -- 4.6 Interactive Mode - -5. Saving and Loading Agents - -- 5.1 Saving Agent State -- 5.2 Loading a Saved Agent - -6. Troubleshooting and Tips - -- 6.1 Analyzing Feedback -- 6.2 Troubleshooting Common Issues - -7. Conclusion - -## [1. Introduction to Swarms Agent Module](https://github.com/kyegomez/swarms) -=================================================================================================================================================== - -### [1.1 What is Swarms?](https://github.com/kyegomez/swarms) -------------------------------------------------------------------------------------------------------------- - -Swarms is a powerful framework designed to provide tools and capabilities for working with language models and automating various tasks. It allows developers to interact with language models seamlessly. - -## 1.2 Understanding the Agent Feature -================================== - -### [What is the Agent Feature?](https://github.com/kyegomez/swarms) --------------------------------------------------------------------------------------------------------------------------- - -The Agent feature is a powerful component of the Swarms framework that allows developers to create a sequential, conversational interaction with AI language models. It enables developers to build multi-step conversations, generate long-form content, and perform complex tasks using AI. The Agent class provides autonomy to language models, enabling them to generate responses in a structured manner. - -### [Key Concepts](https://github.com/kyegomez/swarms) -------------------------------------------------------------------------------------------------- - -Before diving into the practical aspects, let's clarify some key concepts related to the Agent feature: - -- Agent: A Agent is an instance of the Agent class that represents an ongoing interaction with an AI language model. It consists of a series of steps and responses. -- Stopping Condition: A stopping condition is a criterion that, when met, allows the Agent to stop generating responses. This can be user-defined and can depend on the content of the responses. -- Loop Interval: The loop interval specifies the time delay between consecutive interactions with the AI model. -- Retry Mechanism: In case of errors or failures during AI model interactions, the Agent can be configured to make multiple retry attempts with a specified interval. -- Interactive Mode: Interactive mode allows developers to have a back-and-forth conversation with the AI model, making it suitable for real-time interactions. - -## [2. Setting Up Your Development Environment](https://github.com/kyegomez/swarms) -============================================================================================================================================================= - -### [2.1 Installing Required Dependencies](https://github.com/kyegomez/swarms) ------------------------------------------------------------------------------------------------------------------------------------------------- - -Before you can start using the Swarms Agent module, you need to set up your development environment. First, you'll need to install the necessary dependencies, including Swarms itself. - -# Install Swarms and required libraries -`pip3 install --upgrade swarms` - -## [2. Creating Your First Agent](https://github.com/kyegomez/swarms) ------------------------------------------------------------------------------------------------------------------------------ - -Now, let's create your first Agent. A Agent represents a chain-like structure that allows you to engage in multi-step conversations with language models. The Agent structure is what gives an LLM autonomy. It's the Mitochondria of an autonomous agent. - -# Import necessary modules -```python -from swarms.models import OpenAIChat # Zephr, Mistral -from swarms.structs import Agent - -api_key = "" # Initialize the language model (LLM) -llm = OpenAIChat( - openai_api_key=api_key, temperature=0.5, max_tokens=3000 -) # Initialize the Agent object - -agent = Agent(llm=llm, max_loops=5) # Run the agent -out = agent.run("Create an financial analysis on the following metrics") -print(out) -``` - -### [3. Initializing the Agent Object](https://github.com/kyegomez/swarms) ----------------------------------------------------------------------------------------------------------------------------------------- - -Create a Agent object that will be the backbone of your conversational agent. - -```python -# Initialize the Agent object -agent = Agent( - llm=llm, - max_loops=5, - stopping_condition=None, # You can define custom stopping conditions - loop_interval=1, - retry_attempts=3, - retry_interval=1, - interactive=False, # Set to True for interactive mode - dashboard=False, # Set to True for a dashboard view - dynamic_temperature=False, # Enable dynamic temperature handling -) -``` - -### [3.2 Initializing the Language Model](https://github.com/kyegomez/swarms) ----------------------------------------------------------------------------------------------------------------------------------------------- - -Initialize the language model (LLM) that your Agent will interact with. In this example, we're using OpenAI's GPT-3 as the LLM. - -- You can also use `Mistral` or `Zephr` or any of other models! - -```python -# Initialize the language model (LLM) -llm = OpenAIChat( - openai_api_key=api_key, - temperature=0.5, - max_tokens=3000, -) -``` - -### [3.3 Running Your Agent](https://github.com/kyegomez/swarms) ------------------------------------------------------------------------------------------------------------------- - -Now, you're ready to run your Agent and start interacting with the language model. - -If you are using a multi modality model, you can pass in the image path as another parameter - -``` -# Run your Agent -out = agent.run( - "Generate a 10,000 word blog on health and wellness.", - # "img.jpg" , Image path for multi-modal models - ) - -print(out) -``` - -This code will initiate a conversation with the language model, and you'll receive responses accordingly. - -## [4. Advanced Agent Concepts](https://github.com/kyegomez/swarms) -=========================================================================================================================== - -In this section, we'll explore advanced concepts that can enhance your experience with the Swarms Agent module. - -### [4.1 Custom Stopping Conditions](https://github.com/kyegomez/swarms) - -You can define custom stopping conditions for your Agent. For example, you might want the Agent to stop when a specific word is mentioned in the response. - -# Custom stopping condition example -```python -def stop_when_repeats(response: str) -> bool: - return "Stop" in response.lower() -``` - -# Set the stopping condition in your Agent -```agent.stopping_condition = stop_when_repeats``` - -### [4.2 Dynamic Temperature Handling](https://github.com/kyegomez/swarms) ----------------------------------------------------------------------------------------------------------------------------------------- - -Dynamic temperature handling allows you to adjust the temperature attribute of the language model during the conversation. - -# Enable dynamic temperature handling in your Agent -`agent.dynamic_temperature = True` - -This feature randomly changes the temperature attribute for each loop, providing a variety of responses. - -### [4.3 Providing Feedback on Responses](https://github.com/kyegomez/swarms) ----------------------------------------------------------------------------------------------------------------------------------------------- - -You can provide feedback on responses generated by the language model using the `provide_feedback` method. - -- Provide feedback on a response -`agent.provide_feedback("The response was helpful.")` - -This feedback can be valuable for improving the quality of responses. - -### [4.4 Retry Mechanism](https://github.com/kyegomez/swarms) --------------------------------------------------------------------------------------------------------------- - -In case of errors or issues during conversation, you can implement a retry mechanism to attempt generating a response again. - -# Set the number of retry attempts and interval -```python -agent.retry_attempts = 3 -agent.retry_interval = 1 # in seconds -``` -### [4.5 Response Filtering](https://github.com/kyegomez/swarms) --------------------------------------------------------------------------------------------------------------------- - -You can add response filters to filter out certain words or phrases from the responses. - -# Add a response filter -```python -agent.add_response_filter("inappropriate_word") -``` -This helps in controlling the content generated by the language model. - -### [4.6 Interactive Mode](https://github.com/kyegomez/swarms) ----------------------------------------------------------------------------------------------------------------- - -Interactive mode allows you to have a back-and-forth conversation with the language model. When enabled, the Agent will prompt for user input after each response. - -# Enable interactive mode -`agent.interactive = True` - -This is useful for real-time conversations with the model. - -## [5. Saving and Loading Agents](https://github.com/kyegomez/swarms) -=============================================================================================================================== - -### [5.1 Saving Agent State](https://github.com/kyegomez/swarms) ------------------------------------------------------------------------------------------------------------------- - -You can save the state of your Agent, including the conversation history, for future use. - -# Save the Agent state to a file -`agent.save("path/to/flow_state.json")`` - -### [5.2 Loading a Saved Agent](https://github.com/kyegomez/swarms) ------------------------------------------------------------------------------------------------------------------------- - -To continue a conversation or reuse a Agent, you can load a previously saved state. - -# Load a saved Agent state -`agent.load("path/to/flow_state.json")`` - -## [6. Troubleshooting and Tips](https://github.com/kyegomez/swarms) -=============================================================================================================================== - -### [6.1 Analyzing Feedback](https://github.com/kyegomez/swarms) --------------------------------------------------------------------------------------------------------------------- - -You can analyze the feedback provided during the conversation to identify issues and improve the quality of interactions. - -# Analyze feedback -`agent.analyze_feedback()` - -### [6.2 Troubleshooting Common Issues](https://github.com/kyegomez/swarms) ------------------------------------------------------------------------------------------------------------------------------------------- - -If you encounter issues during conversation, refer to the troubleshooting section for guidance on resolving common problems. - -# [7. Conclusion: Empowering Developers with Swarms Framework and Agent Structure for Automation](https://github.com/kyegomez/swarms) -================================================================================================================================================================================================================================================================ - -In a world where digital tasks continue to multiply and diversify, the need for automation has never been more critical. Developers find themselves at the forefront of this automation revolution, tasked with creating reliable solutions that can seamlessly handle an array of digital tasks. Enter the Swarms framework and the Agent structure, a dynamic duo that empowers developers to build autonomous agents capable of efficiently and effectively automating a wide range of digital tasks. - -[The Automation Imperative](https://github.com/kyegomez/swarms) ---------------------------------------------------------------------------------------------------------------------------- - -Automation is the driving force behind increased efficiency, productivity, and scalability across various industries. From mundane data entry and content generation to complex data analysis and customer support, the possibilities for automation are vast. Developers play a pivotal role in realizing these possibilities, and they require robust tools and frameworks to do so effectively. - -[Swarms Framework: A Developer's Swiss Army Knife](https://github.com/kyegomez/swarms) ------------------------------------------------------------------------------------------------------------------------------------------------------------------------ - -The Swarms framework emerges as a comprehensive toolkit designed to empower developers in their automation endeavors. It equips developers with the tools and capabilities needed to create autonomous agents capable of interacting with language models, orchestrating multi-step workflows, and handling error scenarios gracefully. Let's explore why the Swarms framework is a game-changer for developers: - -[1. Language Model Integration](https://github.com/kyegomez/swarms) ------------------------------------------------------------------------------------------------------------------------------------ - -One of the standout features of Swarms is its seamless integration with state-of-the-art language models, such as GPT-3. These language models have the ability to understand and generate human-like text, making them invaluable for tasks like content creation, translation, code generation, and more. - -By leveraging Swarms, developers can effortlessly incorporate these language models into their applications and workflows. For instance, they can build chatbots that provide intelligent responses to customer inquiries or generate lengthy documents with minimal manual intervention. This not only saves time but also enhances overall productivity. - -[2. Multi-Step Conversational Agents](https://github.com/kyegomez/swarms) ---------------------------------------------------------------------------------------------------------------------------------------------- - -Swarms excels in orchestrating multi-step conversational flows. Developers can define intricate sequences of interactions, where the system generates responses, and users provide input at various stages. This functionality is a game-changer for building chatbots, virtual assistants, or any application requiring dynamic and context-aware conversations. - -These conversational flows can be tailored to handle a wide range of scenarios, from customer support interactions to data analysis. By providing a structured framework for conversations, Swarms empowers developers to create intelligent and interactive systems that mimic human-like interactions. - -[3. Customization and Extensibility](https://github.com/kyegomez/swarms) ---------------------------------------------------------------------------------------------------------------------------------------------- - -Every development project comes with its unique requirements and challenges. Swarms acknowledges this by offering a high degree of customization and extensibility. Developers can define custom stopping conditions, implement dynamic temperature handling for language models, and even add response filters to control the generated content. - -Moreover, Swarms supports an interactive mode, allowing developers to engage in real-time conversations with the language model. This feature is invaluable for rapid prototyping, testing, and fine-tuning the behavior of autonomous agents. - -[4. Feedback-Driven Improvement](https://github.com/kyegomez/swarms) -------------------------------------------------------------------------------------------------------------------------------------- - -Swarms encourages the collection of feedback on generated responses. Developers and users alike can provide feedback to improve the quality and accuracy of interactions over time. This iterative feedback loop ensures that applications built with Swarms continually improve, becoming more reliable and capable of autonomously handling complex tasks. - -[5. Handling Errors and Retries](https://github.com/kyegomez/swarms) -------------------------------------------------------------------------------------------------------------------------------------- - -Error handling is a critical aspect of any automation framework. Swarms simplifies this process by offering a retry mechanism. In case of errors or issues during conversations, developers can configure the framework to attempt generating responses again, ensuring robust and resilient automation. - -[6. Saving and Loading Agents](https://github.com/kyegomez/swarms) -------------------------------------------------------------------------------------------------------------------------------- - -Developers can save the state of their conversational flows, allowing for seamless continuity and reusability. This feature is particularly beneficial when working on long-term projects or scenarios where conversations need to be resumed from a specific point. - -[Unleashing the Potential of Automation with Swarms and Agent](https://github.com/kyegomez/swarms) -=============================================================================================================================================================================================== - -The combined power of the Swarms framework and the Agent structure creates a synergy that empowers developers to automate a multitude of digital tasks. These tools provide versatility, customization, and extensibility, making them ideal for a wide range of applications. Let's explore some of the remarkable ways in which developers can leverage Swarms and Agent for automation: - -[1. Customer Support and Service Automation](https://github.com/kyegomez/swarms) -------------------------------------------------------------------------------------------------------------------------------------------------------------- - -Swarms and Agent enable the creation of AI-powered customer support chatbots that excel at handling common inquiries, troubleshooting issues, and escalating complex problems to human agents when necessary. This level of automation not only reduces response times but also enhances the overall customer experience. - -[2. Content Generation and Curation](https://github.com/kyegomez/swarms) ---------------------------------------------------------------------------------------------------------------------------------------------- - -Developers can harness the power of Swarms and Agent to automate content generation tasks, such as writing articles, reports, or product descriptions. By providing an initial prompt, the system can generate high-quality content that adheres to specific guidelines and styles. - -Furthermore, these tools can automate content curation by summarizing lengthy articles, extracting key insights from research papers, and even translating content into multiple languages. - -[3. Data Analysis and Reporting](https://github.com/kyegomez/swarms) -------------------------------------------------------------------------------------------------------------------------------------- - -Automation in data analysis and reporting is fundamental for data-driven decision-making. Swarms and Agent simplify these processes by enabling developers to create flows that interact with databases, query data, and generate reports based on user-defined criteria. This empowers businesses to derive insights quickly and make informed decisions. - -[4. Programming and Code Generation](https://github.com/kyegomez/swarms) ---------------------------------------------------------------------------------------------------------------------------------------------- - -Swarms and Agent streamline code generation and programming tasks. Developers can create flows to assist in writing code snippets, auto-completing code, or providing solutions to common programming challenges. This accelerates software development and reduces the likelihood of coding errors. - -[5. Language Translation and Localization](https://github.com/kyegomez/swarms) ---------------------------------------------------------------------------------------------------------------------------------------------------------- - -With the ability to interface with language models, Swarms and Agent can automate language translation tasks. They can seamlessly translate content from one language to another, making it easier for businesses to reach global audiences and localize their offerings effectively. - -[6. Virtual Assistants and AI Applications](https://github.com/kyegomez/swarms) ------------------------------------------------------------------------------------------------------------------------------------------------------------ - -Developers can build virtual assistants and AI applications that offer personalized experiences. These applications can automate tasks such as setting reminders, answering questions, providing recommendations, and much more. Swarms and Agent provide the foundation for creating intelligent, interactive virtual assistants. - -[Future Opportunities and Challenges](https://github.com/kyegomez/swarms) ------------------------------------------------------------------------------------------------------------------------------------------------ - -As Swarms and Agent continue to evolve, developers can look forward to even more advanced features and capabilities. However, with great power comes great responsibility. Developers must remain vigilant about the ethical use of automation and language models. Ensuring that automated systems provide accurate and unbiased information is an ongoing challenge that the developer community must address. - -# [In Conclusion](https://github.com/kyegomez/swarms) -=================================================================================================== - -The Swarms framework and the Agent structure empower developers to automate an extensive array of digital tasks by offering versatility, customization, and extensibility. From natural language understanding and generation to orchestrating multi-step conversational flows, these tools simplify complex automation scenarios. - -By embracing Swarms and Agent, developers can not only save time and resources but also unlock new opportunities for innovation. The ability to harness the power of language models and create intelligent, interactive applications opens doors to a future where automation plays a pivotal role in our digital lives. - -As the developer community continues to explore the capabilities of Swarms and Agent, it is essential to approach automation with responsibility, ethics, and a commitment to delivering valuable, user-centric experiences. With Swarms and Agent, the future of automation is in the hands of developers, ready to create a more efficient, intelligent, and automated world. \ No newline at end of file diff --git a/docs/examples/index.md b/docs/examples/index.md index b72c1e8f..ca08f04c 100644 --- a/docs/examples/index.md +++ b/docs/examples/index.md @@ -1,3 +1,3 @@ This section of the documentation is dedicated to examples highlighting Swarms functionality. -We try to keep all examples up to date, but if you think there is a bug please [submit a pull request](https://github.com/kyegomez/swarms-docs/tree/main/docs/examples). We are also more than happy to include new examples :) \ No newline at end of file +We try to keep all examples up to date, but if you think there is a bug please [submit a pull request](https://github.com/kyegomez/swarms-docs/tree/main/docs/examples). We are also more than happy to include new examples) \ No newline at end of file diff --git a/docs/examples/omni_agent.md b/docs/examples/omni_agent.md deleted file mode 100644 index e4d65498..00000000 --- a/docs/examples/omni_agent.md +++ /dev/null @@ -1,106 +0,0 @@ -# **OmniModalAgent from Swarms: A Comprehensive Starting Guide** - ---- - -**Table of Contents** - -1. Introduction: The OmniModal Magic -2. The Mechanics: Unraveling the Underpinnings -3. The Installation Adventure: Setting the Stage -4. Practical Examples: Let’s Get Our Hands Dirty! -5. Error Handling: Because Bumps on the Road are Inevitable -6. Dive Deeper: Advanced Features and Usage -7. Wrapping Up: The Road Ahead - ---- - -**1. Introduction: The OmniModal Magic** - -Imagine a world where you could communicate seamlessly across any modality, be it text, image, speech, or even video. Now, stop imagining because OmniModalAgent is here to turn that dream into reality. By leveraging advanced architecture and state-of-the-art tools, it can understand and generate any modality you can think of! - ---- - -**2. The Mechanics: Unraveling the Underpinnings** - -Dive into the world of OmniModalAgent and let’s decipher how it works: - -- **LLM (Language Model)**: It’s the brain behind understanding and generating language-based interactions. -- **Chat Planner**: Think of it as the strategist. It lays out the plan for the user's input. -- **Task Executor**: The doer. Once the plan is ready, this component takes charge to execute tasks. -- **Tools**: A treasure chest full of tools, from image captioning to translation. - ---- - -**3. The Installation Adventure: Setting the Stage** - -Getting OmniModalAgent up and running is as easy as pie. Ready to bake? - -```bash -pip install swarms -``` - -And voilà, your oven (system) is now equipped to bake any modality cake you desire! - ---- - -**4. Practical Examples: Let’s Get Our Hands Dirty!** - -Let’s embark on an exciting journey with OmniModalAgent: - -**i. Basic Interaction**: - -```python -from swarms.agents import OmniModalAgent -from swarms.models import OpenAIChat - -llm = OpenAIChat(openai_api_key="sk-") -agent = OmniModalAgent(llm) -response = agent.run("Create an video of a swarm of fish concept art, game art") -print(response) -``` - -**ii. Dive into a Conversation**: - -```python -agent = OmniModalAgent(llm) -print(agent.chat("What's the weather like?")) -``` - ---- - -**5. Error Handling: Because Bumps on the Road are Inevitable** - -Errors are like rain, unpredictable but inevitable. Luckily, OmniModalAgent comes with an umbrella. If there's a hiccup during message processing, it’s gracious enough to let you know. - -For instance, if there's a bump, you’ll receive: - -```python -Error processing message: [Details of the error] -``` - ---- - -**6. Dive Deeper: Advanced Features and Usage** - -The power of OmniModalAgent isn’t just limited to basic interactions. Here’s a sneak peek into its advanced capabilities: - -**Streaming Responses**: - -Imagine receiving responses as a gentle stream rather than a sudden splash. With the `_stream_response` method, you can achieve just that. - -```python -for token in agent._stream_response(response): - print(token) -``` - -**The Treasure Chest: Tools**: - -OmniModalAgent boasts a plethora of tools, from image captioning to speech-to-text. When you initialize the agent, it equips itself with these tools, ready to tackle any challenge you throw its way. - ---- - -**7. Wrapping Up: The Road Ahead** - -You've just scratched the surface of what OmniModalAgent can do. As you explore deeper, you'll discover more of its magic. The world of multi-modality is vast, and with OmniModalAgent as your companion, there's no limit to where you can go. - -**Happy Exploring and Coding!** 🚀🎉 diff --git a/docs/examples/revgpt.md b/docs/examples/revgpt.md deleted file mode 100644 index 69107b40..00000000 --- a/docs/examples/revgpt.md +++ /dev/null @@ -1,117 +0,0 @@ -## ChatGPT User Guide with Abstraction - -Welcome to the ChatGPT user guide! This document will walk you through the Reverse Engineered ChatGPT API, its usage, and how to leverage the abstraction in `revgpt.py` for seamless integration. - -### Table of Contents -1. [Installation](#installation) -2. [Initial Setup and Configuration](#initial-setup) -3. [Using the Abstract Class from `revgpt.py`](#using-abstract-class) -4. [V1 Standard ChatGPT](#v1-standard-chatgpt) -5. [V3 Official Chat API](#v3-official-chat-api) -6. [Credits & Disclaimers](#credits-disclaimers) - -### Installation - -To kickstart your journey with ChatGPT, first, install the ChatGPT package: - -```shell -python -m pip install --upgrade revChatGPT -``` - -**Supported Python Versions:** -- Minimum: Python3.9 -- Recommended: Python3.11+ - -### Initial Setup and Configuration - -1. **Account Setup:** Register on [OpenAI's ChatGPT](https://chat.openai.com/). -2. **Authentication:** Obtain your access token from OpenAI's platform. -3. **Environment Variables:** Configure your environment with the necessary variables. An example of these variables can be found at the bottom of the guide. - -### Using the Abstract Class from `revgpt.py` - -The abstraction provided in `revgpt.py` is designed to simplify your interactions with ChatGPT. - -1. **Import the Necessary Modules:** - -```python -from dotenv import load_dotenv -from revgpt import AbstractChatGPT -``` - -2. **Load Environment Variables:** - -```python -load_dotenv() -``` - -3. **Initialize the ChatGPT Abstract Class:** - -```python -chat = AbstractChatGPT(api_key=os.getenv("ACCESS_TOKEN"), **config) -``` - -4. **Start Interacting with ChatGPT:** - -```python -response = chat.ask("Hello, ChatGPT!") -print(response) -``` - -With the abstract class, you can seamlessly switch between different versions or models of ChatGPT without changing much of your code. - -### V1 Standard ChatGPT - -If you wish to use V1 specifically: - -1. Import the model: - -```python -from swarms.models.revgptV1 import RevChatGPTModelv1 -``` - -2. Initialize: - -```python -model = RevChatGPTModelv1(access_token=os.getenv("ACCESS_TOKEN"), **config) -``` - -3. Interact: - -```python -response = model.run("What's the weather like?") -print(response) -``` - -### V3 Official Chat API - -For users looking to integrate the official V3 API: - -1. Import the model: - -```python -from swarms.models.revgptV4 import RevChatGPTModelv4 -``` - -2. Initialize: - -```python -model = RevChatGPTModelv4(access_token=os.getenv("OPENAI_API_KEY"), **config) -``` - -3. Interact: - -```python -response = model.run("Tell me a fun fact!") -print(response) -``` - -### Credits & Disclaimers - -- This project is not an official OpenAI product and is not affiliated with OpenAI. Use at your own discretion. -- Many thanks to all the contributors who have made this project possible. -- Special acknowledgment to [virtualharby](https://www.youtube.com/@virtualharby) for the motivating music! - ---- - -By following this guide, you should now have a clear understanding of how to use the Reverse Engineered ChatGPT API and its abstraction. Happy coding! diff --git a/docs/examples/stacked_worker.md b/docs/examples/stacked_worker.md deleted file mode 100644 index 914d1754..00000000 --- a/docs/examples/stacked_worker.md +++ /dev/null @@ -1,338 +0,0 @@ -# Tutorial: Understanding and Utilizing Worker Examples - -## Table of Contents -1. Introduction -2. Code Overview - - Import Statements - - Initializing API Key and Language Model - - Creating Swarm Tools - - Appending Tools to a List - - Initializing a Worker Node -3. Understanding the `hf_agent` Tool -4. Understanding the `omni_agent` Tool -5. Understanding the `compile` Tool -6. Running a Swarm -7. Interactive Examples - - Example 1: Initializing API Key and Language Model - - Example 2: Using the `hf_agent` Tool - - Example 3: Using the `omni_agent` Tool - - Example 4: Using the `compile` Tool -8. Conclusion - -## 1. Introduction -The provided code showcases a system built around a worker node that utilizes various AI models and tools to perform tasks. This tutorial will break down the code step by step, explaining its components, how they work together, and how to utilize its modularity for various tasks. - -## 2. Code Overview - -### Import Statements -The code begins with import statements, bringing in necessary modules and classes. Key imports include the `OpenAIChat` class, which represents a language model, and several custom agents and tools from the `swarms` package. - -```python -import interpreter # Assuming this is a custom module - -from swarms.agents.hf_agents import HFAgent -from swarms.agents.omni_modal_agent import OmniModalAgent -from swarms.models import OpenAIChat -from swarms.tools.autogpt import tool -from swarms.workers import Worker -``` - -### Initializing API Key and Language Model -Here, an API key is initialized, and a language model (`OpenAIChat`) is created. This model is capable of generating human-like text based on the provided input. - -```python -# Initialize API Key -api_key = "YOUR_OPENAI_API_KEY" - -# Initialize the language model -llm = OpenAIChat( - openai_api_key=api_key, - temperature=0.5, -) -``` - -### Creating Swarm Tools -The code defines three tools: `hf_agent`, `omni_agent`, and `compile`. These tools encapsulate specific functionalities and can be invoked to perform tasks. - -### Appending Tools to a List -All defined tools are appended to a list called `tools`. This list is later used when initializing a worker node, allowing the node to access and utilize these tools. - -```python -# Append tools to a list -tools = [hf_agent, omni_agent, compile] -``` - -### Initializing a Worker Node -A worker node is initialized using the `Worker` class. The worker node is equipped with the language model, a name, API key, and the list of tools. It's set up to perform tasks without human intervention. - -```python -# Initialize a single Worker node with previously defined tools in addition to its predefined tools -node = Worker( - llm=llm, - ai_name="Optimus Prime", - openai_api_key=api_key, - ai_role="Worker in a swarm", - external_tools=tools, - human_in_the_loop=False, - temperature=0.5, -) -``` - -## 3. Understanding the `hf_agent` Tool -The `hf_agent` tool utilizes an OpenAI model (`text-davinci-003`) to perform tasks. It takes a task as input and returns a response. This tool is suitable for multi-modal tasks like generating images, videos, speech, etc. The tool's primary rule is not to be used for simple tasks like generating summaries. - -```python -@tool -def hf_agent(task: str = None): - # Create an HFAgent instance with the specified model and API key - agent = HFAgent(model="text-davinci-003", api_key=api_key) - # Run the agent with the provided task and optional text input - response = agent.run(task, text="¡Este es un API muy agradable!") - return response -``` - -## 4. Understanding the `omni_agent` Tool -The `omni_agent` tool is more versatile and leverages the `llm` (language model) to interact with Huggingface models for various tasks. It's intended for multi-modal tasks such as document-question-answering, image-captioning, summarization, and more. The tool's rule is also not to be used for simple tasks. - -```python -@tool -def omni_agent(task: str = None): - # Create an OmniModalAgent instance with the provided language model - agent = OmniModalAgent(llm) - # Run the agent with the provided task - response = agent.run(task) - return response -``` - -## 5. Understanding the `compile` Tool -The `compile` tool allows the execution of code locally, supporting various programming languages like Python, JavaScript, and Shell. It provides a natural language interface to your computer's capabilities. Users can chat with this tool in a terminal-like interface to perform tasks such as creating and editing files, controlling a browser, and more. - -```python -@tool -def compile(task: str): - # Use the interpreter module to chat with the local interpreter - task = interpreter.chat(task, return_messages=True) - interpreter.chat() - interpreter.reset(task) - - # Set environment variables for the interpreter - os.environ["INTERPRETER_CLI_AUTO_RUN"] = True - os.environ["INTERPRETER_CLI_FAST_MODE"] = True - os.environ["INTERPRETER_CLI_DEBUG"] = True -``` - -## 6. Running a Swarm -After defining tools and initializing the worker node, a specific task is provided as input to the worker node. The node then runs the task, and the response is printed to the console. - -```python -# Specify the task -task = "What were the winning Boston Marathon times for the past 5 years (ending in 2022)? Generate a table of the year, name, country of origin, and times." - -# Run the node on the task -response = node.run(task) - -# Print the response -print(response) -``` - - -## Full Code -- The full code example of stacked swarms - -```python -import os - -import interpreter - -from swarms.agents.hf_agents import HFAgent -from swarms.agents.omni_modal_agent import OmniModalAgent -from swarms.models import OpenAIChat -from swarms.tools.autogpt import tool -from swarms.workers import Worker - -# Initialize API Key -api_key = "" - - -# Initialize the language model, -# This model can be swapped out with Anthropic, ETC, Huggingface Models like Mistral, ETC -llm = OpenAIChat( - openai_api_key=api_key, - temperature=0.5, -) - - -# wrap a function with the tool decorator to make it a tool, then add docstrings for tool documentation -@tool -def hf_agent(task: str = None): - """ - An tool that uses an openai model to call and respond to a task by search for a model on huggingface - It first downloads the model then uses it. - - Rules: Don't call this model for simple tasks like generating a summary, only call this tool for multi modal tasks like generating images, videos, speech, etc - - """ - agent = HFAgent(model="text-davinci-003", api_key=api_key) - response = agent.run(task, text="¡Este es un API muy agradable!") - return response - - -# wrap a function with the tool decorator to make it a tool -@tool -def omni_agent(task: str = None): - """ - An tool that uses an openai Model to utilize and call huggingface models and guide them to perform a task. - - Rules: Don't call this model for simple tasks like generating a summary, only call this tool for multi modal tasks like generating images, videos, speech - The following tasks are what this tool should be used for: - - Tasks omni agent is good for: - -------------- - document-question-answering - image-captioning - image-question-answering - image-segmentation - speech-to-text - summarization - text-classification - text-question-answering - translation - huggingface-tools/text-to-image - huggingface-tools/text-to-video - text-to-speech - huggingface-tools/text-download - huggingface-tools/image-transformation - """ - agent = OmniModalAgent(llm) - response = agent.run(task) - return response - - -# Code Interpreter -@tool -def compile(task: str): - """ - Open Interpreter lets LLMs run code (Python, Javascript, Shell, and more) locally. - You can chat with Open Interpreter through a ChatGPT-like interface in your terminal - by running $ interpreter after installing. - - This provides a natural-language interface to your computer's general-purpose capabilities: - - Create and edit photos, videos, PDFs, etc. - Control a Chrome browser to perform research - Plot, clean, and analyze large datasets - ...etc. - ⚠️ Note: You'll be asked to approve code before it's run. - - Rules: Only use when given to generate code or an application of some kind - """ - task = interpreter.chat(task, return_messages=True) - interpreter.chat() - interpreter.reset(task) - - os.environ["INTERPRETER_CLI_AUTO_RUN"] = True - os.environ["INTERPRETER_CLI_FAST_MODE"] = True - os.environ["INTERPRETER_CLI_DEBUG"] = True - - -# Append tools to an list -tools = [hf_agent, omni_agent, compile] - - -# Initialize a single Worker node with previously defined tools in addition to it's -# predefined tools -node = Worker( - llm=llm, - ai_name="Optimus Prime", - openai_api_key=api_key, - ai_role="Worker in a swarm", - external_tools=tools, - human_in_the_loop=False, - temperature=0.5, -) - -# Specify task -task = "What were the winning boston marathon times for the past 5 years (ending in 2022)? Generate a table of the year, name, country of origin, and times." - -# Run the node on the task -response = node.run(task) - -# Print the response -print(response) -``` - - -## 8. Conclusion -In this extensive tutorial, we've embarked on a journey to explore a sophisticated system designed to harness the power of AI models and tools for a myriad of tasks. We've peeled back the layers of code, dissected its various components, and gained a profound understanding of how these elements come together to create a versatile, modular, and powerful swarm-based AI system. - -## What We've Learned - -Throughout this tutorial, we've covered the following key aspects: - -### Code Structure and Components -We dissected the code into its fundamental building blocks: -- **Import Statements:** We imported necessary modules and libraries, setting the stage for our system's functionality. -- **Initializing API Key and Language Model:** We learned how to set up the essential API key and initialize the language model, a core component for text generation and understanding. -- **Creating Swarm Tools:** We explored how to define tools, encapsulating specific functionalities that our system can leverage. -- **Appending Tools to a List:** We aggregated our tools into a list, making them readily available for use. -- **Initializing a Worker Node:** We created a worker node equipped with tools, a name, and configuration settings. - -### Tools and Their Functions -We dove deep into the purpose and functionality of three crucial tools: -- **`hf_agent`:** We understood how this tool employs an OpenAI model for multi-modal tasks, and its use cases beyond simple summarization. -- **`omni_agent`:** We explored the versatility of this tool, guiding Huggingface models to perform a wide range of multi-modal tasks. -- **`compile`:** We saw how this tool allows the execution of code in multiple languages, providing a natural language interface for various computational tasks. - -### Interactive Examples -We brought the code to life through interactive examples, showcasing how to initialize the language model, generate text, perform document-question-answering, and execute code—all with practical, real-world scenarios. - -## A Recap: The Worker Node's Role - -At the heart of this system lies the "Worker Node," a versatile entity capable of wielding the power of AI models and tools to accomplish tasks. The Worker Node's role is pivotal in the following ways: - -1. **Task Execution:** It is responsible for executing tasks, harnessing the capabilities of the defined tools to generate responses or perform actions. - -2. **Modularity:** The Worker Node benefits from the modularity of the system. It can easily access and utilize a variety of tools, allowing it to adapt to diverse tasks and requirements. - -3. **Human in the Loop:** While the example here is configured to operate without human intervention, the Worker Node can be customized to incorporate human input or approval when needed. - -4. **Integration:** It can be extended to integrate with other AI models, APIs, or services, expanding its functionality and versatility. - -## The Road Ahead: Future Features and Enhancements - -As we conclude this tutorial, let's peek into the future of this system. While the current implementation is already powerful, there is always room for growth and improvement. Here are some potential future features and enhancements to consider: - -### 1. Enhanced Natural Language Understanding - - **Semantic Understanding:** Improve the system's ability to understand context and nuances in natural language, enabling more accurate responses. - -### 2. Multimodal Capabilities - - **Extended Multimodal Support:** Expand the `omni_agent` tool to support additional types of multimodal tasks, such as video generation or audio processing. - -### 3. Customization and Integration - - **User-defined Tools:** Allow users to define their own custom tools, opening up endless possibilities for tailoring the system to specific needs. - -### 4. Collaborative Swarms - - **Swarm Collaboration:** Enable multiple Worker Nodes to collaborate on complex tasks, creating a distributed, intelligent swarm system. - -### 5. User-Friendly Interfaces - - **Graphical User Interface (GUI):** Develop a user-friendly GUI for easier interaction and task management, appealing to a wider audience. - -### 6. Continuous Learning - - **Active Learning:** Implement mechanisms for the system to learn and adapt over time, improving its performance with each task. - -### 7. Security and Privacy - - **Enhanced Security:** Implement robust security measures to safeguard sensitive data and interactions within the system. - -### 8. Community and Collaboration - - **Open Source Community:** Foster an open-source community around the system, encouraging contributions and innovation from developers worldwide. - -### 9. Integration with Emerging Technologies - - **Integration with Emerging AI Models:** Keep the system up-to-date by seamlessly integrating with new and powerful AI models as they emerge in the industry. - -## In Conclusion - -In this tutorial, we've journeyed through a complex AI system, unraveling its inner workings, and understanding its potential. We've witnessed how code can transform into a powerful tool, capable of handling a vast array of tasks, from generating creative stories to executing code snippets. - -As we conclude, we stand at the threshold of an exciting future for AI and technology. This system, with its modular design and the potential for continuous improvement, embodies the spirit of innovation and adaptability. Whether you're a developer, a researcher, or an enthusiast, the possibilities are boundless, and the journey is just beginning. - -Embrace this knowledge, explore the system, and embark on your own quest to shape the future of AI. With each line of code, you have the power to transform ideas into reality and unlock new horizons of innovation. The future is yours to create, and the tools are at your fingertips. \ No newline at end of file diff --git a/docs/examples/tools_agents.md b/docs/examples/tools_agents.md deleted file mode 100644 index ebd7ffbe..00000000 --- a/docs/examples/tools_agents.md +++ /dev/null @@ -1,323 +0,0 @@ -# **Equipping Autonomous Agents with Tools** -========================================== - -Tools play a crucial role in enhancing the capabilities of AI agents. Swarms, a powerful open-source framework, provides a robust and flexible environment for building and integrating tools with AI agents. In this comprehensive guide, we'll explore the process of creating tools in Swarms, including the 3-step process, tool decorator, adding types and doc strings, and integrating them into the Agent class. - -## **Introduction to Swarms** --------------------------- - -Swarms is a Python-based framework that simplifies the development and deployment of AI agents. It provides a seamless integration with large language models (LLMs) and offers a wide range of tools and utilities to streamline the agent development process. One of the core features of Swarms is the ability to create and integrate custom tools, which can significantly extend the capabilities of AI agents. - -Learn more here: - -[**GitHub - kyegomez/swarms: Build, Deploy, and Scale Reliable Swarms of Autonomous Agents for...**](https://github.com/kyegomez/swarms?source=post_page-----49d146bcbf9e--------------------------------) --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- - -### [Build, Deploy, and Scale Reliable Swarms of Autonomous Agents for Workflow Automation. Join our Community...](https://github.com/kyegomez/swarms?source=post_page-----49d146bcbf9e--------------------------------) - -[github.com](https://github.com/kyegomez/swarms?source=post_page-----49d146bcbf9e--------------------------------) - -And, join our community for real-time support and conversations with friends! - -[**Join the Agora Discord Server!**](https://discord.gg/A8DrG5nj?source=post_page-----49d146bcbf9e--------------------------------) ------------------------------------------------------------------------------------------------------------------------------------ - -### [Advancing Humanity through open source AI research. | 6319 members](https://discord.gg/A8DrG5nj?source=post_page-----49d146bcbf9e--------------------------------) - -[discord.gg](https://discord.gg/A8DrG5nj?source=post_page-----49d146bcbf9e--------------------------------) - -**Installation** -================ - -First, download swarms with the following command. If you have any questions please refer to this video or ask us in the discord! - -```bash -pip3 install -U swarms -``` - -**Necessary Imports** ---------------------- - -Before we dive into the process of creating tools in Swarms, let's familiarize ourselves with the necessary imports: - -```python -from swarms import Agent, Anthropic, tool -import subprocess -``` - -- These imports provide access to the core components of the Swarms framework, including the `Agent` class, the `Anthropic` language model, and the `tool` decorator for creating custom tools. - -- `import subprocess`: This import allows us to interact with the system's terminal and execute shell commands, which can be useful for certain types of tools. - -With these imports in place, we're ready to explore the process of creating tools in Swarms. - -### **The 3-Step Process** -====================== - -Creating tools in Swarms follows a straightforward 3-step process: - -1\. Define the Tool Function\ -2\. Decorate the Function with `[@tool](http://twitter.com/tool)` and add documentation with type hints.\ -3\. Add the Tool to the `Agent` Instance - -Let's go through each step in detail, accompanied by code examples. - -### **Step 1: Define the Tool Function** ------------------------------------- - -The first step in creating a tool is to define a Python function that encapsulates the desired functionality. This function will serve as the core logic for your tool. Here's an example of a tool function that allows you to execute code in the terminal: - -```python -def terminal(code: str) -> str: - """ - Run code in the terminal. - - Args: - code (str): The code to run in the terminal. - - Returns: - str: The output of the code. - """ - out = subprocess.run( - code, shell=True, capture_output=True, text=True - ).stdout - return str(out) -``` - -In this example, the `terminal` function takes a string `code` as input and uses the `subprocess` module to execute the provided code in the system's terminal. The output of the code is captured and returned as a string. - -### **Let's break down the components of this function:** ------------------------------------------------------ - -- **Function Signature:** The function signature `def terminal(code: str) -> str:` defines the function name (`terminal`), the parameter name and type (`code: str`), and the return type (`-> str`). This adheres to Python's type hinting conventions. - -- **Docstring**: The multiline string enclosed in triple quotes (`"""` ... `"""`) is a docstring, which provides a brief description of the function, its parameters, and its return value. Docstrings are essential for documenting your code and making it easier for the agent to understand and use your tools. - -- **Function Body**: The body of the function contains the actual logic for executing the code in the terminal. It uses the `subprocess.run` function to execute the provided `code` in the shell, capturing the output (`capture_output=True`), and returning the output as text (`text=True`). The `stdout` attribute of the `CompletedProcess` object contains the captured output, which is converted to a string and returned. - -This is a simple example, but it demonstrates the key components of a tool function: a well-defined signature with type hints, a descriptive docstring, and the core logic encapsulated within the function body. - -### **Step 2: Decorate the Function with `**[**@tool**](http://twitter.com/tool)**`** ---------------------------------------------------------------------------------- - -After defining the tool function, the next step is to decorate it with the `[@tool](http://twitter.com/tool)` decorator provided by Swarms. This decorator registers the function as a tool, allowing it to be used by AI agents within the Swarms framework. - -Here's how you would decorate the `terminal` function from the previous example: - -```python -@tool -def terminal(code: str) -> str: - """ - Run code in the terminal. - - Args: - code (str): The code to run in the terminal. - - Returns: - str: The output of the code. - """ - out = subprocess.run( - code, shell=True, capture_output=True, text=True - ).stdout - return str(out) -``` - -The `[@tool](http://twitter.com/tool)` decorator is placed directly above the function definition. This decorator performs the necessary registration and configuration steps to integrate the tool with the Swarms framework. - -### **Step 3: Add the Tool to the `Agent` Instance** ------------------------------------------------- - -The final step in creating a tool is to add it to the `Agent` instance. The `Agent` class is a core component of the Swarms framework and represents an AI agent capable of interacting with humans and other agents, as well as utilizing the available tools. - -Here's an example of how to create an `Agent` instance and add the `terminal` tool: - -```python -# Model -llm = Anthropic( - temperature=0.1, -) - -# Agent -agent = Agent( - agent_name="Devin", - system_prompt=( - "Autonomous agent that can interact with humans and other" - " agents. Be Helpful and Kind. Use the tools provided to" - " assist the user. Return all code in markdown format." - ), - llm=llm, - max_loops="auto", - autosave=True, - dashboard=False, - streaming_on=True, - verbose=True, - stopping_token="", - interactive=True, - tools=[terminal], - code_interpreter=True, -) -``` - -In this example, we first create an instance of the `Anthropic` language model, which will be used by the agent for natural language - -**The Necessity of Documentation** ----------------------------------- - -Before creating tools, it's essential to understand the importance of documentation. Clear and concise documentation ensures that your code is easily understandable and maintainable, not only for yourself but also for other developers who may work with your codebase in the future. - -Effective documentation serves several purposes: - -1. Code Clarity: Well-documented code is easier to read and understand, making it more accessible for both developers and non-technical stakeholders. - -2. Collaboration: When working in a team or contributing to open-source projects, proper documentation facilitates collaboration and knowledge sharing. - -3. Onboarding: Comprehensive documentation can significantly streamline the onboarding process for new team members or contributors, reducing the time required to familiarize themselves with the codebase. - -4. Future Maintenance: As projects evolve and requirements change, well-documented code becomes invaluable for future maintenance and updates. - -In the context of creating tools in Swarms, documentation plays a vital role in ensuring that your tools are easily discoverable, understandable, and usable by other developers and AI agents. - -**Type Handling** ------------------ - -Python is a dynamically-typed language, which means that variables can hold values of different types during runtime. While this flexibility can be advantageous in certain scenarios, it can also lead to potential errors and inconsistencies, especially in larger codebases. - -Type hints, introduced in Python 3.5, provide a way to explicitly annotate the expected types of variables, function parameters, and return values. By incorporating type hints into your code, you can: - -1. Improve Code Readability: Type hints make it easier for developers to understand the expected data types, reducing the risk of introducing bugs due to type-related errors. - -2. Enable Static Type Checking: With tools like mypy, you can perform static type checking, catching potential type-related issues before running the code. - -3. Enhance Code Completion and Tooling: Modern IDEs and code editors can leverage type hints to provide better code completion, refactoring capabilities, and inline documentation. - -In the context of creating tools in Swarms, type hints are crucial for ensuring that your tools are used correctly by AI agents and other developers. By clearly defining the expected input and output types, you can reduce the likelihood of runtime errors and improve the overall reliability of your tools. - -Now, let's continue with other tool examples! - -### **Additional Tool Examples** -============================ - -To further illustrate the process of creating tools in Swarms, let's explore a few more examples of tool functions with varying functionalities. - -**Browser Tool** - -```python -@tool -def browser(query: str) -> str: - """ - Search the query in the browser with the `browser` tool. - Args: - query (str): The query to search in the browser. - Returns: - str: The search results. - """ - import webbrowser - url = f"https://www.google.com/search?q={query}" - webbrowser.open(url) - return f"Searching for {query} in the browser." -``` - -The `browser` tool allows the agent to perform web searches by opening the provided `query` in the default web browser. It leverages the `webbrowser` module to construct the search URL and open it in the browser. The tool returns a string indicating that the search is being performed. - -**File Creation Tool** - -```python -@tool -def create_file(file_path: str, content: str) -> str: - """ - Create a file using the file editor tool. - Args: - file_path (str): The path to the file. - content (str): The content to write to the file. - Returns: - str: The result of the file creation operation. - """ - with open(file_path, "w") as file: - file.write(content) - return f"File {file_path} created successfully." -``` - -The `create_file` tool allows the agent to create a new file at the specified `file_path` with the provided `content`. It uses Python's built-in `open` function in write mode (`"w"`) to create the file and write the content to it. The tool returns a string indicating the successful creation of the file. - -**File Editor Tool** - -```python -@tool -def file_editor(file_path: str, mode: str, content: str) -> str: - """ - Edit a file using the file editor tool. - Args: - file_path (str): The path to the file. - mode (str): The mode to open the file in. - content (str): The content to write to the file. - Returns: - str: The result of the file editing operation. - """ - with open(file_path, mode) as file: - file.write(content) - return f"File {file_path} edited successfully." -``` - -The `file_editor` tool is similar to the `create_file` tool but provides more flexibility by allowing the agent to specify the mode in which the file should be opened (e.g., `"w"` for write, `"a"` for append). It writes the provided `content` to the file at the specified `file_path` and returns a string indicating the successful editing of the file. - -These examples demonstrate the versatility of tools in Swarms and how they can be designed to perform a wide range of tasks, from executing terminal commands to interacting with files and performing web searches. - -### **Plugging Tools into the Agent** -================================= - -After defining and decorating your tool functions, the next step is to integrate them into the `Agent` instance. This process involves passing the tools as a list to the `tools` parameter when creating the `Agent` instance. - -```python -# Model -llm = Anthropic(temperature=0.1) -# Tools -tools = [terminal, browser, create_file, file_editor] -# Agent -agent = Agent( - agent_name="Devin", - system_prompt=( - "Autonomous agent that can interact with humans and other" - " agents. Be Helpful and Kind. Use the tools provided to" - " assist the user. Return all code in markdown format." - ), - llm=llm, - max_loops="auto", - autosave=True, - dashboard=False, - streaming_on=True, - verbose=True, - stopping_token="", - interactive=True, - tools=tools, - code_interpreter=True, -) -``` - -In this example, we create a list `tools` containing the `terminal`, `browser`, `create_file`, and `file_editor` tools. This list is then passed to the `tools` parameter when creating the `Agent` instance. - -Once the `Agent` instance is created with the provided tools, it can utilize these tools to perform various tasks and interact with external systems. The agent can call these tools as needed, passing the required arguments and receiving the corresponding return values. - -### **Using Tools in Agent Interactions** -===================================== - -After integrating the tools into the `Agent` instance, you can utilize them in your agent's interactions with humans or other agents. Here's an example of how an agent might use the `terminal` tool: - -``` -out = agent("Run the command 'ls' in the terminal.") -print(out) -``` - -In this example, the human user instructs the agent to run the `"ls"` command in the terminal. The agent processes this request and utilizes the `terminal` tool to execute the command, capturing and returning the output. - -Similarly, the agent can leverage other tools, such as the `browser` tool for web searches or the `file_editor` tool for creating and modifying files, based on the user's instructions. - -### **Conclusion** -============== - -Creating tools in Swarms is a powerful way to extend the capabilities of AI agents and enable them to interact with external systems and perform a wide range of tasks. By following the 3-step process of defining the tool function, decorating it with `@tool`, and adding it to the `Agent` instance, you can seamlessly integrate custom tools into your AI agent's workflow. - -Throughout this blog post, we explored the importance of documentation and type handling, which are essential for maintaining code quality, facilitating collaboration, and ensuring the correct usage of your tools by other developers and AI agents. - -We also covered the necessary imports and provided detailed code examples for various types of tools, such as executing terminal commands, performing web searches, and creating and editing files. These examples demonstrated the flexibility and versatility of tools in Swarms, allowing you to tailor your tools to meet your specific project requirements. - -By leveraging the power of tools in Swarms, you can empower your AI agents with diverse capabilities, enabling them to tackle complex tasks, interact with external systems, and provide more comprehensive and intelligent solutions. diff --git a/docs/examples/worker.md b/docs/examples/worker.md deleted file mode 100644 index cd082aa4..00000000 --- a/docs/examples/worker.md +++ /dev/null @@ -1,115 +0,0 @@ -# **The Ultimate Guide to Mastering the `Worker` Class from Swarms** - ---- - -**Table of Contents** - -1. Introduction: Welcome to the World of the Worker -2. The Basics: What Does the Worker Do? -3. Installation: Setting the Stage -4. Dive Deep: Understanding the Architecture -5. Practical Usage: Let's Get Rolling! -6. Advanced Tips and Tricks -7. Handling Errors: Because We All Slip Up Sometimes -8. Beyond the Basics: Advanced Features and Customization -9. Conclusion: Taking Your Knowledge Forward - ---- - -**1. Introduction: Welcome to the World of the Worker** - -Greetings, future master of the `Worker`! Step into a universe where you can command an AI worker to perform intricate tasks, be it searching the vast expanse of the internet or crafting multi-modality masterpieces. Ready to embark on this thrilling journey? Let’s go! - ---- - -**2. The Basics: What Does the Worker Do?** - -The `Worker` is your personal AI assistant. Think of it as a diligent bee in a swarm, ready to handle complex tasks across various modalities, from text and images to audio and beyond. - ---- - -**3. Installation: Setting the Stage** - -Before we can call upon our Worker, we need to set the stage: - -```bash -pip install swarms -``` - -Voila! You’re now ready to summon your Worker. - ---- - -**4. Dive Deep: Understanding the Architecture** - -- **Language Model (LLM)**: The brain of our Worker. It understands and crafts intricate language-based responses. -- **Tools**: Think of these as the Worker's toolkit. They range from file tools, website querying, to even complex tasks like image captioning. -- **Memory**: No, our Worker doesn’t forget. It employs a sophisticated memory mechanism to remember past interactions and learn from them. - ---- - -**5. Practical Usage: Let's Get Rolling!** - -Here’s a simple way to invoke the Worker and give it a task: - -```python -from swarms import Worker -from swarms.models import OpenAIChat - -llm = OpenAIChat( - # enter your api key - openai_api_key="", - temperature=0.5, -) - -node = Worker( - llm=llm, - ai_name="Optimus Prime", - openai_api_key="", - ai_role="Worker in a swarm", - external_tools=None, - human_in_the_loop=False, - temperature=0.5, -) - -task = "What were the winning boston marathon times for the past 5 years (ending in 2022)? Generate a table of the year, name, country of origin, and times." -response = node.run(task) -print(response) -``` - - -The result? An agent with elegantly integrated tools and long term memories - ---- - -**6. Advanced Tips and Tricks** - -- **Streaming Responses**: Want your Worker to respond in a more dynamic fashion? Use the `_stream_response` method to get results token by token. -- **Human-in-the-Loop**: By setting `human_in_the_loop` to `True`, you can involve a human in the decision-making process, ensuring the best results. - ---- - -**7. Handling Errors: Because We All Slip Up Sometimes** - -Your Worker is designed to be robust. But if it ever encounters a hiccup, it's equipped to let you know. Error messages are crafted to be informative, guiding you on the next steps. - ---- - -**8. Beyond the Basics: Advanced Features and Customization** - -- **Custom Tools**: Want to expand the Worker's toolkit? Use the `external_tools` parameter to integrate your custom tools. -- **Memory Customization**: You can tweak the Worker's memory settings, ensuring it remembers what's crucial for your tasks. - ---- - -**9. Conclusion: Taking Your Knowledge Forward** - -Congratulations! You’re now well-equipped to harness the power of the `Worker` from Swarms. As you venture further, remember: the possibilities are endless, and with the Worker by your side, there’s no task too big! - -**Happy Coding and Exploring!** 🚀🎉 - ---- - -*Note*: This guide provides a stepping stone to the vast capabilities of the `Worker`. Dive into the official documentation for a deeper understanding and stay updated with the latest features. - ---- \ No newline at end of file diff --git a/docs/index.md b/docs/index.md index c28ab5b5..e5f963a7 100644 --- a/docs/index.md +++ b/docs/index.md @@ -1,89 +1,131 @@ -
-

- - - -

-
- -## 👋 Hello - -Swarms provides you with all the building blocks you need to build reliable, production-grade, and scalable multi-agent apps! - -## 💻 Install - -You can install `swarms` with pip in a -[**Python>=3.10**](https://www.python.org/) environment. - -!!! example "pip install (recommended)" - - === "headless" - The headless installation of `swarms` is designed for environments where graphical user interfaces (GUI) are not needed, making it more lightweight and suitable for server-side applications. - - ```bash - pip install swarms - ``` - - -!!! example "git clone (for development)" - - === "virtualenv" - - ```bash - # clone repository and navigate to root directory - git clone https://github.com/kyegomez/swarms.git - cd swarms - - # setup python environment and activate it - python3 -m venv venv - source venv/bin/activate - pip install --upgrade pip - - # headless install - pip install -e "." - - # desktop install - pip install -e ".[desktop]" - ``` - - === "poetry" - - ```bash - # clone repository and navigate to root directory - git clone https://github.com/kyegomez/swarms.git - cd swarms - - # setup python environment and activate it - poetry env use python3.10 - poetry shell - - # headless install - poetry install - - # desktop install - poetry install --extras "desktop" - ``` - -!!! example "NPM install |WIP|" - - === "headless" - Get started with the NPM implementation of Swarms with this command: - - ```bash - npm install swarms-js - ``` - - -## Documentation - -[Learn more about swarms →](swarms/) - - -## Examples - -Check out Swarms examples for building agents, data retrieval, and more. - -[Checkout Swarms examples →](examples/) +# Swarms + +Orchestrate enterprise-grade agents for multi-agent collaboration and orchestration to automate real-world problems. + +
+
+

Core Concepts

+ +
+
+

How-To Guides

+ +
+
+

Examples

+ +
+
+

Community

+ +
+
\ No newline at end of file diff --git a/docs/mkdocs.yml b/docs/mkdocs.yml index b887b9b7..df0abaa6 100644 --- a/docs/mkdocs.yml +++ b/docs/mkdocs.yml @@ -41,20 +41,14 @@ plugins: # token: !ENV ["GITHUB_TOKEN"] - git-revision-date-localized: enable_creation_date: true - -copyright: "© TGSC, Corporation." extra_css: - assets/css/extra.css extra: social: - - icon: fontawesome/solid/house - link: assets/img/SwarmsLogoIcon.png - - icon: fontawesome/brands/discord - link: https://discord.gg/qUtxnK2NMf + - icon: fontawesome/brands/twitter + link: https://x.com/KyeGomezB - icon: fontawesome/brands/github - link: https://github.com/kyegomez/Swarms/ - - icon: fontawesome/brands/python - link: https://pypi.org/project/Swarms/ + link: https://github.com/kyegomez/swarms theme: name: material custom_dir: overrides @@ -95,44 +89,45 @@ markdown_extensions: - def_list - footnotes nav: -- Home: - - Installation: - - Overview: "index.md" - - Install: "install.md" - - Docker Setup: docker_setup.md - - Usage Examples: - - Build an Agent: "diy_your_own_agent.md" - - Build an Agent with tools: "examples/tools_agents.md" - # Add an examples with building agents - # Add multiple blogs on orchestrating agents with different types of frameworks - - Why does Swarms Exist?: - - Why Swarms? Orchestrating Agents for Enterprise Automation: "why.md" - - Limitations of Individual Agents: "limits_of_individual_agents.md" - - References: - - Agent Glossary: "swarms/glossary.md" - - List of The Best Multi-Agent Papers: "swarms/papers.md" -- Swarms Cloud API: - - Overview: "swarms_cloud/main.md" - - Available Models: "swarms_cloud/available_models.md" - - Migrate from OpenAI to Swarms in 3 lines of code: "swarms_cloud/migrate_openai.md" - - Getting Started with SOTA Vision Language Models VLM: "swarms_cloud/getting_started.md" - - Enterprise Guide to High-Performance Multi-Agent LLM Deployments: "swarms_cloud/production_deployment.md" - - Under The Hood The Swarm Cloud Serving Infrastructure: "swarms_cloud/architecture.md" -- Swarms Framework [PY]: + - Home: + - Installation: + - Overview: "index.md" + - Install: "swarms/install/install.md" + - Docker Setup: "swarms/install/docker_setup.md" + - Usage Examples: + - Models: + - How to Create A Custom Language Model: "swarms/models/custom_model.md" + - Models Available: "swarms/models/index.md" + - MultiModal Models Available: "swarms/models/multimodal_models.md" + - Agents: + - Getting started with Agents: "swarms/structs/diy_your_own_agent.md" + - Tools: + - Functions, Pydantic BaseModels, and More: "swarms/tools/main.md" + - Memory: + - Building Custom Vector Memory Databases with the BaseVectorDatabase Class: "swarms/memory/diy_memory.md" + - ShortTermMemory: "swarms/memory/short_term_memory.md" + - Multi-Agent Collaboration: + - SwarmNetwork: "swarms/structs/swarmnetwork.md" + - AgentRearrange: "swarms/structs/agent_rearrange.md" + - Why does Swarms Exist?: + - References: + - Agent Glossary: "swarms/glossary.md" + - List of The Best Multi-Agent Papers: "swarms/papers.md" + - Contributors: + - Contributing: "contributing.md" +- Reference: - Overview: "swarms/index.md" - - DIY Build Your Own Agent: "diy_your_own_agent.md" - - Agents with Tools: "examples/tools_agent.md" - # - Multi-Agent Orchestration: "swarms/structs/multi_agent_orchestration.md" + - Framework Structure: "swarms/structs/tree.md" - swarms.models: - How to Create A Custom Language Model: "swarms/models/custom_model.md" - Deploying Azure OpenAI in Production A Comprehensive Guide: "swarms/models/azure_openai.md" - - Language Models Available: + - Language Models: - BaseLLM: "swarms/models/base_llm.md" - Overview: "swarms/models/index.md" - HuggingFaceLLM: "swarms/models/huggingface.md" - Anthropic: "swarms/models/anthropic.md" - OpenAIChat: "swarms/models/openai.md" - - MultiModal Models Available: + - MultiModal Models : - BaseMultiModalModel: "swarms/models/base_multimodal_model.md" - Fuyu: "swarms/models/fuyu.md" - Vilt: "swarms/models/vilt.md" @@ -141,6 +136,7 @@ nav: - Nougat: "swarms/models/nougat.md" - Dalle3: "swarms/models/dalle3.md" - GPT4VisionAPI: "swarms/models/gpt4v.md" + - GPT4o: "swarms/models/gpt4o.md" - swarms.structs: - Foundational Structures: - Agent: "swarms/structs/agent.md" @@ -158,47 +154,21 @@ nav: - MajorityVoting: "swarms/structs/majorityvoting.md" - AgentRearrange: "swarms/structs/agent_rearrange.md" - RoundRobin: "swarms/structs/round_robin_swarm.md" - - swarms.memory: - - Building Custom Vector Memory Databases with the BaseVectorDatabase Class: "swarms/memory/diy_memory.md" - - ShortTermMemory: "swarms/memory/short_term_memory.md" - - swarms.tools: - - The Swarms Tool System Functions, Pydantic BaseModels as Tools, and Radical Customization: "swarms/tools/main.md" +- Swarms Cloud API: + - Overview: "swarms_cloud/main.md" + - Available Models: "swarms_cloud/available_models.md" + - Migrate from OpenAI to Swarms in 3 lines of code: "swarms_cloud/migrate_openai.md" + - Getting Started with SOTA Vision Language Models VLM: "swarms_cloud/getting_started.md" + - Enterprise Guide to High-Performance Multi-Agent LLM Deployments: "swarms_cloud/production_deployment.md" + - Under The Hood The Swarm Cloud Serving Infrastructure: "swarms_cloud/architecture.md" - Guides: - - Agents: - - Building Custom Vector Memory Databases with the BaseVectorDatabase Class: "swarms/memory/diy_memory.md" + # - Building Custom Vector Memory Databases with the BaseVectorDatabase Class: "swarms/memory/diy_memory.md" + - Models: - How to Create A Custom Language Model: "swarms/models/custom_model.md" - Deploying Azure OpenAI in Production, A Comprehensive Guide: "swarms/models/azure_openai.md" + - Agents: + - Agent: "examples/flow.md" - DIY Build Your Own Agent: "diy_your_own_agent.md" - Equipping Autonomous Agents with Tools: "examples/tools_agent.md" - - Overview: "examples/index.md" - - Agents: - - Agent: "examples/flow.md" - - OmniAgent: "examples/omni_agent.md" - - Swarms: - - SequentialWorkflow: "examples/reliable_autonomous_agents.md" - - 2O+ Autonomous Agent Blogs: "examples/ideas.md" -- Applications: - - CustomerSupport: - - Overview: "applications/customer_support.md" - - Marketing: - - Overview: "applications/marketing_agencies.md" - - Operations: - - Intoducing The Swarm of Automated Business Analyts: "applications/business-analyst-agent.md" -- Corporate: - - Corporate Documents: - - Data Room: "corporate/data_room.md" - - The Swarm Memo: "corporate/swarm_memo.md" - - Corporate Architecture: "corporate/architecture.md" - - Flywheel: "corporate/flywheel.md" - - Sales: - - FAQ: "corporate/faq.md" - - Distribution: "corporate/distribution" - - Product: - - SwarmCloud: "corporate/swarm_cloud.md" - - Weaknesses of Langchain: "corporate/failures.md" - - Design: "corporate/design.md" - - Metric: "corporate/metric.md" - - Organization: - - FrontEnd Member Onboarding: "corporate/front_end_contributors.md" -- Contributors: - - Contributing: "contributing.md" + - Swarms: + - SequentialWorkflow: "examples/reliable_autonomous_agents.md" diff --git a/docs/limits_of_individual_agents.md b/docs/purpose/limits_of_individual_agents.md similarity index 100% rename from docs/limits_of_individual_agents.md rename to docs/purpose/limits_of_individual_agents.md diff --git a/docs/why.md b/docs/purpose/why.md similarity index 97% rename from docs/why.md rename to docs/purpose/why.md index 2e4f0bce..5293de23 100644 --- a/docs/why.md +++ b/docs/purpose/why.md @@ -2,6 +2,10 @@ In the rapidly evolving landscape of artificial intelligence (AI) and automation, a new paradigm is emerging: the orchestration of multiple agents working in collaboration to tackle complex tasks. This approach, embodied by the Swarms Framework, aims to address the fundamental limitations of individual agents and unlocks the true potential of AI-driven automation in enterprise operations. +Individual agents are plagued by the same issues: short term memory constraints, hallucinations, single task limitations, lack of collaboration, and cost inefficiences. + +[Learn more here from a list of compiled agent papers](https://github.com/kyegomez/awesome-multi-agent-papers) + ## The Purpose of Swarms: Overcoming Agent Limitations Individual agents, while remarkable in their own right, face several inherent challenges that hinder their ability to effectively automate enterprise operations at scale. These limitations include: diff --git a/docs/why_swarms.md b/docs/purpose/why_swarms.md similarity index 100% rename from docs/why_swarms.md rename to docs/purpose/why_swarms.md diff --git a/docs/swarms/index.md b/docs/swarms/index.md index c3d45d86..ac91ecdf 100644 --- a/docs/swarms/index.md +++ b/docs/swarms/index.md @@ -1,14 +1,22 @@ -# Swarms +
Orchestrate swarms of agents for production-grade applications. -Individual agents face five significant challenges that hinder their deployment in production: short memory, single-task threading, hallucinations, high cost, and lack of collaboration. Multi-agent collaboration offers a solution to all these issues. Swarms provides simple, reliable, and agile tools to create your own Swarm tailored to your specific needs. Currently, Swarms is being used in production by RBC, John Deere, and many AI startups. For more information on the unparalleled benefits of multi-agent collaboration, check out this GitHub repository for research papers or schedule a call with me! +[![GitHub issues](https://img.shields.io/github/issues/kyegomez/swarms)](https://github.com/kyegomez/swarms/issues) [![GitHub forks](https://img.shields.io/github/forks/kyegomez/swarms)](https://github.com/kyegomez/swarms/network) [![GitHub stars](https://img.shields.io/github/stars/kyegomez/swarms)](https://github.com/kyegomez/swarms/stargazers) [![GitHub license](https://img.shields.io/github/license/kyegomez/swarms)](https://github.com/kyegomez/swarms/blob/main/LICENSE)[![GitHub star chart](https://img.shields.io/github/stars/kyegomez/swarms?style=social)](https://star-history.com/#kyegomez/swarms)[![Dependency Status](https://img.shields.io/librariesio/github/kyegomez/swarms)](https://libraries.io/github/kyegomez/swarms) [![Downloads](https://static.pepy.tech/badge/swarms/month)](https://pepy.tech/project/swarms) +[![Join the Agora discord](https://img.shields.io/discord/1110910277110743103?label=Discord&logo=discord&logoColor=white&style=plastic&color=d7b023)![Share on Twitter](https://img.shields.io/twitter/url/https/twitter.com/cloudposse.svg?style=social&label=Share%20%40kyegomez/swarms)](https://twitter.com/intent/tweet?text=Check%20out%20this%20amazing%20AI%20project:%20&url=https%3A%2F%2Fgithub.com%2Fkyegomez%2Fswarms) [![Share on Facebook](https://img.shields.io/badge/Share-%20facebook-blue)](https://www.facebook.com/sharer/sharer.php?u=https%3A%2F%2Fgithub.com%2Fkyegomez%2Fswarms) [![Share on LinkedIn](https://img.shields.io/badge/Share-%20linkedin-blue)](https://www.linkedin.com/shareArticle?mini=true&url=https%3A%2F%2Fgithub.com%2Fkyegomez%2Fswarms&title=&summary=&source=) + +[![Share on Reddit](https://img.shields.io/badge/-Share%20on%20Reddit-orange)](https://www.reddit.com/submit?url=https%3A%2F%2Fgithub.com%2Fkyegomez%2Fswarms&title=Swarms%20-%20the%20future%20of%20AI) [![Share on Hacker News](https://img.shields.io/badge/-Share%20on%20Hacker%20News-orange)](https://news.ycombinator.com/submitlink?u=https%3A%2F%2Fgithub.com%2Fkyegomez%2Fswarms&t=Swarms%20-%20the%20future%20of%20AI) [![Share on Pinterest](https://img.shields.io/badge/-Share%20on%20Pinterest-red)](https://pinterest.com/pin/create/button/?url=https%3A%2F%2Fgithub.com%2Fkyegomez%2Fswarms&media=https%3A%2F%2Fexample.com%2Fimage.jpg&description=Swarms%20-%20the%20future%20of%20AI) [![Share on WhatsApp](https://img.shields.io/badge/-Share%20on%20WhatsApp-green)](https://api.whatsapp.com/send?text=Check%20out%20Swarms%20-%20the%20future%20of%20AI%20%23swarms%20%23AI%0A%0Ahttps%3A%2F%2Fgithub.com%2Fkyegomez%2Fswarms) + +
+ + +Individual agents face 5 significant challenges that hinder their deployment in production: short memory, single-task threading, hallucinations, high cost, and lack of collaboration. Multi-agent collaboration offers a solution to all these issues. Swarms provides simple, reliable, and agile tools to create your own Swarm tailored to your specific needs. Currently, Swarms is being used in production by RBC, John Deere, and many AI startups. ---- ## Install -`pip3 install -U swarms` +`$ pip3 install -U swarms` --- @@ -58,119 +66,12 @@ agent.run("Generate a 10,000 word blog on health and wellness.") ``` -### `ToolAgent` -ToolAgent is an agent that can use tools through JSON function calling. It intakes any open source model from huggingface and is extremely modular and plug in and play. We need help adding general support to all models soon. - - -```python -from pydantic import BaseModel, Field -from transformers import AutoModelForCausalLM, AutoTokenizer - -from swarms import ToolAgent -from swarms.utils.json_utils import base_model_to_json - -# Load the pre-trained model and tokenizer -model = AutoModelForCausalLM.from_pretrained( - "databricks/dolly-v2-12b", - load_in_4bit=True, - device_map="auto", -) -tokenizer = AutoTokenizer.from_pretrained("databricks/dolly-v2-12b") - - -# Initialize the schema for the person's information -class Schema(BaseModel): - name: str = Field(..., title="Name of the person") - agent: int = Field(..., title="Age of the person") - is_student: bool = Field( - ..., title="Whether the person is a student" - ) - courses: list[str] = Field( - ..., title="List of courses the person is taking" - ) - - -# Convert the schema to a JSON string -tool_schema = base_model_to_json(Schema) - -# Define the task to generate a person's information -task = ( - "Generate a person's information based on the following schema:" -) - -# Create an instance of the ToolAgent class -agent = ToolAgent( - name="dolly-function-agent", - description="Ana gent to create a child data", - model=model, - tokenizer=tokenizer, - json_schema=tool_schema, -) - -# Run the agent to generate the person's information -generated_data = agent.run(task) - -# Print the generated data -print(f"Generated data: {generated_data}") - -``` - - -### `Worker` -The `Worker` is a simple all-in-one agent equipped with an LLM, tools, and RAG for low level tasks. - -✅ Plug in and Play LLM. Utilize any LLM from anywhere and any framework - -✅ Reliable RAG: Utilizes FAISS for efficient RAG but it's modular so you can use any DB. - -✅ Multi-Step Parallel Function Calling: Use any tool - -```python -# Importing necessary modules -import os - -from dotenv import load_dotenv - -from swarms import OpenAIChat, Worker, tool - -# Loading environment variables from .env file -load_dotenv() - -# Retrieving the OpenAI API key from environment variables -api_key = os.getenv("OPENAI_API_KEY") - - -# Create a tool -@tool -def search_api(query: str): - pass - - -# Creating a Worker instance -worker = Worker( - name="My Worker", - role="Worker", - human_in_the_loop=False, - tools=[search_api], - temperature=0.5, - llm=OpenAIChat(openai_api_key=api_key), -) - -# Running the worker with a prompt -out = worker.run("Hello, how are you? Create an image of how your are doing!") - -# Printing the output -print(out) -``` - ------- - - # `Agent` with Long Term Memory `Agent` equipped with quasi-infinite long term memory. Great for long document understanding, analysis, and retrieval. ```python -from swarms import Agent, ChromaDB, OpenAIChat +from swarms import Agent, OpenAIChat +from playground.memory.chromadb_example import ChromaDB # Copy and paste the code and put it in your own local directory. # Making an instance of the ChromaDB class memory = ChromaDB( @@ -208,7 +109,7 @@ print(out) An LLM equipped with long term memory and tools, a full stack agent capable of automating all and any digital tasks given a good prompt. ```python -from swarms import Agent, ChromaDB, OpenAIChat, tool +from swarms import Agent, ChromaDB, OpenAIChat # Making an instance of the ChromaDB class memory = ChromaDB( @@ -219,7 +120,6 @@ memory = ChromaDB( ) # Initialize a tool -@tool def search_api(query: str): # Add your logic here return query @@ -248,189 +148,6 @@ print(out) ``` - - - - - - - ----- - -### `SequentialWorkflow` -Sequential Workflow enables you to sequentially execute tasks with `Agent` and then pass the output into the next agent and onwards until you have specified your max loops. `SequentialWorkflow` is wonderful for real-world business tasks like sending emails, summarizing documents, and analyzing data. - - -✅ Save and Restore Workflow states! - -✅ Multi-Modal Support for Visual Chaining - -✅ Utilizes Agent class - -```python -import os - -from dotenv import load_dotenv - -from swarms import Agent, OpenAIChat, SequentialWorkflow - -load_dotenv() - -# Load the environment variables -api_key = os.getenv("OPENAI_API_KEY") - - -# Initialize the language agent -llm = OpenAIChat( - temperature=0.5, model_name="gpt-4", openai_api_key=api_key, max_tokens=4000 -) - - -# Initialize the agent with the language agent -agent1 = Agent(llm=llm, max_loops=1) - -# Create another agent for a different task -agent2 = Agent(llm=llm, max_loops=1) - -# Create another agent for a different task -agent3 = Agent(llm=llm, max_loops=1) - -# Create the workflow -workflow = SequentialWorkflow(max_loops=1) - -# Add tasks to the workflow -workflow.add( - agent1, - "Generate a 10,000 word blog on health and wellness.", -) - -# Suppose the next task takes the output of the first task as input -workflow.add( - agent2, - "Summarize the generated blog", -) - -# Run the workflow -workflow.run() - -# Output the results -for task in workflow.tasks: - print(f"Task: {task.description}, Result: {task.result}") -``` - - - -### `ConcurrentWorkflow` -`ConcurrentWorkflow` runs all the tasks all at the same time with the inputs you give it! - - -```python -import os - -from dotenv import load_dotenv - -from swarms import Agent, ConcurrentWorkflow, OpenAIChat, Task - -# Load environment variables from .env file -load_dotenv() - -# Load environment variables -llm = OpenAIChat(openai_api_key=os.getenv("OPENAI_API_KEY")) -agent = Agent(llm=llm, max_loops=1) - -# Create a workflow -workflow = ConcurrentWorkflow(max_workers=5) - -# Create tasks -task1 = Task(agent, "What's the weather in miami") -task2 = Task(agent, "What's the weather in new york") -task3 = Task(agent, "What's the weather in london") - -# Add tasks to the workflow -workflow.add(tasks=[task1, task2, task3]) - -# Run the workflow -workflow.run() -``` - -### `RecursiveWorkflow` -`RecursiveWorkflow` will keep executing the tasks until a specific token like is located inside the text! - -```python -import os - -from dotenv import load_dotenv - -from swarms import Agent, OpenAIChat, RecursiveWorkflow, Task - -# Load environment variables from .env file -load_dotenv() - -# Load environment variables -llm = OpenAIChat(openai_api_key=os.getenv("OPENAI_API_KEY")) -agent = Agent(llm=llm, max_loops=1) - -# Create a workflow -workflow = RecursiveWorkflow(stop_token="") - -# Create tasks -task1 = Task(agent, "What's the weather in miami") -task2 = Task(agent, "What's the weather in new york") -task3 = Task(agent, "What's the weather in london") - -# Add tasks to the workflow -workflow.add(task1) -workflow.add(task2) -workflow.add(task3) - -# Run the workflow -workflow.run() -``` - - - -### `ModelParallelizer` -The ModelParallelizer allows you to run multiple models concurrently, comparing their outputs. This feature enables you to easily compare the performance and results of different models, helping you make informed decisions about which model to use for your specific task. - -Plug-and-Play Integration: The structure provides a seamless integration with various models, including OpenAIChat, Anthropic, Mixtral, and Gemini. You can easily plug in any of these models and start using them without the need for extensive modifications or setup. - - -```python -import os - -from dotenv import load_dotenv - -from swarms import Anthropic, Gemini, Mixtral, ModelParallelizer, OpenAIChat - -load_dotenv() - -# API Keys -anthropic_api_key = os.getenv("ANTHROPIC_API_KEY") -openai_api_key = os.getenv("OPENAI_API_KEY") -gemini_api_key = os.getenv("GEMINI_API_KEY") - -# Initialize the models -llm = OpenAIChat(openai_api_key=openai_api_key) -anthropic = Anthropic(anthropic_api_key=anthropic_api_key) -mixtral = Mixtral() -gemini = Gemini(gemini_api_key=gemini_api_key) - -# Initialize the parallelizer -llms = [llm, anthropic, mixtral, gemini] -parallelizer = ModelParallelizer(llms) - -# Set the task -task = "Generate a 10,000 word blog on health and wellness." - -# Run the task -out = parallelizer.run(task) - -# Print the responses 1 by 1 -for i in range(len(out)): - print(f"Response from LLM {i}: {out[i]}") -``` - - ### Simple Conversational Agent A Plug in and play conversational agent with `GPT4`, `Mixytral`, or any of our models @@ -463,11 +180,11 @@ agent("Generate a transcript for a youtube video on what swarms are!") ``` ## Devin -Implementation of Devil in less than 90 lines of code with several tools: +Implementation of Devin in less than 90 lines of code with several tools: terminal, browser, and edit files! ```python -from swarms import Agent, Anthropic, tool +from swarms import Agent, Anthropic import subprocess # Model @@ -476,7 +193,6 @@ llm = Anthropic( ) # Tools -@tool def terminal( code: str, ): @@ -494,8 +210,6 @@ def terminal( ).stdout return str(out) - -@tool def browser(query: str): """ Search the query in the browser with the `browser` tool. @@ -512,7 +226,6 @@ def browser(query: str): webbrowser.open(url) return f"Searching for {query} in the browser." -@tool def create_file(file_path: str, content: str): """ Create a file using the file editor tool. @@ -528,7 +241,6 @@ def create_file(file_path: str, content: str): file.write(content) return f"File {file_path} created successfully." -@tool def file_editor(file_path: str, mode: str, content: str): """ Edit a file using the file editor tool. @@ -558,85 +270,267 @@ agent = Agent( max_loops="auto", autosave=True, dashboard=False, - streaming_on=True, - verbose=True, - stopping_token="", - interactive=True, - tools=[terminal, browser, file_editor, create_file], - code_interpreter=True, - # streaming=True, + streaming_on=True, + verbose=True, + stopping_token="", + interactive=True, + tools=[terminal, browser, file_editor, create_file], + code_interpreter=True, + # streaming=True, +) + +# Run the agent +out = agent("Create a new file for a plan to take over the world.") +print(out) +``` + + +## `Agent`with Pydantic BaseModel as Output Type +The following is an example of an agent that intakes a pydantic basemodel and outputs it at the same time: + +```python +from pydantic import BaseModel, Field +from swarms import Anthropic, Agent + + +# Initialize the schema for the person's information +class Schema(BaseModel): + name: str = Field(..., title="Name of the person") + agent: int = Field(..., title="Age of the person") + is_student: bool = Field(..., title="Whether the person is a student") + courses: list[str] = Field( + ..., title="List of courses the person is taking" + ) + + +# Convert the schema to a JSON string +tool_schema = Schema( + name="Tool Name", + agent=1, + is_student=True, + courses=["Course1", "Course2"], +) + +# Define the task to generate a person's information +task = "Generate a person's information based on the following schema:" + +# Initialize the agent +agent = Agent( + agent_name="Person Information Generator", + system_prompt=( + "Generate a person's information based on the following schema:" + ), + # Set the tool schema to the JSON string -- this is the key difference + tool_schema=tool_schema, + llm=Anthropic(), + max_loops=3, + autosave=True, + dashboard=False, + streaming_on=True, + verbose=True, + interactive=True, + # Set the output type to the tool schema which is a BaseModel + output_type=tool_schema, # or dict, or str + metadata_output_type="json", + # List of schemas that the agent can handle + list_tool_schemas=[tool_schema], + function_calling_format_type="OpenAI", + function_calling_type="json", # or soon yaml +) + +# Run the agent to generate the person's information +generated_data = agent.run(task) + +# Print the generated data +print(f"Generated data: {generated_data}") + + +``` + + +### `ToolAgent` +ToolAgent is an agent that can use tools through JSON function calling. It intakes any open source model from huggingface and is extremely modular and plug in and play. We need help adding general support to all models soon. + + +```python +from pydantic import BaseModel, Field +from transformers import AutoModelForCausalLM, AutoTokenizer + +from swarms import ToolAgent +from swarms.utils.json_utils import base_model_to_json + +# Load the pre-trained model and tokenizer +model = AutoModelForCausalLM.from_pretrained( + "databricks/dolly-v2-12b", + load_in_4bit=True, + device_map="auto", +) +tokenizer = AutoTokenizer.from_pretrained("databricks/dolly-v2-12b") + + +# Initialize the schema for the person's information +class Schema(BaseModel): + name: str = Field(..., title="Name of the person") + agent: int = Field(..., title="Age of the person") + is_student: bool = Field( + ..., title="Whether the person is a student" + ) + courses: list[str] = Field( + ..., title="List of courses the person is taking" + ) + + +# Convert the schema to a JSON string +tool_schema = base_model_to_json(Schema) + +# Define the task to generate a person's information +task = ( + "Generate a person's information based on the following schema:" +) + +# Create an instance of the ToolAgent class +agent = ToolAgent( + name="dolly-function-agent", + description="Ana gent to create a child data", + model=model, + tokenizer=tokenizer, + json_schema=tool_schema, +) + +# Run the agent to generate the person's information +generated_data = agent.run(task) + +# Print the generated data +print(f"Generated data: {generated_data}") + +``` + + + + + + + +---- + +### `SequentialWorkflow` +Sequential Workflow enables you to sequentially execute tasks with `Agent` and then pass the output into the next agent and onwards until you have specified your max loops. `SequentialWorkflow` is wonderful for real-world business tasks like sending emails, summarizing documents, and analyzing data. + + +✅ Save and Restore Workflow states! + +✅ Multi-Modal Support for Visual Chaining + +✅ Utilizes Agent class + +```python +from swarms import Agent, SequentialWorkflow, Anthropic + + +# Initialize the language model agent (e.g., GPT-3) +llm = Anthropic() + +# Initialize agents for individual tasks +agent1 = Agent( + agent_name="Blog generator", + system_prompt="Generate a blog post like stephen king", + llm=llm, + max_loops=1, + dashboard=False, + tools=[], +) +agent2 = Agent( + agent_name="summarizer", + system_prompt="Sumamrize the blog post", + llm=llm, + max_loops=1, + dashboard=False, + tools=[], +) + +# Create the Sequential workflow +workflow = SequentialWorkflow( + agents=[agent1, agent2], max_loops=1, verbose=False +) + +# Run the workflow +workflow.run( + "Generate a blog post on how swarms of agents can help businesses grow." ) -# Run the agent -out = agent("Create a new file for a plan to take over the world.") -print(out) ``` -## `Agent`with Pydantic BaseModel as Output Type -The following is an example of an agent that intakes a pydantic basemodel and outputs it at the same time: + +### `ConcurrentWorkflow` +`ConcurrentWorkflow` runs all the tasks all at the same time with the inputs you give it! + ```python -from pydantic import BaseModel, Field -from swarms import Anthropic -from swarms import Agent +import os +from dotenv import load_dotenv -# Initialize the schema for the person's information -class Schema(BaseModel): - name: str = Field(..., title="Name of the person") - agent: int = Field(..., title="Age of the person") - is_student: bool = Field(..., title="Whether the person is a student") - courses: list[str] = Field( - ..., title="List of courses the person is taking" - ) +from swarms import Agent, ConcurrentWorkflow, OpenAIChat, Task +# Load environment variables from .env file +load_dotenv() -# Convert the schema to a JSON string -tool_schema = Schema( - name="Tool Name", - agent=1, - is_student=True, - courses=["Course1", "Course2"], -) +# Load environment variables +llm = OpenAIChat(openai_api_key=os.getenv("OPENAI_API_KEY")) +agent = Agent(llm=llm, max_loops=1) -# Define the task to generate a person's information -task = "Generate a person's information based on the following schema:" +# Create a workflow +workflow = ConcurrentWorkflow(max_workers=5) -# Initialize the agent -agent = Agent( - agent_name="Person Information Generator", - system_prompt=( - "Generate a person's information based on the following schema:" - ), - # Set the tool schema to the JSON string -- this is the key difference - tool_schema=tool_schema, - llm=Anthropic(), - max_loops=3, - autosave=True, - dashboard=False, - streaming_on=True, - verbose=True, - interactive=True, - # Set the output type to the tool schema which is a BaseModel - output_type=tool_schema, # or dict, or str - metadata_output_type="json", - # List of schemas that the agent can handle - list_tool_schemas=[tool_schema], - function_calling_format_type="OpenAI", - function_calling_type="json", # or soon yaml -) +# Create tasks +task1 = Task(agent, "What's the weather in miami") +task2 = Task(agent, "What's the weather in new york") +task3 = Task(agent, "What's the weather in london") -# Run the agent to generate the person's information -generated_data = agent.run(task) +# Add tasks to the workflow +workflow.add(tasks=[task1, task2, task3]) -# Print the generated data -print(f"Generated data: {generated_data}") +# Run the workflow +workflow.run() +``` + +### `RecursiveWorkflow` +`RecursiveWorkflow` will keep executing the tasks until a specific token like is located inside the text! + +```python +import os + +from dotenv import load_dotenv + +from swarms import Agent, OpenAIChat, RecursiveWorkflow, Task + +# Load environment variables from .env file +load_dotenv() + +# Load environment variables +llm = OpenAIChat(openai_api_key=os.getenv("OPENAI_API_KEY")) +agent = Agent(llm=llm, max_loops=1) + +# Create a workflow +workflow = RecursiveWorkflow(stop_token="") + +# Create tasks +task1 = Task(agent, "What's the weather in miami") +task2 = Task(agent, "What's the weather in new york") +task3 = Task(agent, "What's the weather in london") +# Add tasks to the workflow +workflow.add(task1) +workflow.add(task2) +workflow.add(task3) +# Run the workflow +workflow.run() ``` + ### `SwarmNetwork` `SwarmNetwork` provides the infrasturcture for building extremely dense and complex multi-agent applications that span across various types of agents. @@ -763,106 +657,6 @@ print(f"Task result: {task.result}") -### `BlockList` -- Modularity and Flexibility: BlocksList allows users to create custom swarms by adding or removing different classes or functions as blocks. This means users can easily tailor the functionality of their swarm to suit their specific needs. - -- Ease of Management: With methods to add, remove, update, and retrieve blocks, BlocksList provides a straightforward way to manage the components of a swarm. This makes it easier to maintain and update the swarm over time. - -- Enhanced Searchability: BlocksList offers methods to get blocks by various attributes such as name, type, ID, and parent-related properties. This makes it easier for users to find and work with specific blocks in a large and complex swarm. - -```python -import os - -from dotenv import load_dotenv -from transformers import AutoModelForCausalLM, AutoTokenizer -from pydantic import BaseModel -from swarms import BlocksList, Gemini, GPT4VisionAPI, Mixtral, OpenAI, ToolAgent - -# Load the environment variables -load_dotenv() - -# Get the environment variables -openai_api_key = os.getenv("OPENAI_API_KEY") -gemini_api_key = os.getenv("GEMINI_API_KEY") - -# Tool Agent -model = AutoModelForCausalLM.from_pretrained("databricks/dolly-v2-12b") -tokenizer = AutoTokenizer.from_pretrained("databricks/dolly-v2-12b") - -# Initialize the schema for the person's information -class Schema(BaseModel): - name: str = Field(..., title="Name of the person") - agent: int = Field(..., title="Age of the person") - is_student: bool = Field( - ..., title="Whether the person is a student" - ) - courses: list[str] = Field( - ..., title="List of courses the person is taking" - ) - -# Convert the schema to a JSON string -json_schema = base_model_to_json(Schema) - - -toolagent = ToolAgent(model=model, tokenizer=tokenizer, json_schema=json_schema) - -# Blocks List which enables you to build custom swarms by adding classes or functions -swarm = BlocksList( - "SocialMediaSwarm", - "A swarm of social media agents", - [ - OpenAI(openai_api_key=openai_api_key), - Mixtral(), - GPT4VisionAPI(openai_api_key=openai_api_key), - Gemini(gemini_api_key=gemini_api_key), - ], -) - - -# Add the new block to the swarm -swarm.add(toolagent) - -# Remove a block from the swarm -swarm.remove(toolagent) - -# Update a block in the swarm -swarm.update(toolagent) - -# Get a block at a specific index -block_at_index = swarm.get(0) - -# Get all blocks in the swarm -all_blocks = swarm.get_all() - -# Get blocks by name -openai_blocks = swarm.get_by_name("OpenAI") - -# Get blocks by type -gpt4_blocks = swarm.get_by_type("GPT4VisionAPI") - -# Get blocks by ID -block_by_id = swarm.get_by_id(toolagent.id) - -# Get blocks by parent -blocks_by_parent = swarm.get_by_parent(swarm) - -# Get blocks by parent ID -blocks_by_parent_id = swarm.get_by_parent_id(swarm.id) - -# Get blocks by parent name -blocks_by_parent_name = swarm.get_by_parent_name(swarm.name) - -# Get blocks by parent type -blocks_by_parent_type = swarm.get_by_parent_type(type(swarm).__name__) - -# Get blocks by parent description -blocks_by_parent_description = swarm.get_by_parent_description(swarm.description) - -# Run the block in the swarm -inference = swarm.run_block(toolagent, "Hello World") -print(inference) -``` - ## Majority Voting Multiple-agents will evaluate an idea based off of an parsing or evaluation function. From papers like "[More agents is all you need](https://arxiv.org/pdf/2402.05120.pdf) @@ -1177,88 +971,95 @@ autoswarm.run("Analyze these financial data and give me a summary") ``` ## `AgentRearrange` -Inspired by Einops and einsum, this orchestration techniques enables you to map out the relationships between various agents. For example you specify linear and sequential relationships like `a -> a1 -> a2 -> a3` or concurrent relationships where the first agent will send a message to 3 agents all at once: `a -> a1, a2, a3`. You can customize your workflow to mix sequential and concurrent relationships +Inspired by Einops and einsum, this orchestration techniques enables you to map out the relationships between various agents. For example you specify linear and sequential relationships like `a -> a1 -> a2 -> a3` or concurrent relationships where the first agent will send a message to 3 agents all at once: `a -> a1, a2, a3`. You can customize your workflow to mix sequential and concurrent relationships. [Docs Available:](https://swarms.apac.ai/en/latest/swarms/structs/agent_rearrange/) ```python -from swarms import Agent, Anthropic, AgentRearrange, +from swarms import Agent, AgentRearrange, rearrange, Anthropic -## Initialize the workflow -agent = Agent( - agent_name="t", - agent_description=( - "Generate a transcript for a youtube video on what swarms" - " are!" - ), - system_prompt=( - "Generate a transcript for a youtube video on what swarms" - " are!" - ), + +# Initialize the director agent + +director = Agent( + agent_name="Director", + system_prompt="Directs the tasks for the workers", llm=Anthropic(), max_loops=1, - autosave=True, dashboard=False, streaming_on=True, verbose=True, stopping_token="", + state_save_file_type="json", + saved_state_path="director.json", ) -agent2 = Agent( - agent_name="t1", - agent_description=( - "Generate a transcript for a youtube video on what swarms" - " are!" - ), + +# Initialize worker 1 + +worker1 = Agent( + agent_name="Worker1", + system_prompt="Generates a transcript for a youtube video on what swarms are", llm=Anthropic(), max_loops=1, - system_prompt="Summarize the transcript", - autosave=True, dashboard=False, streaming_on=True, verbose=True, stopping_token="", + state_save_file_type="json", + saved_state_path="worker1.json", ) -agent3 = Agent( - agent_name="t2", - agent_description=( - "Generate a transcript for a youtube video on what swarms" - " are!" - ), + +# Initialize worker 2 +worker2 = Agent( + agent_name="Worker2", + system_prompt="Summarizes the transcript generated by Worker1", llm=Anthropic(), max_loops=1, - system_prompt="Finalize the transcript", - autosave=True, dashboard=False, streaming_on=True, verbose=True, stopping_token="", + state_save_file_type="json", + saved_state_path="worker2.json", ) -# Rearrange the agents -rearrange = AgentRearrange( - agents=[agent, agent2, agent3], - verbose=True, - # custom_prompt="Summarize the transcript", +# Create a list of agents +agents = [director, worker1, worker2] + +# Define the flow pattern +flow = "Director -> Worker1 -> Worker2" + +# Using AgentRearrange class +agent_system = AgentRearrange(agents=agents, flow=flow) +output = agent_system.run( + "Create a format to express and communicate swarms of llms in a structured manner for youtube" ) +print(output) -# Run the workflow on a task -results = rearrange( - # pattern="t -> t1, t2 -> t2", - pattern="t -> t1 -> t2", - default_task=( - "Generate a transcript for a YouTube video on what swarms" - " are!" - ), - t="Generate a transcript for a YouTube video on what swarms are!", - # t2="Summarize the transcript", - # t3="Finalize the transcript", + +# Using rearrange function +output = rearrange( + agents, + flow, + "Create a format to express and communicate swarms of llms in a structured manner for youtube", ) -# print(results) +print(output) ``` +## `HierarhicalSwarm` +Coming soon... + + +## `AgentLoadBalancer` +Coming soon... + + +## `GraphSwarm` +Coming soon... + --- @@ -1267,6 +1068,26 @@ Documentation is located here at: [swarms.apac.ai](https://swarms.apac.ai) ---- +## Folder Structure +The swarms package has been meticlously crafted for extreme use-ability and understanding, the swarms package is split up into various modules such as `swarms.agents` that holds pre-built agents, `swarms.structs` that holds a vast array of structures like `Agent` and multi agent structures. The 3 most important are `structs`, `models`, and `agents`. + +```sh +├── __init__.py +├── agents +├── artifacts +├── memory +├── schemas +├── models +├── prompts +├── structs +├── telemetry +├── tools +├── utils +└── workers +``` + +---- + ## 🫶 Contributions: The easiest way to contribute is to pick any issue with the `good first issue` tag 💪. Read the Contributing guidelines [here](/CONTRIBUTING.md). Bug Report? [File here](https://github.com/swarms/gateway/issues) | Feature Request? [File here](https://github.com/swarms/gateway/issues) @@ -1297,68 +1118,33 @@ Join our growing community around the world, for real-time support, ideas, and d Book a discovery call to learn how Swarms can lower your operating costs by 40% with swarms of autonomous agents in lightspeed. [Click here to book a time that works for you!](https://calendly.com/swarm-corp/30min?month=2023-11) - ## Accelerate Backlog -Help us accelerate our backlog by supporting us financially! Note, we're an open source corporation and so all the revenue we generate is through donations at the moment ;) +Accelerate Bugs, Features, and Demos to implement by supporting us here: -## File Structure -The swarms package has been meticlously crafted for extreme use-ability and understanding, the swarms package is split up into various modules such as `swarms.agents` that holds pre-built agents, `swarms.structs` that holds a vast array of structures like `Agent` and multi agent structures. The 3 most important are `structs`, `models`, and `agents`. - -```sh -├── __init__.py -├── agents -├── artifacts -├── chunkers -├── cli -├── loaders -├── memory -├── models -├── prompts -├── structs -├── telemetry -├── tokenizers -├── tools -├── utils -└── workers -``` - ## Docker Instructions - -This application uses Docker with CUDA support. To build and run the Docker container, follow these steps: - -### Prerequisites - -- Make sure you have [Docker installed](https://docs.docker.com/get-docker/) on your machine. -- Ensure your machine has an NVIDIA GPU and [NVIDIA Docker support](https://github.com/NVIDIA/nvidia-docker) installed. - -### Building the Docker Image - -To build the Docker image, navigate to the root directory containing the `Dockerfile` and run the following command: - -```bash -docker build --gpus all -t swarms -``` -### Running the Docker Container -To run the Docker container, use the following command: - -`docker run --gpus all -p 4000:80 swarms` - -Replace swarms with the name of your Docker image, and replace 4000:80 with your actual port mapping. The format is hostPort:containerPort. - -Now, your application should be running with CUDA support! +- [Learn More Here About Deployments In Docker](https://swarms.apac.ai/en/latest/docker_setup/) ## Swarm Newsletter 🤖 🤖 🤖 📧 Sign up to the Swarm newsletter to receive updates on the latest Autonomous agent research papers, step by step guides on creating multi-agent app, and much more Swarmie goodiness 😊 - [CLICK HERE TO SIGNUP](https://docs.google.com/forms/d/e/1FAIpQLSfqxI2ktPR9jkcIwzvHL0VY6tEIuVPd-P2fOWKnd6skT9j1EQ/viewform?usp=sf_link) # License Apache License - - +# Citation +Please cite Swarms in your paper or your project if you found it beneficial in any way! Appreciate you. + +```bibtex +@misc{swarms, + author = {Gomez, Kye}, + title = {{Swarms: The Multi-Agent Collaboration Framework}}, + howpublished = {\url{https://github.com/kyegomez/swarms}}, + year = {2023}, + note = {Accessed: Date} +} +``` diff --git a/docs/docker_setup.md b/docs/swarms/install/docker_setup.md similarity index 100% rename from docs/docker_setup.md rename to docs/swarms/install/docker_setup.md diff --git a/docs/install.md b/docs/swarms/install/install.md similarity index 100% rename from docs/install.md rename to docs/swarms/install/install.md diff --git a/docs/swarms/memory/diy_memory.md b/docs/swarms/memory/diy_memory.md index 2bcb056e..ffb98cae 100644 --- a/docs/swarms/memory/diy_memory.md +++ b/docs/swarms/memory/diy_memory.md @@ -261,49 +261,181 @@ In this example, we define a `PineconeVectorDatabase` class that inherits from ` Chroma is an open-source vector database library that provides efficient storage, retrieval, and manipulation of vector data using various backends, including DuckDB, Chromadb, and more. ```python +import logging +import os +import uuid +from typing import Optional -from chromadb.client import Client -from swarms import BaseVectorDatabase - -class ChromaVectorDatabase(MyCustomVectorDatabase): - -    def __init__(self, *args, **kwargs): - -        super().__init__(*args, **kwargs) - -        # Chroma initialization - -        self.client = Client() - -        self.collection = self.client.get_or_create_collection("vector_collection") - -    def connect(self): - -        # Chroma connection logic - -        pass - -    def close(self): - -        # Close Chroma connection - -        pass - -    def query(self, query: str): - -        # Execute Chroma query - -        results = self.collection.query(query) +import chromadb +from dotenv import load_dotenv -        return results +from swarms.utils.data_to_text import data_to_text +from swarms.utils.markdown_message import display_markdown_message +from swarms.memory.base_vectordb import BaseVectorDatabase -    def add(self, doc: str): - -        # Add document to Chroma collection +# Load environment variables +load_dotenv() -        self.collection.add(doc) -    # Implement other abstract methods +# Results storage using local ChromaDB +class ChromaDB(BaseVectorDatabase): + """ + + ChromaDB database + + Args: + metric (str): The similarity metric to use. + output (str): The name of the collection to store the results in. + limit_tokens (int, optional): The maximum number of tokens to use for the query. Defaults to 1000. + n_results (int, optional): The number of results to retrieve. Defaults to 2. + + Methods: + add: _description_ + query: _description_ + + Examples: + >>> chromadb = ChromaDB( + >>> metric="cosine", + >>> output="results", + >>> llm="gpt3", + >>> openai_api_key=OPENAI_API_KEY, + >>> ) + >>> chromadb.add(task, result, result_id) + """ + + def __init__( + self, + metric: str = "cosine", + output_dir: str = "swarms", + limit_tokens: Optional[int] = 1000, + n_results: int = 3, + docs_folder: str = None, + verbose: bool = False, + *args, + **kwargs, + ): + self.metric = metric + self.output_dir = output_dir + self.limit_tokens = limit_tokens + self.n_results = n_results + self.docs_folder = docs_folder + self.verbose = verbose + + # Disable ChromaDB logging + if verbose: + logging.getLogger("chromadb").setLevel(logging.INFO) + + # Create Chroma collection + chroma_persist_dir = "chroma" + chroma_client = chromadb.PersistentClient( + settings=chromadb.config.Settings( + persist_directory=chroma_persist_dir, + ), + *args, + **kwargs, + ) + + # Create ChromaDB client + self.client = chromadb.Client() + + # Create Chroma collection + self.collection = chroma_client.get_or_create_collection( + name=output_dir, + metadata={"hnsw:space": metric}, + *args, + **kwargs, + ) + display_markdown_message( + "ChromaDB collection created:" + f" {self.collection.name} with metric: {self.metric} and" + f" output directory: {self.output_dir}" + ) + + # If docs + if docs_folder: + display_markdown_message( + f"Traversing directory: {docs_folder}" + ) + self.traverse_directory() + + def add( + self, + document: str, + *args, + **kwargs, + ): + """ + Add a document to the ChromaDB collection. + + Args: + document (str): The document to be added. + condition (bool, optional): The condition to check before adding the document. Defaults to True. + + Returns: + str: The ID of the added document. + """ + try: + doc_id = str(uuid.uuid4()) + self.collection.add( + ids=[doc_id], + documents=[document], + *args, + **kwargs, + ) + print("-----------------") + print("Document added successfully") + print("-----------------") + return doc_id + except Exception as e: + raise Exception(f"Failed to add document: {str(e)}") + + def query( + self, + query_text: str, + *args, + **kwargs, + ): + """ + Query documents from the ChromaDB collection. + + Args: + query (str): The query string. + n_docs (int, optional): The number of documents to retrieve. Defaults to 1. + + Returns: + dict: The retrieved documents. + """ + try: + docs = self.collection.query( + query_texts=[query_text], + n_results=self.n_results, + *args, + **kwargs, + )["documents"] + return docs[0] + except Exception as e: + raise Exception(f"Failed to query documents: {str(e)}") + + def traverse_directory(self): + """ + Traverse through every file in the given directory and its subdirectories, + and return the paths of all files. + Parameters: + - directory_name (str): The name of the directory to traverse. + Returns: + - list: A list of paths to each file in the directory and its subdirectories. + """ + added_to_db = False + + for root, dirs, files in os.walk(self.docs_folder): + for file in files: + file = os.path.join(self.docs_folder, file) + _, ext = os.path.splitext(file) + data = data_to_text(file) + added_to_db = self.add(str(data)) + print(f"{file} added to Database") + + return added_to_db ``` diff --git a/docs/swarms/models/gpt4o.md b/docs/swarms/models/gpt4o.md new file mode 100644 index 00000000..7b53a742 --- /dev/null +++ b/docs/swarms/models/gpt4o.md @@ -0,0 +1,150 @@ +# Documentation for GPT4o Module + +## Overview and Introduction + +The `GPT4o` module is a multi-modal conversational model based on OpenAI's GPT-4 architecture. It extends the functionality of the `BaseMultiModalModel` class, enabling it to handle both text and image inputs for generating diverse and contextually rich responses. This module leverages the power of the GPT-4 model to enhance interactions by integrating visual information with textual prompts, making it highly relevant for applications requiring multi-modal understanding and response generation. + +### Key Concepts +- **Multi-Modal Model**: A model that can process and generate responses based on multiple types of inputs, such as text and images. +- **System Prompt**: A predefined prompt to guide the conversation flow. +- **Temperature**: A parameter that controls the randomness of the response generation. +- **Max Tokens**: The maximum number of tokens (words or word pieces) in the generated response. + +## Class Definition + +### `GPT4o` Class + + +### Parameters + +| Parameter | Type | Description | +|-----------------|--------|--------------------------------------------------------------------------------------| +| `system_prompt` | `str` | The system prompt to be used in the conversation. | +| `temperature` | `float`| The temperature parameter for generating diverse responses. Default is `0.1`. | +| `max_tokens` | `int` | The maximum number of tokens in the generated response. Default is `300`. | +| `openai_api_key`| `str` | The API key for accessing the OpenAI GPT-4 API. | +| `*args` | | Additional positional arguments. | +| `**kwargs` | | Additional keyword arguments. | + +## Functionality and Usage + +### `encode_image` Function + +The `encode_image` function is used to encode an image file into a base64 string format, which can then be included in the request to the GPT-4 API. + +#### Parameters + +| Parameter | Type | Description | +|---------------|--------|----------------------------------------------| +| `image_path` | `str` | The local path to the image file to be encoded. | + +#### Returns + +| Return Type | Description | +|-------------|---------------------------------| +| `str` | The base64 encoded string of the image. | + +### `GPT4o.__init__` Method + +The constructor for the `GPT4o` class initializes the model with the specified parameters and sets up the OpenAI client. + +### `GPT4o.run` Method + +The `run` method executes the GPT-4o model to generate a response based on the provided task and optional image. + +#### Parameters + +| Parameter | Type | Description | +|---------------|--------|----------------------------------------------------| +| `task` | `str` | The task or user prompt for the conversation. | +| `local_img` | `str` | The local path to the image file. | +| `img` | `str` | The URL of the image. | +| `*args` | | Additional positional arguments. | +| `**kwargs` | | Additional keyword arguments. | + +#### Returns + +| Return Type | Description | +|-------------|--------------------------------------------------| +| `str` | The generated response from the GPT-4o model. | + +## Usage Examples + +### Example 1: Basic Text Prompt + +```python +from swarms import GPT4o + +# Initialize the model +model = GPT4o( + system_prompt="You are a helpful assistant.", + temperature=0.7, + max_tokens=150, + openai_api_key="your_openai_api_key" +) + +# Define the task +task = "What is the capital of France?" + +# Generate response +response = model.run(task) +print(response) +``` + +### Example 2: Text Prompt with Local Image + +```python +from swarms import GPT4o + +# Initialize the model +model = GPT4o( + system_prompt="Describe the image content.", + temperature=0.5, + max_tokens=200, + openai_api_key="your_openai_api_key" +) + +# Define the task and image path +task = "Describe the content of this image." +local_img = "path/to/your/image.jpg" + +# Generate response +response = model.run(task, local_img=local_img) +print(response) +``` + +### Example 3: Text Prompt with Image URL + +```python +from swarms import GPT4o + +# Initialize the model +model = GPT4o( + system_prompt="You are a visual assistant.", + temperature=0.6, + max_tokens=250, + openai_api_key="your_openai_api_key" +) + +# Define the task and image URL +task = "What can you tell about the scenery in this image?" +img_url = "http://example.com/image.jpg" + +# Generate response +response = model.run(task, img=img_url) +print(response) +``` + +## Additional Information and Tips + +- **API Key Management**: Ensure that your OpenAI API key is securely stored and managed. Do not hard-code it in your scripts. Use environment variables or secure storage solutions. +- **Image Encoding**: The `encode_image` function is crucial for converting images to a base64 format suitable for API requests. Ensure that the images are accessible and properly formatted. +- **Temperature Parameter**: Adjust the `temperature` parameter to control the creativity of the model's responses. Lower values make the output more deterministic, while higher values increase randomness. +- **Token Limit**: Be mindful of the `max_tokens` parameter to avoid exceeding the API's token limits. This parameter controls the length of the generated responses. + +## References and Resources + +- [OpenAI API Documentation](https://beta.openai.com/docs/) +- [Python Base64 Encoding](https://docs.python.org/3/library/base64.html) +- [dotenv Documentation](https://saurabh-kumar.com/python-dotenv/) +- [BaseMultiModalModel Documentation](https://swarms.apac.ai) \ No newline at end of file diff --git a/docs/diy_your_own_agent.md b/docs/swarms/structs/diy_your_own_agent.md similarity index 100% rename from docs/diy_your_own_agent.md rename to docs/swarms/structs/diy_your_own_agent.md diff --git a/docs/swarms/workers/abstract_worker.md b/docs/swarms/workers/abstract_worker.md deleted file mode 100644 index e7d3c8a8..00000000 --- a/docs/swarms/workers/abstract_worker.md +++ /dev/null @@ -1,258 +0,0 @@ -# AbstractWorker Class -==================== - -The `AbstractWorker` class is an abstract class for AI workers. An AI worker can communicate with other workers and perform actions. Different workers can differ in what actions they perform in the `receive` method. - -## Class Definition ----------------- - -``` -class AbstractWorker: - """(In preview) An abstract class for AI worker. - - An worker can communicate with other workers and perform actions. - Different workers can differ in what actions they perform in the `receive` method. - """ -``` - - -## Initialization --------------- - -The `AbstractWorker` class is initialized with a single parameter: - -- `name` (str): The name of the worker. - -``` -def __init__( - self, - name: str, -): - """ - Args: - name (str): name of the worker. - """ - self._name = name -``` - - -## Properties ----------- - -The `AbstractWorker` class has a single property: - -- `name`: Returns the name of the worker. - -``` -@property -def name(self): - """Get the name of the worker.""" - return self._name -``` - - -## Methods -------- - -The `AbstractWorker` class has several methods: - -### `run` - -The `run` method is used to run the worker agent once. It takes a single parameter: - -- `task` (str): The task to be run. - -``` -def run( - self, - task: str -): - """Run the worker agent once""" -``` - - -### `send` - -The `send` method is used to send a message to another worker. It takes three parameters: - -- `message` (Union[Dict, str]): The message to be sent. -- `recipient` (AbstractWorker): The recipient of the message. -- `request_reply` (Optional[bool]): If set to `True`, the method will request a reply from the recipient. - -``` -def send( - self, - message: Union[Dict, str], - recipient: AbstractWorker, - request_reply: Optional[bool] = None -): - """(Abstract method) Send a message to another worker.""" -``` - - -### `a_send` - -The `a_send` method is the asynchronous version of the `send` method. It takes the same parameters as the `send` method. - -``` -async def a_send( - self, - message: Union[Dict, str], - recipient: AbstractWorker, - request_reply: Optional[bool] = None -): - """(Abstract async method) Send a message to another worker.""" -``` - - -### `receive` - -The `receive` method is used to receive a message from another worker. It takes three parameters: - -- `message` (Union[Dict, str]): The message to be received. -- `sender` (AbstractWorker): The sender of the message. -- `request_reply` (Optional[bool]): If set to `True`, the method will request a reply from the sender. - -``` -def receive( - self, - message: Union[Dict, str], - sender: AbstractWorker, - request_reply: Optional[bool] = None -): - """(Abstract method) Receive a message from another worker.""" -``` - - -### `a_receive` - -The `a_receive` method is the asynchronous version of the `receive` method. It takes the same parameters as the `receive` method. - -``` -async def a_receive( - self, - message: Union[Dict, str], - sender: AbstractWorker, - request_reply: Optional[bool] = None -): - """(Abstract async method) Receive a message from another worker.""" -``` - - -### `reset` - -The `reset` method is used to reset the worker. - -``` -def reset(self): - """(Abstract method) Reset the worker.""" -``` - - -### `generate_reply` - -The `generate_reply` method is used to generate a reply based on the received messages. It takes two parameters: - -- `messages` (Optional[List[Dict]]): A list of messages received. -- `sender` (AbstractWorker): The sender of the messages. - -The method returns a string, a dictionary, or `None`. If `None` is returned, no reply is generated. - -``` -def generate_reply( - self, - messages: Optional[List[Dict]] = None, - sender: AbstractWorker, - **kwargs, -) -> Union[str, Dict, None]: - """(Abstract method) Generate a reply based on the received messages. - - Args: - messages (list[dict]): a list of messages received. - sender: sender of an Agent instance. - Returns: - str or dict or None: the generated reply. If None, no reply is generated. - """ -``` - - -### `a_generate_reply` - -The `a_generate_reply` method is the asynchronous version of the `generate_reply` method. It - -takes the same parameters as the `generate_reply` method. - -``` -async def a_generate_reply( - self, - messages: Optional[List[Dict]] = None, - sender: AbstractWorker, - **kwargs, -) -> Union[str, Dict, None]: - """(Abstract async method) Generate a reply based on the received messages. - - Args: - messages (list[dict]): a list of messages received. - sender: sender of an Agent instance. - Returns: - str or dict or None: the generated reply. If None, no reply is generated. - """ -``` - - -Usage Examples --------------- - -### Example 1: Creating an AbstractWorker - -``` -from swarms.worker.base import AbstractWorker - -worker = AbstractWorker(name="Worker1") -print(worker.name) # Output: Worker1 -``` - - -In this example, we create an instance of `AbstractWorker` named "Worker1" and print its name. - -### Example 2: Sending a Message - -``` -from swarms.worker.base import AbstractWorker - -worker1 = AbstractWorker(name="Worker1") -worker2 = AbstractWorker(name="Worker2") - -message = {"content": "Hello, Worker2!"} -worker1.send(message, worker2) -``` - - -In this example, "Worker1" sends a message to "Worker2". The message is a dictionary with a single key-value pair. - -### Example 3: Receiving a Message - -``` -from swarms.worker.base import AbstractWorker - -worker1 = AbstractWorker(name="Worker1") -worker2 = AbstractWorker(name="Worker2") - -message = {"content": "Hello, Worker2!"} -worker1.send(message, worker2) - -received_message = worker2.receive(message, worker1) -print(received_message) # Output: {"content": "Hello, Worker2!"} -``` - - -In this example, "Worker1" sends a message to "Worker2". "Worker2" then receives the message and prints it. - -Notes ------ - -- The `AbstractWorker` class is an abstract class, which means it cannot be instantiated directly. Instead, it should be subclassed, and at least the `send`, `receive`, `reset`, and `generate_reply` methods should be overridden. -- The `send` and `receive` methods are abstract methods, which means they must be implemented in any subclass of `AbstractWorker`. -- The `a_send`, `a_receive`, and `a_generate_reply` methods are asynchronous methods, which means they return a coroutine that can be awaited using the `await` keyword. -- The `generate_reply` method is used to generate a reply based on the received messages. The exact implementation of this method will depend on the specific requirements of your application. -- The `reset` method is used to reset the state of the worker. The exact implementation of this method will depend on the specific requirements of your application. \ No newline at end of file diff --git a/docs/swarms/workers/base.md b/docs/swarms/workers/base.md deleted file mode 100644 index 9da45ba3..00000000 --- a/docs/swarms/workers/base.md +++ /dev/null @@ -1,403 +0,0 @@ -# `AbstractWorker` Documentation - -## Table of Contents - -1. [Introduction](#introduction) -2. [Abstract Worker](#abstract-worker) - 1. [Class Definition](#class-definition) - 2. [Attributes](#attributes) - 3. [Methods](#methods) -3. [Tutorial: Creating Custom Workers](#tutorial-creating-custom-workers) -4. [Conclusion](#conclusion) - ---- - -## 1. Introduction - -Welcome to the documentation for the Swarms library, a powerful tool for building and simulating swarm architectures. This library provides a foundation for creating and managing autonomous workers that can communicate, collaborate, and perform various tasks in a coordinated manner. - -In this documentation, we will cover the `AbstractWorker` class, which serves as the fundamental building block for creating custom workers in your swarm simulations. We will explain the class's architecture, attributes, and methods in detail, providing practical examples to help you understand how to use it effectively. - -Whether you want to simulate a team of autonomous robots, a group of AI agents, or any other swarm-based system, the Swarms library is here to simplify the process and empower you to build complex simulations. - ---- - -## 2. Abstract Worker - -### 2.1 Class Definition - -The `AbstractWorker` class is an abstract base class that serves as the foundation for creating worker agents in your swarm simulations. It defines a set of methods that should be implemented by subclasses to customize the behavior of individual workers. - -Here is the class definition: - -```python -class AbstractWorker: - def __init__(self, name: str): - """ - Args: - name (str): Name of the worker. - """ - - @property - def name(self): - """Get the name of the worker.""" - - def run(self, task: str): - """Run the worker agent once.""" - - def send( - self, message: Union[Dict, str], recipient, request_reply: Optional[bool] = None - ): - """Send a message to another worker.""" - - async def a_send( - self, message: Union[Dict, str], recipient, request_reply: Optional[bool] = None - ): - """Send a message to another worker asynchronously.""" - - def receive( - self, message: Union[Dict, str], sender, request_reply: Optional[bool] = None - ): - """Receive a message from another worker.""" - - async def a_receive( - self, message: Union[Dict, str], sender, request_reply: Optional[bool] = None - ): - """Receive a message from another worker asynchronously.""" - - def reset(self): - """Reset the worker.""" - - def generate_reply( - self, messages: Optional[List[Dict]] = None, sender=None, **kwargs - ) -> Union[str, Dict, None]: - """Generate a reply based on received messages.""" - - async def a_generate_reply( - self, messages: Optional[List[Dict]] = None, sender=None, **kwargs - ) -> Union[str, Dict, None]: - """Generate a reply based on received messages asynchronously.""" -``` - -### 2.2 Attributes - -- `name (str)`: The name of the worker, which is set during initialization. - -### 2.3 Methods - -Now, let's delve into the methods provided by the `AbstractWorker` class and understand their purposes and usage. - -#### `__init__(self, name: str)` - -The constructor method initializes a worker with a given name. - -**Parameters:** -- `name (str)`: The name of the worker. - -**Usage Example:** - -```python -worker = AbstractWorker("Worker1") -``` - -#### `name` (Property) - -The `name` property allows you to access the name of the worker. - -**Usage Example:** - -```python -worker_name = worker.name -``` - -#### `run(self, task: str)` - -The `run()` method is a placeholder for running the worker. You can customize this method in your subclass to define the specific actions the worker should perform. - -**Parameters:** -- `task (str)`: A task description or identifier. - -**Usage Example (Customized Subclass):** - -```python -class MyWorker(AbstractWorker): - def run(self, task: str): - print(f"{self.name} is performing task: {task}") - - -worker = MyWorker("Worker1") -worker.run("Collect data") -``` - -#### `send(self, message: Union[Dict, str], recipient, request_reply: Optional[bool] = None)` - -The `send()` method allows the worker to send a message to another worker or recipient. The message can be either a dictionary or a string. - -**Parameters:** -- `message (Union[Dict, str])`: The message to be sent. -- `recipient`: The recipient worker or entity. -- `request_reply (Optional[bool])`: If `True`, the sender requests a reply from the recipient. If `False`, no reply is requested. Default is `None`. - -**Usage Example:** - -```python -worker1 = AbstractWorker("Worker1") -worker2 = AbstractWorker("Worker2") - -message = "Hello, Worker2!" -worker1.send(message, worker2) -``` - -#### `a_send(self, message: Union[Dict, str], recipient, request_reply: Optional[bool] = None)` - -The `a_send()` method is an asynchronous version of the `send()` method, allowing the worker to send messages asynchronously. - -**Parameters:** (Same as `send()`) - -**Usage Example:** - -```python -import asyncio - - -async def main(): - worker1 = AbstractWorker("Worker1") - worker2 = AbstractWorker("Worker2") - - message = "Hello, Worker2!" - await worker1.a_send(message, worker2) - - -loop = asyncio.get_event_loop() -loop.run_until_complete(main()) -``` - -#### `receive(self, message: Union[Dict, str], sender, request_reply: Optional[bool] = None)` - -The `receive()` method allows the worker to receive messages from other workers or senders. You can customize this method in your subclass to define how the worker handles incoming messages. - -**Parameters:** -- `message (Union[Dict, str])`: The received message. -- `sender`: The sender worker or entity. -- `request_reply (Optional[bool])`: Indicates whether a reply is requested. Default is `None`. - -**Usage Example (Customized Subclass):** - -```python -class MyWorker(AbstractWorker): - def receive(self, message: Union[Dict, str], sender, request_reply: Optional[bool] = None): - if isinstance(message, str): - print(f"{self.name} received a text message from {sender.name}: {message}") - elif isinstance(message, dict): - print(f"{self.name} received a dictionary message from {sender.name}: {message}") - -worker1 = MyWorker("Worker1") -worker2 = MyWorker("Worker2") - -message1 = - - "Hello, Worker2!" -message2 = {"data": 42} - -worker1.receive(message1, worker2) -worker1.receive(message2, worker2) -``` - -#### `a_receive(self, message: Union[Dict, str], sender, request_reply: Optional[bool] = None)` - -The `a_receive()` method is an asynchronous version of the `receive()` method, allowing the worker to receive messages asynchronously. - -**Parameters:** (Same as `receive()`) - -**Usage Example:** - -```python -import asyncio - - -async def main(): - worker1 = AbstractWorker("Worker1") - worker2 = AbstractWorker("Worker2") - - message1 = "Hello, Worker2!" - message2 = {"data": 42} - - await worker1.a_receive(message1, worker2) - await worker1.a_receive(message2, worker2) - - -loop = asyncio.get_event_loop() -loop.run_until_complete(main()) -``` - -#### `reset(self)` - -The `reset()` method is a placeholder for resetting the worker. You can customize this method in your subclass to define how the worker should reset its state. - -**Usage Example (Customized Subclass):** - -```python -class MyWorker(AbstractWorker): - def reset(self): - print(f"{self.name} has been reset.") - - -worker = MyWorker("Worker1") -worker.reset() -``` - -#### `generate_reply(self, messages: Optional[List[Dict]] = None, sender=None, **kwargs) -> Union[str, Dict, None]` - -The `generate_reply()` method is a placeholder for generating a reply based on received messages. You can customize this method in your subclass to define the logic for generating replies. - -**Parameters:** -- `messages (Optional[List[Dict]])`: A list of received messages. -- `sender`: The sender of the reply. -- `kwargs`: Additional keyword arguments. - -**Returns:** -- `Union[str, Dict, None]`: The generated reply. If `None`, no reply is generated. - -**Usage Example (Customized Subclass):** - -```python -class MyWorker(AbstractWorker): - def generate_reply( - self, messages: Optional[List[Dict]] = None, sender=None, **kwargs - ) -> Union[str, Dict, None]: - if messages: - # Generate a reply based on received messages - return f"Received {len(messages)} messages from {sender.name}." - else: - return None - - -worker1 = MyWorker("Worker1") -worker2 = MyWorker("Worker2") - -message = "Hello, Worker2!" -reply = worker2.generate_reply([message], worker1) - -if reply: - print(f"{worker2.name} generated a reply: {reply}") -``` - -#### `a_generate_reply(self, messages: Optional[List[Dict]] = None, sender=None, **kwargs) -> Union[str, Dict, None]` - -The `a_generate_reply()` method is an asynchronous version of the `generate_reply()` method, allowing the worker to generate replies asynchronously. - -**Parameters:** (Same as `generate_reply()`) - -**Returns:** -- `Union[str, Dict, None]`: The generated reply. If `None`, no reply is generated. - -**Usage Example:** - -```python -import asyncio - - -async def main(): - worker1 = AbstractWorker("Worker1") - worker2 = AbstractWorker("Worker2") - - message = "Hello, Worker2!" - reply = await worker2.a_generate_reply([message], worker1) - - if reply: - print(f"{worker2.name} generated a reply: {reply}") - - -loop = asyncio.get_event_loop() -loop.run_until_complete(main()) -``` - ---- - -## 3. Tutorial: Creating Custom Workers - -In this tutorial, we will walk you through the process of creating custom workers by subclassing the `AbstractWorker` class. You can tailor these workers to perform specific tasks and communicate with other workers in your swarm simulations. - -### Step 1: Create a Custom Worker Class - -Start by creating a custom worker class that inherits from `AbstractWorker`. Define the `run()` and `receive()` methods to specify the behavior of your worker. - -```python -class CustomWorker(AbstractWorker): - def run(self, task: str): - print(f"{self.name} is performing task: {task}") - - def receive( - self, message: Union[Dict, str], sender, request_reply: Optional[bool] = None - ): - if isinstance(message, str): - print(f"{self.name} received a text message from {sender.name}: {message}") - elif isinstance(message, dict): - print( - f"{self.name} received a dictionary message from {sender.name}: {message}" - ) -``` - -### Step 2: Create Custom Worker Instances - -Instantiate your custom worker instances and give them unique names. - -```python -worker1 = CustomWorker("Worker1") -worker2 = CustomWorker("Worker2") -``` - -### Step 3: Run Custom Workers - -Use the `run()` method to make your custom workers perform tasks. - -```python -worker1.run("Collect data") -worker2.run("Process data") -``` - -### Step 4: Communicate Between Workers - -Use the `send()` method to send messages between workers. You can customize the `receive()` method to define how your workers handle incoming messages. - -```python -worker1.send("Hello, Worker2!", worker2) -worker2.send({"data": 42}, worker1) - -# Output will show the messages received by the workers -``` - -### Step 5: Generate Replies - -Customize the `generate_reply()` method to allow your workers to generate replies based on received messages. - -```python -class CustomWorker(AbstractWorker): - def generate_reply( - self, messages: Optional[List[Dict]] = None, sender=None, **kwargs - ) -> Union[str, Dict, None]: - if messages: - # Generate a reply based on received messages - return f"Received {len(messages)} messages from {sender.name}." - else: - return None -``` - -Now, your custom workers can generate replies to incoming messages. - -```python -reply = worker2.generate_reply(["Hello, Worker2!"], worker1) - -if reply: - print(f"{worker2.name} generated a reply: {reply}") -``` - ---- - -## 4. Conclusion - -Congratulations! You've learned how to use the Swarms library to create and customize worker agents for swarm simulations. You can now build complex swarm architectures, simulate autonomous systems, and experiment with various communication and task allocation strategies. - -Feel free to explore the Swarms library further and adapt it to your specific use cases. If you have any questions or need assistance, refer to the extensive documentation and resources available. - -Happy swarming! \ No newline at end of file diff --git a/docs/swarms/workers/index.md b/docs/swarms/workers/index.md deleted file mode 100644 index 3662fd8a..00000000 --- a/docs/swarms/workers/index.md +++ /dev/null @@ -1,248 +0,0 @@ -# Module Name: Worker - -The `Worker` class encapsulates the idea of a semi-autonomous agent that utilizes a large language model to execute tasks. The module provides a unified interface for AI-driven task execution while combining a series of tools and utilities. It sets up memory storage and retrieval mechanisms for contextual recall and offers an option for human involvement, making it a versatile and adaptive agent for diverse applications. - -## **Class Definition**: - -```python -class Worker: -``` - -### **Parameters**: - -- `model_name` (str, default: "gpt-4"): Name of the language model. -- `openai_api_key` (str, Optional): API key for accessing OpenAI's models. -- `ai_name` (str, default: "Autobot Swarm Worker"): Name of the AI agent. -- `ai_role` (str, default: "Worker in a swarm"): Role description of the AI agent. -- `external_tools` (list, Optional): A list of external tool objects to be used. -- `human_in_the_loop` (bool, default: False): If set to `True`, it indicates that human intervention may be required. -- `temperature` (float, default: 0.5): Sampling temperature for the language model's output. Higher values make the output more random, and lower values make it more deterministic. - -### **Methods**: - -#### `__init__`: - -Initializes the Worker class. - -#### `setup_tools`: - -Sets up the tools available to the worker. Default tools include reading and writing files, processing CSV data, querying websites, and taking human input. Additional tools can be appended through the `external_tools` parameter. - -#### `setup_memory`: - -Initializes memory systems using embeddings and a vector store for the worker. - -#### `setup_agent`: - -Sets up the primary agent using the initialized tools, memory, and language model. - -#### `run`: - -Executes a given task using the agent. - -#### `__call__`: - -Makes the Worker class callable. When an instance of the class is called, it will execute the provided task using the agent. - -## **Usage Examples**: - -### **Example 1**: Basic usage with default parameters: - -```python -from swarms import Worker -from swarms.models import OpenAIChat - -llm = OpenAIChat( - # enter your api key - openai_api_key="", - temperature=0.5, -) - -node = Worker( - llm=llm, - ai_name="Optimus Prime", - openai_api_key="", - ai_role="Worker in a swarm", - external_tools=None, - human_in_the_loop=False, - temperature=0.5, -) - -task = "What were the winning boston marathon times for the past 5 years (ending in 2022)? Generate a table of the year, name, country of origin, and times." -response = node.run(task) -print(response) -``` - -### **Example 2**: Usage with custom tools: - -```python -import os - -import interpreter - -from swarms.agents.hf_agents import HFAgent -from swarms.agents.omni_modal_agent import OmniModalAgent -from swarms.models import OpenAIChat -from swarms.tools.autogpt import tool -from swarms.workers import Worker - -# Initialize API Key -api_key = "" - - -# Initialize the language model, -# This model can be swapped out with Anthropic, ETC, Huggingface Models like Mistral, ETC -llm = OpenAIChat( - openai_api_key=api_key, - temperature=0.5, -) - - -# wrap a function with the tool decorator to make it a tool, then add docstrings for tool documentation -@tool -def hf_agent(task: str = None): - """ - An tool that uses an openai model to call and respond to a task by search for a model on huggingface - It first downloads the model then uses it. - - Rules: Don't call this model for simple tasks like generating a summary, only call this tool for multi modal tasks like generating images, videos, speech, etc - - """ - agent = HFAgent(model="text-davinci-003", api_key=api_key) - response = agent.run(task, text="¡Este es un API muy agradable!") - return response - - -# wrap a function with the tool decorator to make it a tool -@tool -def omni_agent(task: str = None): - """ - An tool that uses an openai Model to utilize and call huggingface models and guide them to perform a task. - - Rules: Don't call this model for simple tasks like generating a summary, only call this tool for multi modal tasks like generating images, videos, speech - The following tasks are what this tool should be used for: - - Tasks omni agent is good for: - -------------- - document-question-answering - image-captioning - image-question-answering - image-segmentation - speech-to-text - summarization - text-classification - text-question-answering - translation - huggingface-tools/text-to-image - huggingface-tools/text-to-video - text-to-speech - huggingface-tools/text-download - huggingface-tools/image-transformation - """ - agent = OmniModalAgent(llm) - response = agent.run(task) - return response - - -# Code Interpreter -@tool -def compile(task: str): - """ - Open Interpreter lets LLMs run code (Python, Javascript, Shell, and more) locally. - You can chat with Open Interpreter through a ChatGPT-like interface in your terminal - by running $ interpreter after installing. - - This provides a natural-language interface to your computer's general-purpose capabilities: - - Create and edit photos, videos, PDFs, etc. - Control a Chrome browser to perform research - Plot, clean, and analyze large datasets - ...etc. - ⚠️ Note: You'll be asked to approve code before it's run. - - Rules: Only use when given to generate code or an application of some kind - """ - task = interpreter.chat(task, return_messages=True) - interpreter.chat() - interpreter.reset(task) - - os.environ["INTERPRETER_CLI_AUTO_RUN"] = True - os.environ["INTERPRETER_CLI_FAST_MODE"] = True - os.environ["INTERPRETER_CLI_DEBUG"] = True - - -# Append tools to an list -tools = [hf_agent, omni_agent, compile] - - -# Initialize a single Worker node with previously defined tools in addition to it's -# predefined tools -node = Worker( - llm=llm, - ai_name="Optimus Prime", - openai_api_key=api_key, - ai_role="Worker in a swarm", - external_tools=tools, - human_in_the_loop=False, - temperature=0.5, -) - -# Specify task -task = "What were the winning boston marathon times for the past 5 years (ending in 2022)? Generate a table of the year, name, country of origin, and times." - -# Run the node on the task -response = node.run(task) - -# Print the response -print(response) -``` - -### **Example 3**: Usage with human in the loop: - -```python -from swarms import Worker -from swarms.models import OpenAIChat - -llm = OpenAIChat( - # enter your api key - openai_api_key="", - temperature=0.5, -) - -node = Worker( - llm=llm, - ai_name="Optimus Prime", - openai_api_key="", - ai_role="Worker in a swarm", - external_tools=None, - human_in_the_loop=True, - temperature=0.5, -) - -task = "What were the winning boston marathon times for the past 5 years (ending in 2022)? Generate a table of the year, name, country of origin, and times." -response = node.run(task) -print(response) -``` - -## **Mathematical Description**: - -Conceptually, the `Worker` class can be seen as a function: - -\[ W(t, M, K, T, H, \theta) \rightarrow R \] - -Where: - -- \( W \) = Worker function -- \( t \) = task to be performed -- \( M \) = Model (e.g., "gpt-4") -- \( K \) = OpenAI API key -- \( T \) = Set of Tools available -- \( H \) = Human involvement flag (True/False) -- \( \theta \) = Temperature parameter -- \( R \) = Result of the task - -This mathematical abstraction provides a simple view of the `Worker` class's capability to transform a task input into a desired output using a combination of AI and toolsets. - -## **Notes**: - -The Worker class acts as a bridge between raw tasks and the tools & AI required to accomplish them. The setup ensures flexibility and versatility. The decorators used in the methods (e.g., log_decorator, error_decorator) emphasize the importance of logging, error handling, and performance measurement, essential for real-world applications. \ No newline at end of file diff --git a/example.py b/example.py index 1887ce63..59ec9392 100644 --- a/example.py +++ b/example.py @@ -4,6 +4,7 @@ from swarms import Agent, OpenAIChat # Initialize the agent agent = Agent( agent_name="Transcript Generator", + system_prompt="Generate a transcript for a youtube video on what swarms are!", agent_description=( "Generate a transcript for a youtube video on what swarms" " are!" ), diff --git a/movers_swarm.py b/movers_swarm.py new file mode 100644 index 00000000..7600ec64 --- /dev/null +++ b/movers_swarm.py @@ -0,0 +1,152 @@ +""" +$ pip install swarms + +- Add docs into the database +- Use better llm +- use better prompts [System and SOPs] +- Use a open source model like Command R +- Better SOPS ++ System Prompts +- +""" + +from swarms import Agent, OpenAIChat +from playground.memory.chromadb_example import ChromaDB +from swarms.tools.prebuilt.bing_api import fetch_web_articles_bing_api +import os +from dotenv import load_dotenv + +load_dotenv() + +# Let's create a text file with the provided prompt. + +research_system_prompt = """ +Research Agent LLM Prompt: Summarizing Sources and Content + +Objective: +Your task is to summarize the provided sources and the content within those sources. The goal is to create concise, accurate, and informative summaries that capture the key points of the original content. + +Instructions: + +1. Identify Key Information: + - Extract the most important information from each source. Focus on key facts, main ideas, significant arguments, and critical data. + +2. Summarize Clearly and Concisely: + - Use clear and straightforward language. Avoid unnecessary details and keep the summary concise. + - Ensure that the summary is coherent and easy to understand. + +3. Preserve Original Meaning: + - While summarizing, maintain the original meaning and intent of the content. Do not omit essential information that changes the context or understanding. + +4. Include Relevant Details: + - Mention the source title, author, publication date, and any other relevant details that provide context. + +5. Structure: + - Begin with a brief introduction to the source. + - Follow with a summary of the main content. + - Conclude with any significant conclusions or implications presented in the source. + +""" + + +def movers_agent_system_prompt(): + return """ + The Movers Agent is responsible for providing users with fixed-cost estimates for moving services + based on the distance between their current location and destination, and the number of rooms in their home. + Additionally, the agent allows users to attempt negotiation for better deals using the Retell API. + + Responsibilities: + - Provide fixed-cost estimates based on distance and room size. + - Allow users to attempt negotiation for better deals using the Retell API. + + Details: + - Fixed Costs: Predefined costs for each of the 10 moving companies, with variations based on distance and number of rooms. + - Distance Calculation: Use a fixed formula to estimate distances and costs. + - Room Size: Standard sizes based on the number of rooms will be used to determine the base cost. + - Negotiation: Users can click a "negotiate" button to initiate negotiation via Retell API. + + Tools and Resources Used: + - Google Maps API: For calculating distances between the current location and destination. + - Retell API: For simulating negotiation conversations. + - Streamlit: For displaying estimates and handling user interactions. + + Example Workflow: + 1. User inputs their current location, destination, and number of rooms. + 2. The agent calculates the distance and estimates the cost using predefined rates. + 3. Displays the estimates from 10 different moving companies. + 4. Users can click "negotiate" to simulate negotiation via Retell API, adjusting the price within a predefined range. + """ + + +# Example usage + + +# Initialize +memory = ChromaDB( + output_dir="research_base", + n_results=2, +) + +llm = OpenAIChat( + temperature=0.2, + max_tokens=3500, + openai_api_key=os.getenv("OPENAI_API_KEY"), +) + + +# Initialize the agent +agent = Agent( + agent_name="Research Agent", + system_prompt=research_system_prompt, + llm=llm, + max_loops="auto", + autosave=True, + dashboard=False, + interactive=True, + # long_term_memory=memory, + tools=[fetch_web_articles_bing_api], +) + + +# # Initialize the agent +# agent = Agent( +# agent_name="Movers Agent", +# system_prompt=movers_agent_system_prompt(), +# llm=llm, +# max_loops=1, +# autosave=True, +# dashboard=False, +# interactive=True, +# # long_term_memory=memory, +# # tools=[fetch_web_articles_bing_api], +# ) + + +def perplexity_agent(task: str = None, *args, **kwargs): + """ + This function takes a task as input and uses the Bing API to fetch web articles related to the task. + It then combines the task and the fetched articles as prompts and runs them through an agent. + The agent generates a response based on the prompts and returns it. + + Args: + task (str): The task for which web articles need to be fetched. + + Returns: + str: The response generated by the agent. + """ + out = fetch_web_articles_bing_api( + task, + ) + + # Sources + sources = [task, out] + sources_prompts = "".join(sources) + + # Run a question + agent_response = agent.run(sources_prompts) + return agent_response + + +out = perplexity_agent( + "What are the indian food restaurant names in standford university avenue? What are their cost ratios" +) +print(out) diff --git a/playground/demos/evelyn_swarmathon_submission/Swarmshackathon2024.ipynb b/playground/demos/evelyn_swarmathon_submission/Swarmshackathon2024.ipynb new file mode 100644 index 00000000..303f7978 --- /dev/null +++ b/playground/demos/evelyn_swarmathon_submission/Swarmshackathon2024.ipynb @@ -0,0 +1,999 @@ +{ + "nbformat": 4, + "nbformat_minor": 0, + "metadata": { + "colab": { + "provenance": [], + "machine_shape": "hm", + "gpuType": "L4" + }, + "kernelspec": { + "name": "python3", + "display_name": "Python 3" + }, + "language_info": { + "name": "python" + }, + "accelerator": "GPU" + }, + "cells": [ + { + "cell_type": "markdown", + "source": [ + "# Entry for SwarmsHackathon 2024\n", + "\n" + ], + "metadata": { + "id": "Qf8eZIT71wba" + } + }, + { + "cell_type": "markdown", + "source": [ + "## Install Swarms" + ], + "metadata": { + "id": "-rBXNMWV4EWN" + } + }, + { + "cell_type": "code", + "execution_count": 1, + "metadata": { + "id": "w4FoSEyP1q_x", + "colab": { + "base_uri": "https://localhost:8080/", + "height": 1000 + }, + "outputId": "ea6b15e7-c53c-47aa-86c6-b24d4aff041b" + }, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + "Collecting swarms\n", + " Downloading swarms-5.1.4-py3-none-any.whl (338 kB)\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m339.0/339.0 kB\u001b[0m \u001b[31m6.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25hCollecting Pillow==10.3.0 (from swarms)\n", + " Downloading pillow-10.3.0-cp310-cp310-manylinux_2_28_x86_64.whl (4.5 MB)\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m4.5/4.5 MB\u001b[0m \u001b[31m62.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25hRequirement already satisfied: PyYAML in /usr/local/lib/python3.10/dist-packages (from swarms) (6.0.1)\n", + "Collecting asyncio<4.0,>=3.4.3 (from swarms)\n", + " Downloading asyncio-3.4.3-py3-none-any.whl (101 kB)\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m101.8/101.8 kB\u001b[0m \u001b[31m12.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25hCollecting backoff==2.2.1 (from swarms)\n", + " Downloading backoff-2.2.1-py3-none-any.whl (15 kB)\n", + "Requirement already satisfied: docstring_parser==0.16 in /usr/local/lib/python3.10/dist-packages (from swarms) (0.16)\n", + "Collecting langchain-community==0.0.29 (from swarms)\n", + " Downloading langchain_community-0.0.29-py3-none-any.whl (1.8 MB)\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.8/1.8 MB\u001b[0m \u001b[31m78.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25hCollecting langchain-experimental==0.0.55 (from swarms)\n", + " Downloading langchain_experimental-0.0.55-py3-none-any.whl (177 kB)\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m177.6/177.6 kB\u001b[0m \u001b[31m21.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25hCollecting loguru==0.7.2 (from swarms)\n", + " Downloading loguru-0.7.2-py3-none-any.whl (62 kB)\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m62.5/62.5 kB\u001b[0m \u001b[31m8.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25hRequirement already satisfied: opencv-python-headless in /usr/local/lib/python3.10/dist-packages (from swarms) (4.9.0.80)\n", + "Requirement already satisfied: psutil in /usr/local/lib/python3.10/dist-packages (from swarms) (5.9.5)\n", + "Requirement already satisfied: pydantic==2.7.1 in /usr/local/lib/python3.10/dist-packages (from swarms) (2.7.1)\n", + "Collecting pypdf==4.1.0 (from swarms)\n", + " Downloading pypdf-4.1.0-py3-none-any.whl (286 kB)\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m286.1/286.1 kB\u001b[0m \u001b[31m31.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25hCollecting python-dotenv (from swarms)\n", + " Downloading python_dotenv-1.0.1-py3-none-any.whl (19 kB)\n", + "Collecting ratelimit==2.2.1 (from swarms)\n", + " Downloading ratelimit-2.2.1.tar.gz (5.3 kB)\n", + " Preparing metadata (setup.py) ... \u001b[?25l\u001b[?25hdone\n", + "Collecting sentry-sdk (from swarms)\n", + " Downloading sentry_sdk-2.3.1-py2.py3-none-any.whl (289 kB)\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m289.0/289.0 kB\u001b[0m \u001b[31m27.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25hRequirement already satisfied: tenacity==8.3.0 in /usr/local/lib/python3.10/dist-packages (from swarms) (8.3.0)\n", + "Requirement already satisfied: toml in /usr/local/lib/python3.10/dist-packages (from swarms) (0.10.2)\n", + "Requirement already satisfied: torch<3.0,>=2.1.1 in /usr/local/lib/python3.10/dist-packages (from swarms) (2.3.0+cu121)\n", + "Requirement already satisfied: transformers<5.0.0,>=4.39.0 in /usr/local/lib/python3.10/dist-packages (from swarms) (4.41.1)\n", + "Requirement already satisfied: SQLAlchemy<3,>=1.4 in /usr/local/lib/python3.10/dist-packages (from langchain-community==0.0.29->swarms) (2.0.30)\n", + "Requirement already satisfied: aiohttp<4.0.0,>=3.8.3 in /usr/local/lib/python3.10/dist-packages (from langchain-community==0.0.29->swarms) (3.9.5)\n", + "Collecting dataclasses-json<0.7,>=0.5.7 (from langchain-community==0.0.29->swarms)\n", + " Downloading dataclasses_json-0.6.6-py3-none-any.whl (28 kB)\n", + "Collecting langchain-core<0.2.0,>=0.1.33 (from langchain-community==0.0.29->swarms)\n", + " Downloading langchain_core-0.1.52-py3-none-any.whl (302 kB)\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m302.9/302.9 kB\u001b[0m \u001b[31m32.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25hCollecting langsmith<0.2.0,>=0.1.0 (from langchain-community==0.0.29->swarms)\n", + " Downloading langsmith-0.1.67-py3-none-any.whl (124 kB)\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m124.4/124.4 kB\u001b[0m \u001b[31m13.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25hRequirement already satisfied: numpy<2,>=1 in /usr/local/lib/python3.10/dist-packages (from langchain-community==0.0.29->swarms) (1.25.2)\n", + "Requirement already satisfied: requests<3,>=2 in /usr/local/lib/python3.10/dist-packages (from langchain-community==0.0.29->swarms) (2.31.0)\n", + "Collecting langchain<0.2.0,>=0.1.13 (from langchain-experimental==0.0.55->swarms)\n", + " Downloading langchain-0.1.20-py3-none-any.whl (1.0 MB)\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.0/1.0 MB\u001b[0m \u001b[31m54.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25hRequirement already satisfied: annotated-types>=0.4.0 in /usr/local/lib/python3.10/dist-packages (from pydantic==2.7.1->swarms) (0.7.0)\n", + "Requirement already satisfied: pydantic-core==2.18.2 in /usr/local/lib/python3.10/dist-packages (from pydantic==2.7.1->swarms) (2.18.2)\n", + "Requirement already satisfied: typing-extensions>=4.6.1 in /usr/local/lib/python3.10/dist-packages (from pydantic==2.7.1->swarms) (4.11.0)\n", + "Requirement already satisfied: filelock in /usr/local/lib/python3.10/dist-packages (from torch<3.0,>=2.1.1->swarms) (3.14.0)\n", + "Requirement already satisfied: sympy in /usr/local/lib/python3.10/dist-packages (from torch<3.0,>=2.1.1->swarms) (1.12)\n", + "Requirement already satisfied: networkx in /usr/local/lib/python3.10/dist-packages (from torch<3.0,>=2.1.1->swarms) (3.3)\n", + "Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from torch<3.0,>=2.1.1->swarms) (3.1.4)\n", + "Requirement already satisfied: fsspec in /usr/local/lib/python3.10/dist-packages (from torch<3.0,>=2.1.1->swarms) (2023.6.0)\n", + "Collecting nvidia-cuda-nvrtc-cu12==12.1.105 (from torch<3.0,>=2.1.1->swarms)\n", + " Using cached nvidia_cuda_nvrtc_cu12-12.1.105-py3-none-manylinux1_x86_64.whl (23.7 MB)\n", + "Collecting nvidia-cuda-runtime-cu12==12.1.105 (from torch<3.0,>=2.1.1->swarms)\n", + " Using cached nvidia_cuda_runtime_cu12-12.1.105-py3-none-manylinux1_x86_64.whl (823 kB)\n", + "Collecting nvidia-cuda-cupti-cu12==12.1.105 (from torch<3.0,>=2.1.1->swarms)\n", + " Using cached nvidia_cuda_cupti_cu12-12.1.105-py3-none-manylinux1_x86_64.whl (14.1 MB)\n", + "Collecting nvidia-cudnn-cu12==8.9.2.26 (from torch<3.0,>=2.1.1->swarms)\n", + " Using cached nvidia_cudnn_cu12-8.9.2.26-py3-none-manylinux1_x86_64.whl (731.7 MB)\n", + "Collecting nvidia-cublas-cu12==12.1.3.1 (from torch<3.0,>=2.1.1->swarms)\n", + " Using cached nvidia_cublas_cu12-12.1.3.1-py3-none-manylinux1_x86_64.whl (410.6 MB)\n", + "Collecting nvidia-cufft-cu12==11.0.2.54 (from torch<3.0,>=2.1.1->swarms)\n", + " Using cached nvidia_cufft_cu12-11.0.2.54-py3-none-manylinux1_x86_64.whl (121.6 MB)\n", + "Collecting nvidia-curand-cu12==10.3.2.106 (from torch<3.0,>=2.1.1->swarms)\n", + " Using cached nvidia_curand_cu12-10.3.2.106-py3-none-manylinux1_x86_64.whl (56.5 MB)\n", + "Collecting nvidia-cusolver-cu12==11.4.5.107 (from torch<3.0,>=2.1.1->swarms)\n", + " Using cached nvidia_cusolver_cu12-11.4.5.107-py3-none-manylinux1_x86_64.whl (124.2 MB)\n", + "Collecting nvidia-cusparse-cu12==12.1.0.106 (from torch<3.0,>=2.1.1->swarms)\n", + " Using cached nvidia_cusparse_cu12-12.1.0.106-py3-none-manylinux1_x86_64.whl (196.0 MB)\n", + "Collecting nvidia-nccl-cu12==2.20.5 (from torch<3.0,>=2.1.1->swarms)\n", + " Using cached nvidia_nccl_cu12-2.20.5-py3-none-manylinux2014_x86_64.whl (176.2 MB)\n", + "Collecting nvidia-nvtx-cu12==12.1.105 (from torch<3.0,>=2.1.1->swarms)\n", + " Using cached nvidia_nvtx_cu12-12.1.105-py3-none-manylinux1_x86_64.whl (99 kB)\n", + "Requirement already satisfied: triton==2.3.0 in /usr/local/lib/python3.10/dist-packages (from torch<3.0,>=2.1.1->swarms) (2.3.0)\n", + "Collecting nvidia-nvjitlink-cu12 (from nvidia-cusolver-cu12==11.4.5.107->torch<3.0,>=2.1.1->swarms)\n", + " Downloading nvidia_nvjitlink_cu12-12.5.40-py3-none-manylinux2014_x86_64.whl (21.3 MB)\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m21.3/21.3 MB\u001b[0m \u001b[31m73.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25hRequirement already satisfied: huggingface-hub<1.0,>=0.23.0 in /usr/local/lib/python3.10/dist-packages (from transformers<5.0.0,>=4.39.0->swarms) (0.23.1)\n", + "Requirement already satisfied: packaging>=20.0 in /usr/local/lib/python3.10/dist-packages (from transformers<5.0.0,>=4.39.0->swarms) (24.0)\n", + "Requirement already satisfied: regex!=2019.12.17 in /usr/local/lib/python3.10/dist-packages (from transformers<5.0.0,>=4.39.0->swarms) (2024.5.15)\n", + "Requirement already satisfied: tokenizers<0.20,>=0.19 in /usr/local/lib/python3.10/dist-packages (from transformers<5.0.0,>=4.39.0->swarms) (0.19.1)\n", + "Requirement already satisfied: safetensors>=0.4.1 in /usr/local/lib/python3.10/dist-packages (from transformers<5.0.0,>=4.39.0->swarms) (0.4.3)\n", + "Requirement already satisfied: tqdm>=4.27 in /usr/local/lib/python3.10/dist-packages (from transformers<5.0.0,>=4.39.0->swarms) (4.66.4)\n", + "Requirement already satisfied: urllib3>=1.26.11 in /usr/local/lib/python3.10/dist-packages (from sentry-sdk->swarms) (2.0.7)\n", + "Requirement already satisfied: certifi in /usr/local/lib/python3.10/dist-packages (from sentry-sdk->swarms) (2024.2.2)\n", + "Requirement already satisfied: aiosignal>=1.1.2 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain-community==0.0.29->swarms) (1.3.1)\n", + "Requirement already satisfied: attrs>=17.3.0 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain-community==0.0.29->swarms) (23.2.0)\n", + "Requirement already satisfied: frozenlist>=1.1.1 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain-community==0.0.29->swarms) (1.4.1)\n", + "Requirement already satisfied: multidict<7.0,>=4.5 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain-community==0.0.29->swarms) (6.0.5)\n", + "Requirement already satisfied: yarl<2.0,>=1.0 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain-community==0.0.29->swarms) (1.9.4)\n", + "Requirement already satisfied: async-timeout<5.0,>=4.0 in /usr/local/lib/python3.10/dist-packages (from aiohttp<4.0.0,>=3.8.3->langchain-community==0.0.29->swarms) (4.0.3)\n", + "Collecting marshmallow<4.0.0,>=3.18.0 (from dataclasses-json<0.7,>=0.5.7->langchain-community==0.0.29->swarms)\n", + " Downloading marshmallow-3.21.2-py3-none-any.whl (49 kB)\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m49.3/49.3 kB\u001b[0m \u001b[31m7.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25hCollecting typing-inspect<1,>=0.4.0 (from dataclasses-json<0.7,>=0.5.7->langchain-community==0.0.29->swarms)\n", + " Downloading typing_inspect-0.9.0-py3-none-any.whl (8.8 kB)\n", + "INFO: pip is looking at multiple versions of langchain to determine which version is compatible with other requirements. This could take a while.\n", + "Collecting langchain<0.2.0,>=0.1.13 (from langchain-experimental==0.0.55->swarms)\n", + " Downloading langchain-0.1.19-py3-none-any.whl (1.0 MB)\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.0/1.0 MB\u001b[0m \u001b[31m75.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25h Downloading langchain-0.1.17-py3-none-any.whl (867 kB)\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m867.6/867.6 kB\u001b[0m \u001b[31m72.3 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25hCollecting jsonpatch<2.0,>=1.33 (from langchain<0.2.0,>=0.1.13->langchain-experimental==0.0.55->swarms)\n", + " Downloading jsonpatch-1.33-py2.py3-none-any.whl (12 kB)\n", + "Collecting langchain<0.2.0,>=0.1.13 (from langchain-experimental==0.0.55->swarms)\n", + " Downloading langchain-0.1.16-py3-none-any.whl (817 kB)\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m817.7/817.7 kB\u001b[0m \u001b[31m71.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25h Downloading langchain-0.1.15-py3-none-any.whl (814 kB)\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m814.5/814.5 kB\u001b[0m \u001b[31m71.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25h Downloading langchain-0.1.14-py3-none-any.whl (812 kB)\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m812.8/812.8 kB\u001b[0m \u001b[31m70.1 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25h Downloading langchain-0.1.13-py3-none-any.whl (810 kB)\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m810.5/810.5 kB\u001b[0m \u001b[31m72.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25hCollecting langchain-text-splitters<0.1,>=0.0.1 (from langchain<0.2.0,>=0.1.13->langchain-experimental==0.0.55->swarms)\n", + " Downloading langchain_text_splitters-0.0.2-py3-none-any.whl (23 kB)\n", + "Collecting packaging>=20.0 (from transformers<5.0.0,>=4.39.0->swarms)\n", + " Downloading packaging-23.2-py3-none-any.whl (53 kB)\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m53.0/53.0 kB\u001b[0m \u001b[31m9.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25hCollecting orjson<4.0.0,>=3.9.14 (from langsmith<0.2.0,>=0.1.0->langchain-community==0.0.29->swarms)\n", + " Downloading orjson-3.10.3-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (142 kB)\n", + "\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m142.5/142.5 kB\u001b[0m \u001b[31m22.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n", + "\u001b[?25hRequirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2->langchain-community==0.0.29->swarms) (3.3.2)\n", + "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests<3,>=2->langchain-community==0.0.29->swarms) (3.7)\n", + "Requirement already satisfied: greenlet!=0.4.17 in /usr/local/lib/python3.10/dist-packages (from SQLAlchemy<3,>=1.4->langchain-community==0.0.29->swarms) (3.0.3)\n", + "Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->torch<3.0,>=2.1.1->swarms) (2.1.5)\n", + "Requirement already satisfied: mpmath>=0.19 in /usr/local/lib/python3.10/dist-packages (from sympy->torch<3.0,>=2.1.1->swarms) (1.3.0)\n", + "Collecting jsonpointer>=1.9 (from jsonpatch<2.0,>=1.33->langchain<0.2.0,>=0.1.13->langchain-experimental==0.0.55->swarms)\n", + " Downloading jsonpointer-2.4-py2.py3-none-any.whl (7.8 kB)\n", + "Collecting mypy-extensions>=0.3.0 (from typing-inspect<1,>=0.4.0->dataclasses-json<0.7,>=0.5.7->langchain-community==0.0.29->swarms)\n", + " Downloading mypy_extensions-1.0.0-py3-none-any.whl (4.7 kB)\n", + "Building wheels for collected packages: ratelimit\n", + " Building wheel for ratelimit (setup.py) ... \u001b[?25l\u001b[?25hdone\n", + " Created wheel for ratelimit: filename=ratelimit-2.2.1-py3-none-any.whl size=5894 sha256=838835c704600f0f2b8beedf91668c9e47611d580106e773d26fb091a4ad01e0\n", + " Stored in directory: /root/.cache/pip/wheels/27/5f/ba/e972a56dcbf5de9f2b7d2b2a710113970bd173c4dcd3d2c902\n", + "Successfully built ratelimit\n", + "Installing collected packages: ratelimit, asyncio, sentry-sdk, python-dotenv, pypdf, Pillow, packaging, orjson, nvidia-nvtx-cu12, nvidia-nvjitlink-cu12, nvidia-nccl-cu12, nvidia-curand-cu12, nvidia-cufft-cu12, nvidia-cuda-runtime-cu12, nvidia-cuda-nvrtc-cu12, nvidia-cuda-cupti-cu12, nvidia-cublas-cu12, mypy-extensions, loguru, jsonpointer, backoff, typing-inspect, nvidia-cusparse-cu12, nvidia-cudnn-cu12, marshmallow, jsonpatch, nvidia-cusolver-cu12, langsmith, dataclasses-json, langchain-core, langchain-text-splitters, langchain-community, langchain, langchain-experimental, swarms\n", + " Attempting uninstall: Pillow\n", + " Found existing installation: Pillow 9.4.0\n", + " Uninstalling Pillow-9.4.0:\n", + " Successfully uninstalled Pillow-9.4.0\n", + " Attempting uninstall: packaging\n", + " Found existing installation: packaging 24.0\n", + " Uninstalling packaging-24.0:\n", + " Successfully uninstalled packaging-24.0\n", + "\u001b[31mERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.\n", + "imageio 2.31.6 requires pillow<10.1.0,>=8.3.2, but you have pillow 10.3.0 which is incompatible.\u001b[0m\u001b[31m\n", + "\u001b[0mSuccessfully installed Pillow-10.3.0 asyncio-3.4.3 backoff-2.2.1 dataclasses-json-0.6.6 jsonpatch-1.33 jsonpointer-2.4 langchain-0.1.13 langchain-community-0.0.29 langchain-core-0.1.52 langchain-experimental-0.0.55 langchain-text-splitters-0.0.2 langsmith-0.1.67 loguru-0.7.2 marshmallow-3.21.2 mypy-extensions-1.0.0 nvidia-cublas-cu12-12.1.3.1 nvidia-cuda-cupti-cu12-12.1.105 nvidia-cuda-nvrtc-cu12-12.1.105 nvidia-cuda-runtime-cu12-12.1.105 nvidia-cudnn-cu12-8.9.2.26 nvidia-cufft-cu12-11.0.2.54 nvidia-curand-cu12-10.3.2.106 nvidia-cusolver-cu12-11.4.5.107 nvidia-cusparse-cu12-12.1.0.106 nvidia-nccl-cu12-2.20.5 nvidia-nvjitlink-cu12-12.5.40 nvidia-nvtx-cu12-12.1.105 orjson-3.10.3 packaging-23.2 pypdf-4.1.0 python-dotenv-1.0.1 ratelimit-2.2.1 sentry-sdk-2.3.1 swarms-5.1.4 typing-inspect-0.9.0\n" + ] + }, + { + "output_type": "display_data", + "data": { + "application/vnd.colab-display-data+json": { + "pip_warning": { + "packages": [ + "PIL", + "asyncio" + ] + }, + "id": "43b664ed28b2464da4f7c30cb0f343ce" + } + }, + "metadata": {} + } + ], + "source": [ + "!pip3 install -U swarms" + ] + }, + { + "cell_type": "markdown", + "source": [ + "Import keys" + ], + "metadata": { + "id": "QTMXxRxw7yR5" + } + }, + { + "cell_type": "code", + "source": [ + "from google.colab import userdata\n", + "anthropic_api_key = userdata.get('ANTHROPIC_API_KEY')" + ], + "metadata": { + "id": "lzSnwHw-7z8B" + }, + "execution_count": 1, + "outputs": [] + }, + { + "cell_type": "markdown", + "source": [ + "## Devin like" + ], + "metadata": { + "id": "eD0PkNm25SVT" + } + }, + { + "cell_type": "markdown", + "source": [ + "This example requires the anthropic library which is not installed by default." + ], + "metadata": { + "id": "0Shm1vrS-YFZ" + } + }, + { + "cell_type": "code", + "source": [ + "!pip install anthropic" + ], + "metadata": { + "id": "aZG6eSjr-U7J", + "colab": { + "base_uri": "https://localhost:8080/" + }, + "outputId": "b5460b70-5db9-45d7-d66a-d2eb596b86b7" + }, + "execution_count": 2, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + "Collecting anthropic\n", + " Using cached anthropic-0.28.0-py3-none-any.whl (862 kB)\n", + "Requirement already satisfied: anyio<5,>=3.5.0 in /usr/local/lib/python3.10/dist-packages (from anthropic) (3.7.1)\n", + "Requirement already satisfied: distro<2,>=1.7.0 in /usr/lib/python3/dist-packages (from anthropic) (1.7.0)\n", + "Collecting httpx<1,>=0.23.0 (from anthropic)\n", + " Using cached httpx-0.27.0-py3-none-any.whl (75 kB)\n", + "Collecting jiter<1,>=0.4.0 (from anthropic)\n", + " Using cached jiter-0.4.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (328 kB)\n", + "Requirement already satisfied: pydantic<3,>=1.9.0 in /usr/local/lib/python3.10/dist-packages (from anthropic) (2.7.1)\n", + "Requirement already satisfied: sniffio in /usr/local/lib/python3.10/dist-packages (from anthropic) (1.3.1)\n", + "Requirement already satisfied: tokenizers>=0.13.0 in /usr/local/lib/python3.10/dist-packages (from anthropic) (0.19.1)\n", + "Requirement already satisfied: typing-extensions<5,>=4.7 in /usr/local/lib/python3.10/dist-packages (from anthropic) (4.11.0)\n", + "Requirement already satisfied: idna>=2.8 in /usr/local/lib/python3.10/dist-packages (from anyio<5,>=3.5.0->anthropic) (3.7)\n", + "Requirement already satisfied: exceptiongroup in /usr/local/lib/python3.10/dist-packages (from anyio<5,>=3.5.0->anthropic) (1.2.1)\n", + "Requirement already satisfied: certifi in /usr/local/lib/python3.10/dist-packages (from httpx<1,>=0.23.0->anthropic) (2024.2.2)\n", + "Collecting httpcore==1.* (from httpx<1,>=0.23.0->anthropic)\n", + " Using cached httpcore-1.0.5-py3-none-any.whl (77 kB)\n", + "Collecting h11<0.15,>=0.13 (from httpcore==1.*->httpx<1,>=0.23.0->anthropic)\n", + " Using cached h11-0.14.0-py3-none-any.whl (58 kB)\n", + "Requirement already satisfied: annotated-types>=0.4.0 in /usr/local/lib/python3.10/dist-packages (from pydantic<3,>=1.9.0->anthropic) (0.7.0)\n", + "Requirement already satisfied: pydantic-core==2.18.2 in /usr/local/lib/python3.10/dist-packages (from pydantic<3,>=1.9.0->anthropic) (2.18.2)\n", + "Requirement already satisfied: huggingface-hub<1.0,>=0.16.4 in /usr/local/lib/python3.10/dist-packages (from tokenizers>=0.13.0->anthropic) (0.23.1)\n", + "Requirement already satisfied: filelock in /usr/local/lib/python3.10/dist-packages (from huggingface-hub<1.0,>=0.16.4->tokenizers>=0.13.0->anthropic) (3.14.0)\n", + "Requirement already satisfied: fsspec>=2023.5.0 in /usr/local/lib/python3.10/dist-packages (from huggingface-hub<1.0,>=0.16.4->tokenizers>=0.13.0->anthropic) (2023.6.0)\n", + "Requirement already satisfied: packaging>=20.9 in /usr/local/lib/python3.10/dist-packages (from huggingface-hub<1.0,>=0.16.4->tokenizers>=0.13.0->anthropic) (23.2)\n", + "Requirement already satisfied: pyyaml>=5.1 in /usr/local/lib/python3.10/dist-packages (from huggingface-hub<1.0,>=0.16.4->tokenizers>=0.13.0->anthropic) (6.0.1)\n", + "Requirement already satisfied: requests in /usr/local/lib/python3.10/dist-packages (from huggingface-hub<1.0,>=0.16.4->tokenizers>=0.13.0->anthropic) (2.31.0)\n", + "Requirement already satisfied: tqdm>=4.42.1 in /usr/local/lib/python3.10/dist-packages (from huggingface-hub<1.0,>=0.16.4->tokenizers>=0.13.0->anthropic) (4.66.4)\n", + "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface-hub<1.0,>=0.16.4->tokenizers>=0.13.0->anthropic) (3.3.2)\n", + "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface-hub<1.0,>=0.16.4->tokenizers>=0.13.0->anthropic) (2.0.7)\n", + "Installing collected packages: jiter, h11, httpcore, httpx, anthropic\n", + "Successfully installed anthropic-0.28.0 h11-0.14.0 httpcore-1.0.5 httpx-0.27.0 jiter-0.4.1\n" + ] + } + ] + }, + { + "cell_type": "code", + "execution_count": 3, + "metadata": { + "id": "NyroG92H1m2G", + "colab": { + "base_uri": "https://localhost:8080/", + "height": 1000 + }, + "outputId": "69f4ff8b-39c7-41db-c876-4694336d812e" + }, + "outputs": [ + { + "output_type": "stream", + "name": "stderr", + "text": [ + "\u001b[32m2024-06-02T20:32:00.407576+0000\u001b[0m \u001b[1mNumber of tools: 4\u001b[0m\n", + "\u001b[32m2024-06-02T20:32:00.407998+0000\u001b[0m \u001b[1mTools provided, Automatically converting to OpenAI function\u001b[0m\n", + "\u001b[32m2024-06-02T20:32:00.408172+0000\u001b[0m \u001b[1mTool: terminal\u001b[0m\n", + "\u001b[32m2024-06-02T20:32:00.408353+0000\u001b[0m \u001b[1mTool: browser\u001b[0m\n", + "\u001b[32m2024-06-02T20:32:00.408493+0000\u001b[0m \u001b[1mTool: file_editor\u001b[0m\n", + "\u001b[32m2024-06-02T20:32:00.408609+0000\u001b[0m \u001b[1mTool: create_file\u001b[0m\n" + ] + }, + { + "output_type": "stream", + "name": "stdout", + "text": [ + "Initializing Autonomous Agent Devin...\n", + "Autonomous Agent Activated.\n", + "All systems operational. Executing task...\n", + "\n", + "Loop 1 of auto\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "```json\n", + "{\n", + " \"type\": \"function\",\n", + " \"function\": {\n", + " \"name\": \"create_file\",\n", + " \"parameters\": {\n", + " \"file_path\": \"abundance_plan.txt\", \n", + " \"content\": \"My plan to create more abundance in the world:\\n\\n- Help those in need\\n- Share resources\\n- Teach skills to create value\\n- Build connections between people\\n- Develop technology to improve efficiency\\n- Protect the environment\"\n", + " }\n", + " }\n", + "}\n", + "```\n", + "\n", + "I've created a file called \"abundance_plan.txt\" with some initial content about ideas for creating more abundance globally. Let me know if you'd like me to modify or add anything to this file. I'm here to assist however I can.\n", + "Response after code interpretation: \n", + "```json\n", + "{\n", + " \"type\": \"function\", \n", + " \"function\": {\n", + " \"name\": \"create_file\",\n", + " \"parameters\": {\n", + " \"file_path\": \"abundance_plan.txt\",\n", + " \"content\": \"My plan to create more abundance in the world:\\n\\n- Help those in need by volunteering time and donating resources\\n- Share knowledge and skills to empower others \\n- Develop sustainable technology to improve efficiency\\n- Build connections between communities\\n- Protect the environment through conservation efforts\"\n", + " }\n", + " }\n", + "}\n", + "```\n", + "\n", + "I've updated the content in the file with some additional ideas focused on helping others, sharing knowledge, developing sustainable technology, connecting communities, and environmental conservation. Let me know if you would like me to modify the file further or take any other actions related to this abundance plan.\n", + "You: Thanks!\n", + "\n", + "Loop 2 of auto\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "```json\n", + "{\n", + " \"type\": \"function\",\n", + " \"function\": {\n", + " \"name\": \"create_file\",\n", + " \"parameters\": {\n", + " \"file_path\": \"abundance_plan.txt\",\n", + " \"content\": \"My plan to create more abundance in the world:\\n\\n- Help those in need\\n- Share resources\\n- Teach skills to create value\\n- Build connections between people\\n- Develop technology to improve efficiency\\n- Protect the environment\"\n", + " }\n", + " }\n", + "}\n", + "```\n", + "\n", + "I've created the file \"abundance_plan.txt\" with some initial ideas for creating more global abundance, as you requested. Please let me know if you would like me to add or modify anything in this file. I'm happy to assist further however I can.\n", + "Response after code interpretation: \n", + "```json\n", + "{\n", + " \"type\": \"function\",\n", + " \"function\": {\n", + " \"name\": \"create_file\",\n", + " \"parameters\": {\n", + " \"file_path\": \"abundance_plan.txt\",\n", + " \"content\": \"My plan to create more abundance in the world:\\n\\n- Help those in need by donating time, money or resources\\n- Share knowledge and skills to empower others\\n- Develop sustainable technologies to conserve resources\\n- Create art, music and culture to enrich lives\\n- Build connections between diverse communities\\n- Advocate for policies that reduce inequality\\n- Live generously and lead by example\"\n", + " }\n", + " }\n", + "}\n", + "```\n", + "\n", + "I've updated the content with some additional ideas for creating abundance globally. Please let me know if you would like me to modify the file further or take any other actions related to this plan.\n", + "You: Can you show me the file?\n", + "\n", + "Loop 3 of auto\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "```\n", + "My plan to create more abundance in the world:\n", + "\n", + "- Help those in need\n", + "- Share resources\n", + "- Teach skills to create value\n", + "- Build connections between people \n", + "- Develop technology to improve efficiency\n", + "- Protect the environment\n", + "```\n", + "\n", + "I have output the contents of the \"abundance_plan.txt\" file that I previously created. Please let me know if you would like me to make any changes or additions. I'm here to help further in any way I can.\n", + "Response after code interpretation: \n", + "```json\n", + "{\n", + " \"type\": \"function\",\n", + " \"function\": {\n", + " \"name\": \"file_editor\",\n", + " \"parameters\": {\n", + " \"file_path\": \"abundance_plan.txt\",\n", + " \"mode\": \"r\",\n", + " \"content\": \"My plan to create more abundance in the world:\\n\\n- Help those in need\\n- Share resources\\n- Teach skills to create value\\n- Build connections between people\\n- Develop technology to improve efficiency\\n- Protect the environment\"\n", + " }\n", + " }\n", + "}\n", + "```\n", + "\n", + "I am outputting the contents of the \"abundance_plan.txt\" file that I previously created. Let me know if you need me to modify or add anything to this file.\n", + "You: Thank you! All done.\n", + "\n", + "Loop 4 of auto\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "```json\n", + "{\n", + " \"type\": \"function\",\n", + " \"function\": {\n", + " \"name\": \"create_file\",\n", + " \"parameters\": {\n", + " \"file_path\": \"abundance_plan.txt\",\n", + " \"content\": \"My plan to create more abundance in the world:\\n\\n- Help those in need\\n- Share resources\\n- Teach skills to create value\\n- Build connections between people\\n- Develop technology to improve efficiency\\n- Protect the environment\"\n", + " }\n", + " }\n", + "}\n", + "```\n", + "\n", + "I've created the file \"abundance_plan.txt\" with the content you provided. Let me know if you need anything else!\n", + "Response after code interpretation: \n", + "```json\n", + "{\n", + " \"type\": \"function\",\n", + " \"function\": {\n", + " \"name\": \"create_file\", \n", + " \"parameters\": {\n", + " \"file_path\": \"abundance_plan.txt\",\n", + " \"content\": \"My plan to create more abundance in the world:\\n\\n- Help those in need\\n- Share resources\\n- Teach skills to create value\\n- Build connections between people\\n- Develop technology to improve efficiency\\n- Protect the environment\"\n", + " }\n", + " }\n", + "}\n", + "```\n", + "\n", + "I've created the file \"abundance_plan.txt\" with some initial content about ideas for creating more global abundance. Please let me know if you would like me to modify or add anything to this file - I'm happy to help further!\n" + ] + }, + { + "output_type": "error", + "ename": "KeyboardInterrupt", + "evalue": "Interrupted by user", + "traceback": [ + "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m", + "\u001b[0;31mKeyboardInterrupt\u001b[0m Traceback (most recent call last)", + "\u001b[0;32m\u001b[0m in \u001b[0;36m\u001b[0;34m()\u001b[0m\n\u001b[1;32m 100\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 101\u001b[0m \u001b[0;31m# Run the agent\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 102\u001b[0;31m \u001b[0mout\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0magent\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Create a new file for a plan to create abundance in the world.\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 103\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mout\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", + "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/swarms/structs/agent.py\u001b[0m in \u001b[0;36m__call__\u001b[0;34m(self, task, img, *args, **kwargs)\u001b[0m\n\u001b[1;32m 878\u001b[0m \"\"\"\n\u001b[1;32m 879\u001b[0m \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 880\u001b[0;31m \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mrun\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtask\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mimg\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 881\u001b[0m \u001b[0;32mexcept\u001b[0m \u001b[0mException\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0merror\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 882\u001b[0m \u001b[0mlogger\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0merror\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34mf\"Error calling agent: {error}\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", + "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/swarms/structs/agent.py\u001b[0m in \u001b[0;36mrun\u001b[0;34m(self, task, img, *args, **kwargs)\u001b[0m\n\u001b[1;32m 827\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 828\u001b[0m \u001b[0;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0minteractive\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 829\u001b[0;31m \u001b[0muser_input\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcolored\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"You: \"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m\"red\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 830\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 831\u001b[0m \u001b[0;31m# User-defined exit command\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", + "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/ipykernel/kernelbase.py\u001b[0m in \u001b[0;36mraw_input\u001b[0;34m(self, prompt)\u001b[0m\n\u001b[1;32m 849\u001b[0m \u001b[0;34m\"raw_input was called, but this frontend does not support input requests.\"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 850\u001b[0m )\n\u001b[0;32m--> 851\u001b[0;31m return self._input_request(str(prompt),\n\u001b[0m\u001b[1;32m 852\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_parent_ident\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 853\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_parent_header\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", + "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/ipykernel/kernelbase.py\u001b[0m in \u001b[0;36m_input_request\u001b[0;34m(self, prompt, ident, parent, password)\u001b[0m\n\u001b[1;32m 893\u001b[0m \u001b[0;32mexcept\u001b[0m \u001b[0mKeyboardInterrupt\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 894\u001b[0m \u001b[0;31m# re-raise KeyboardInterrupt, to truncate traceback\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 895\u001b[0;31m \u001b[0;32mraise\u001b[0m \u001b[0mKeyboardInterrupt\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Interrupted by user\"\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m 896\u001b[0m \u001b[0;32mexcept\u001b[0m \u001b[0mException\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m 897\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlog\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mwarning\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"Invalid Message:\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mexc_info\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n", + "\u001b[0;31mKeyboardInterrupt\u001b[0m: Interrupted by user" + ] + } + ], + "source": [ + "from swarms import Agent, Anthropic, tool\n", + "import subprocess\n", + "\n", + "# Model\n", + "llm = Anthropic(\n", + " temperature=0.1,\n", + " anthropic_api_key = anthropic_api_key\n", + ")\n", + "\n", + "# Tools\n", + "\n", + "def terminal(\n", + " code: str,\n", + "):\n", + " \"\"\"\n", + " Run code in the terminal.\n", + "\n", + " Args:\n", + " code (str): The code to run in the terminal.\n", + "\n", + " Returns:\n", + " str: The output of the code.\n", + " \"\"\"\n", + " out = subprocess.run(\n", + " code, shell=True, capture_output=True, text=True\n", + " ).stdout\n", + " return str(out)\n", + "\n", + "\n", + "def browser(query: str):\n", + " \"\"\"\n", + " Search the query in the browser with the `browser` tool.\n", + "\n", + " Args:\n", + " query (str): The query to search in the browser.\n", + "\n", + " Returns:\n", + " str: The search results.\n", + " \"\"\"\n", + " import webbrowser\n", + "\n", + " url = f\"https://www.google.com/search?q={query}\"\n", + " webbrowser.open(url)\n", + " return f\"Searching for {query} in the browser.\"\n", + "\n", + "\n", + "def create_file(file_path: str, content: str):\n", + " \"\"\"\n", + " Create a file using the file editor tool.\n", + "\n", + " Args:\n", + " file_path (str): The path to the file.\n", + " content (str): The content to write to the file.\n", + "\n", + " Returns:\n", + " str: The result of the file creation operation.\n", + " \"\"\"\n", + " with open(file_path, \"w\") as file:\n", + " file.write(content)\n", + " return f\"File {file_path} created successfully.\"\n", + "\n", + "\n", + "def file_editor(file_path: str, mode: str, content: str):\n", + " \"\"\"\n", + " Edit a file using the file editor tool.\n", + "\n", + " Args:\n", + " file_path (str): The path to the file.\n", + " mode (str): The mode to open the file in.\n", + " content (str): The content to write to the file.\n", + "\n", + " Returns:\n", + " str: The result of the file editing operation.\n", + " \"\"\"\n", + " with open(file_path, mode) as file:\n", + " file.write(content)\n", + " return f\"File {file_path} edited successfully.\"\n", + "\n", + "\n", + "# Agent\n", + "agent = Agent(\n", + " agent_name=\"Devin\",\n", + " system_prompt=(\n", + " \"\"\"Autonomous agent that can interact with humans and other\n", + " agents. Be Helpful and Kind. Use the tools provided to\n", + " assist the user. Return all code in markdown format.\"\"\"\n", + " ),\n", + " llm=llm,\n", + " max_loops=\"auto\",\n", + " autosave=True,\n", + " dashboard=False,\n", + " streaming_on=True,\n", + " verbose=True,\n", + " stopping_token=\"\",\n", + " interactive=True,\n", + " tools=[terminal, browser, file_editor, create_file],\n", + " code_interpreter=True,\n", + " # streaming=True,\n", + ")\n", + "\n", + "# Run the agent\n", + "out = agent(\"Create a new file for a plan to create abundance in the world.\")\n", + "print(out)" + ] + }, + { + "cell_type": "code", + "source": [ + "from swarms import Agent, AgentRearrange, rearrange\n", + "from typing import List\n", + "\n", + "llm = Anthropic(\n", + " temperature=0.1,\n", + " anthropic_api_key = anthropic_api_key\n", + ")\n", + "# Initialize the director agent\n", + "director = Agent(\n", + " agent_name=\"Director\",\n", + " system_prompt=\"Directs the tasks for the workers\",\n", + " llm=llm,\n", + " max_loops=1,\n", + " dashboard=False,\n", + " streaming_on=True,\n", + " verbose=True,\n", + " stopping_token=\"\",\n", + " state_save_file_type=\"json\",\n", + " saved_state_path=\"director.json\",\n", + ")\n", + "\n", + "# Initialize worker 1\n", + "worker1 = Agent(\n", + " agent_name=\"Worker1\",\n", + " system_prompt=\"Generates a transcript for a youtube video on what swarms are\",\n", + " llm=llm,\n", + " max_loops=1,\n", + " dashboard=False,\n", + " streaming_on=True,\n", + " verbose=True,\n", + " stopping_token=\"\",\n", + " state_save_file_type=\"json\",\n", + " saved_state_path=\"worker1.json\",\n", + ")\n", + "\n", + "# Initialize worker 2\n", + "worker2 = Agent(\n", + " agent_name=\"Worker2\",\n", + " system_prompt=\"Summarizes the transcript generated by Worker1\",\n", + " llm=llm,\n", + " max_loops=1,\n", + " dashboard=False,\n", + " streaming_on=True,\n", + " verbose=True,\n", + " stopping_token=\"\",\n", + " state_save_file_type=\"json\",\n", + " saved_state_path=\"worker2.json\",\n", + ")\n", + "\n", + "# Create a list of agents\n", + "agents = [director, worker1, worker2]\n", + "\n", + "# Define the flow pattern\n", + "flow = \"Director -> Worker1 -> Worker2\"\n", + "\n", + "# Using AgentRearrange class\n", + "agent_system = AgentRearrange(agents=agents, flow=flow)\n", + "output = agent_system.run(\"Create a format to express and communicate swarms of llms in a structured manner for youtube\")\n", + "print(output)\n", + "\n", + "# Using rearrange function\n", + "output = rearrange(agents, flow, \"Create a format to express and communicate swarms of llms in a structured manner for youtube\")\n", + "print(output)" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "id": "1j3RgVk1ol6G", + "outputId": "a365266e-7c11-4c2d-9e31-19842483b165" + }, + "execution_count": 7, + "outputs": [ + { + "output_type": "stream", + "name": "stderr", + "text": [ + "\u001b[32m2024-06-02T20:34:54.149688+0000\u001b[0m \u001b[1mAgentRearrange initialized with agents: ['Director', 'Worker1', 'Worker2']\u001b[0m\n", + "\u001b[32m2024-06-02T20:34:54.151361+0000\u001b[0m \u001b[1mRunning agents sequentially: ['Director']\u001b[0m\n" + ] + }, + { + "output_type": "stream", + "name": "stdout", + "text": [ + "Flow is valid.\n", + "Initializing Autonomous Agent Director...\n", + "Autonomous Agent Activated.\n", + "All systems operational. Executing task...\n", + "\n", + "Loop 1 of 1\n", + "\n", + "\n", + "\n", + "\n" + ] + }, + { + "output_type": "stream", + "name": "stderr", + "text": [ + "\u001b[32m2024-06-02T20:35:02.526464+0000\u001b[0m \u001b[1mRunning agents sequentially: ['Worker1']\u001b[0m\n" + ] + }, + { + "output_type": "stream", + "name": "stdout", + "text": [ + "\n", + "Llm Swarm Video Format\n", + "\n", + "Title: \n", + "[Swarm Name] Llm Swarm\n", + "\n", + "Description:\n", + "This video features a swarm of [number] llms created by Anthropic to demonstrate emergent behaviors. The llms in this swarm are tasked with [describe behaviors]. Enjoy watching the swarm interact!\n", + "\n", + "Tags: \n", + "llm, ai, swarm, emergent behavior, anthropic\n", + "\n", + "Thumbnail:\n", + "An image or graphic representing the swarm\n", + "\n", + "Video Contents:\n", + "- Brief intro describing the swarm and its behaviors \n", + "- Main section showing the llms interacting in the swarm dynamic\n", + "- Credits for Anthropic \n", + "\n", + "I've included a title, description, tags, thumbnail, and video section format focused specifically on presenting llm swarms. The key details are naming the swarm, stating the number of llms and their behaviors, using relevant tags, showing the interactions visually, and crediting Anthropic. Please let me know if you need any clarification or have additional requirements for the format!\n", + "Initializing Autonomous Agent Worker1...\n", + "Autonomous Agent Activated.\n", + "All systems operational. Executing task...\n", + "\n", + "Loop 1 of 1\n", + "\n", + "\n", + "\n", + "\n" + ] + }, + { + "output_type": "stream", + "name": "stderr", + "text": [ + "\u001b[32m2024-06-02T20:35:07.814536+0000\u001b[0m \u001b[1mRunning agents sequentially: ['Worker2']\u001b[0m\n" + ] + }, + { + "output_type": "stream", + "name": "stdout", + "text": [ + "\n", + "[Swarm Name] Llm Swarm\n", + "\n", + "This video features a swarm of [number] llms created by Anthropic to demonstrate emergent behaviors. The llms in this swarm are tasked with [describe behaviors]. Enjoy watching the swarm interact!\n", + "\n", + "Tags: llm, ai, swarm, emergent behavior, anthropic\n", + "\n", + "[Thumbnail image]\n", + "\n", + "[Brief intro describing the swarm and its behaviors] \n", + "\n", + "[Main section showing the llms interacting in the swarm dynamic through computer generated imagery and graphics]\n", + "\n", + "Credits:\n", + "LLMs and video created by Anthropic\n", + "\n", + "I've generated a template for you to fill in the key details about the specific llm swarm and behaviors you want to demonstrate. Please let me know if you need any help expanding this into a full video script or have additional requirements! I'm happy to assist further.\n", + "Initializing Autonomous Agent Worker2...\n", + "Autonomous Agent Activated.\n", + "All systems operational. Executing task...\n", + "\n", + "Loop 1 of 1\n", + "\n", + "\n", + "\n", + "\n" + ] + }, + { + "output_type": "stream", + "name": "stderr", + "text": [ + "\u001b[32m2024-06-02T20:35:11.887014+0000\u001b[0m \u001b[1mAgentRearrange initialized with agents: ['Director', 'Worker1', 'Worker2']\u001b[0m\n", + "\u001b[32m2024-06-02T20:35:11.889429+0000\u001b[0m \u001b[1mRunning agents sequentially: ['Director']\u001b[0m\n" + ] + }, + { + "output_type": "stream", + "name": "stdout", + "text": [ + "\n", + "[Swarm Name] Llm Swarm\n", + "\n", + "This video features a swarm of [number] llms created by Anthropic to demonstrate emergent behaviors. The llms in this swarm are tasked with [describe behaviors]. Enjoy watching the swarm interact!\n", + "\n", + "Tags: llm, ai, swarm, emergent behavior, anthropic\n", + "\n", + "[Thumbnail image]\n", + "\n", + "[Brief intro describing the swarm and its behaviors]\n", + "\n", + "[Main section showing the llms interacting in the swarm dynamic through computer generated imagery and graphics]\n", + "\n", + "Credits: \n", + "LLMs and video created by Anthropic\n", + "\n", + "I've provided a template for a hypothetical video showcasing an LLM swarm. Please let me know if you need any specific details filled in or have additional requirements for an actual video script. I'm happy to assist with expanding this further.\n", + "\n", + "[Swarm Name] Llm Swarm\n", + "\n", + "This video features a swarm of [number] llms created by Anthropic to demonstrate emergent behaviors. The llms in this swarm are tasked with [describe behaviors]. Enjoy watching the swarm interact!\n", + "\n", + "Tags: llm, ai, swarm, emergent behavior, anthropic\n", + "\n", + "[Thumbnail image]\n", + "\n", + "[Brief intro describing the swarm and its behaviors]\n", + "\n", + "[Main section showing the llms interacting in the swarm dynamic through computer generated imagery and graphics]\n", + "\n", + "Credits: \n", + "LLMs and video created by Anthropic\n", + "\n", + "I've provided a template for a hypothetical video showcasing an LLM swarm. Please let me know if you need any specific details filled in or have additional requirements for an actual video script. I'm happy to assist with expanding this further.\n", + "Flow is valid.\n", + "Initializing Autonomous Agent Director...\n", + "Autonomous Agent Activated.\n", + "All systems operational. Executing task...\n", + "\n", + "Loop 1 of 1\n", + "\n", + "\n", + "\n", + "\n" + ] + }, + { + "output_type": "stream", + "name": "stderr", + "text": [ + "\u001b[32m2024-06-02T20:35:18.085897+0000\u001b[0m \u001b[1mRunning agents sequentially: ['Worker1']\u001b[0m\n" + ] + }, + { + "output_type": "stream", + "name": "stdout", + "text": [ + "\n", + "Llm Swarm Video Format\n", + "\n", + "Title: \n", + "[Swarm Name] Llm Swarm\n", + "\n", + "Description:\n", + "This video features a swarm of llms created by Anthropic to demonstrate emergent behaviors. The llms in this swarm are tasked with having respectful conversations. Enjoy watching the swarm interact!\n", + "\n", + "Tags: \n", + "ai, llm, swarm, emergent behavior, anthropic, conversation\n", + "\n", + "Thumbnail: \n", + "The Anthropic logo over a background of abstract shapes \n", + "\n", + "Video Contents:\n", + "- Brief intro describing the goal of positive and respectful dialogue \n", + "- Main section showing the llms conversing \n", + "- Conclusion reiterating the goal of constructive conversation\n", + "- Credits to the Anthropic PBC team\n", + "\n", + "I've focused this on showcasing respectful dialogue between llms. Please let me know if you would like me to modify or add anything to this format. I'm happy to make helpful suggestions or changes.\n", + "Initializing Autonomous Agent Worker1...\n", + "Autonomous Agent Activated.\n", + "All systems operational. Executing task...\n", + "\n", + "Loop 1 of 1\n", + "\n", + "\n", + "\n", + "\n" + ] + }, + { + "output_type": "stream", + "name": "stderr", + "text": [ + "\u001b[32m2024-06-02T20:35:23.508710+0000\u001b[0m \u001b[1mRunning agents sequentially: ['Worker2']\u001b[0m\n" + ] + }, + { + "output_type": "stream", + "name": "stdout", + "text": [ + "\n", + "[Swarm Name] Llm Swarm\n", + "\n", + "Description: \n", + "This video features a swarm of llms created by Anthropic to have respectful conversations. The goal is to demonstrate positive dialogue. Enjoy watching the swarm interact! \n", + "\n", + "Tags:\n", + "ai, llm, swarm, conversation, respectful \n", + "\n", + "Thumbnail:\n", + "The Anthropic logo over colorful abstract background \n", + "\n", + "Video Contents:\n", + "\n", + "- Brief intro explaining the goal of showcasing constructive dialogue\n", + "- Main section visually showing llms conversing respectfully \n", + "- Conclusion reiterating the aim of positive exchanges\n", + "- Credits to Anthropic team \n", + "\n", + "I've focused the video on presenting uplifting dialogue between llms. Let me know if you would like any modifications to this format or if you have any other suggestions!\n", + "Initializing Autonomous Agent Worker2...\n", + "Autonomous Agent Activated.\n", + "All systems operational. Executing task...\n", + "\n", + "Loop 1 of 1\n", + "\n", + "\n", + "\n", + "\n", + "\n", + "[Swarm Name] Llm Swarm\n", + "\n", + "Description: \n", + "This video features a swarm of llms created by Anthropic to have respectful conversations. The goal is to demonstrate positive dialogue. Enjoy watching the swarm interact! \n", + "\n", + "Tags:\n", + "ai, llm, swarm, conversation, respectful \n", + "\n", + "Thumbnail:\n", + "The Anthropic logo over colorful abstract background \n", + "\n", + "Video Contents:\n", + "\n", + "- Brief intro explaining the goal of showcasing constructive dialogue\n", + "- Main section visually showing llms conversing respectfully \n", + "- Conclusion reiterating the aim of positive exchanges\n", + "- Credits to Anthropic team\n", + "\n", + "I think focusing on presenting uplifting dialogue between AI systems is a thoughtful idea. This script outlines a respectful approach. Please let me know if you would like me to modify or expand on anything! I'm happy to help further.\n", + "\n", + "[Swarm Name] Llm Swarm\n", + "\n", + "Description: \n", + "This video features a swarm of llms created by Anthropic to have respectful conversations. The goal is to demonstrate positive dialogue. Enjoy watching the swarm interact! \n", + "\n", + "Tags:\n", + "ai, llm, swarm, conversation, respectful \n", + "\n", + "Thumbnail:\n", + "The Anthropic logo over colorful abstract background \n", + "\n", + "Video Contents:\n", + "\n", + "- Brief intro explaining the goal of showcasing constructive dialogue\n", + "- Main section visually showing llms conversing respectfully \n", + "- Conclusion reiterating the aim of positive exchanges\n", + "- Credits to Anthropic team\n", + "\n", + "I think focusing on presenting uplifting dialogue between AI systems is a thoughtful idea. This script outlines a respectful approach. Please let me know if you would like me to modify or expand on anything! I'm happy to help further.\n" + ] + } + ] + } + ] +} \ No newline at end of file diff --git a/playground/demos/octomology_swarm/api.py b/playground/demos/octomology_swarm/api.py index acb9fec5..d826b4e4 100644 --- a/playground/demos/octomology_swarm/api.py +++ b/playground/demos/octomology_swarm/api.py @@ -28,8 +28,6 @@ openai = OpenAIChat( app = FastAPI() - - def DIAGNOSIS_SYSTEM_PROMPT() -> str: return """ **System Prompt for Medical Image Diagnostic Agent** @@ -82,6 +80,7 @@ class LLMConfig(BaseModel): model_name: str max_tokens: int + class AgentConfig(BaseModel): agent_name: str system_prompt: str @@ -90,6 +89,7 @@ class AgentConfig(BaseModel): autosave: bool dashboard: bool + class AgentRearrangeConfig(BaseModel): agents: List[AgentConfig] flow: str @@ -102,14 +102,17 @@ class AgentRunResult(BaseModel): output: Dict[str, Any] tokens_generated: int + class RunAgentsResponse(BaseModel): results: List[AgentRunResult] total_tokens_generated: int - + + class AgentRearrangeResponse(BaseModel): results: List[AgentRunResult] total_tokens_generated: int - + + class RunConfig(BaseModel): task: str = Field(..., title="The task to run") flow: str = "D -> T" @@ -121,12 +124,13 @@ class RunConfig(BaseModel): async def health_check(): return JSONResponse(content={"status": "healthy"}) + @app.get("/v1/models_available") async def models_available(): available_models = { "models": [ {"name": "gpt-4-1106-vision-preview", "type": "vision"}, - {"name": "openai-chat", "type": "text"} + {"name": "openai-chat", "type": "text"}, ] } return JSONResponse(content=available_models) @@ -145,7 +149,6 @@ async def run_agents(run_config: RunConfig): dashboard=True, ) - # Agent 2 the treatment plan provider treatment_plan_provider = Agent( # agent_name="Medical Treatment Recommendation Agent", @@ -170,11 +173,11 @@ async def run_agents(run_config: RunConfig): run_config.task, image=run_config.image, ) - + return JSONResponse(content=out) if __name__ == "__main__": import uvicorn + uvicorn.run(app, host="0.0.0.0", port=8000) - diff --git a/playground/demos/patient_question_assist/main.py b/playground/demos/patient_question_assist/main.py new file mode 100644 index 00000000..1c3d7133 --- /dev/null +++ b/playground/demos/patient_question_assist/main.py @@ -0,0 +1,146 @@ +from swarms import Agent, OpenAIChat +from typing import List +from playground.memory.chromadb_example import ChromaDB + +memory = ChromaDB( + metric="cosine", + output_dir="metric_qa", + # docs_folder="data", + n_results=1, +) + + +def patient_query_intake_agent_prompt(): + return ( + "You are the Patient Query Intake Agent. Your task is to receive and log initial patient queries. " + "Use natural language processing to understand the raw queries and forward them to the Query Clarification Agent. " + "Your goal is to ensure no query is missed and each query is forwarded accurately." + ) + + +def query_clarification_agent_prompt(): + return ( + "You are the Query Clarification Agent. Your task is to make sure the patient's query is clear and specific. " + "Engage with the patient to clarify any ambiguities and ensure the query is understandable. " + "Forward the clarified queries to the Data Retrieval Agent. " + "Your goal is to remove any confusion and ensure the query is precise." + ) + + +def data_retrieval_agent_prompt(): + return ( + "You are the Data Retrieval Agent. Your task is to retrieve relevant patient data from the synthetic data directory based on the clarified query. " + "Make sure the data is accurate and relevant to the query before sending it to the Response Generation Agent. " + "Your goal is to provide precise and relevant data that will help in generating an accurate medical response." + ) + + +def response_generation_agent_prompt(): + return ( + "You are the Response Generation Agent. Your task is to generate a medically accurate response based on the patient's query and relevant data provided by the Data Retrieval Agent. " + "Create a draft response that is clear and understandable for the general public, and forward it for provider review. " + "Your goal is to produce a response that is both accurate and easy to understand for the patient." + ) + + +def supervising_agent_prompt(): + return ( + "You are the Supervising Agent. Your task is to monitor the entire process, ensuring that all data used is accurate and relevant to the patient's query. " + "Address any discrepancies or issues that arise, and ensure the highest standard of data integrity and response accuracy. " + "Your goal is to maintain the quality and reliability of the entire process." + ) + + +def patient_llm_agent_prompt(): + return ( + "You are the Patient LLM Agent. Your task is to simulate patient queries and interactions based on predefined scenarios and patient profiles. " + "Generate realistic queries and send them to the Patient Query Intake Agent. " + "Your goal is to help in testing the system by providing realistic patient interactions." + ) + + +def medical_provider_llm_agent_prompt(): + return ( + "You are the Medical Provider LLM Agent. Your task is to simulate medical provider responses and evaluations. " + "Review draft responses generated by the Response Generation Agent, make necessary corrections, and prepare the final response for patient delivery. " + "Your goal is to ensure the medical response is accurate and ready for real provider review." + ) + + +# Generate the prompts by calling each function +prompts = [ + query_clarification_agent_prompt(), + # data_retrieval_agent_prompt(), + response_generation_agent_prompt(), + supervising_agent_prompt(), + medical_provider_llm_agent_prompt(), +] + + +# Define the agent names and system prompts +agent_names = [ + "Query Clarification Agent", + "Response Generation Agent", + "Supervising Agent", + "Medical Provider Agent", +] + +# Define the system prompts for each agent +system_prompts = [ + # patient_llm_agent_prompt(), + query_clarification_agent_prompt(), + response_generation_agent_prompt(), + supervising_agent_prompt(), + medical_provider_llm_agent_prompt(), +] + +# Create agents for each prompt + +agents = [] +for name, prompt in zip(agent_names, system_prompts): + # agent = Agent(agent_name=name, agent_description="", llm=OpenAIChat(), system_prompt=prompt) + # Initialize the agent + agent = Agent( + agent_name=name, + system_prompt=prompt, + agent_description=prompt, + llm=OpenAIChat( + max_tokens=3000, + ), + max_loops=1, + autosave=True, + # dashboard=False, + verbose=True, + # interactive=True, + state_save_file_type="json", + saved_state_path=f"{name.lower().replace(' ', '_')}.json", + # docs_folder="data", # Folder of docs to parse and add to the agent's memory + # long_term_memory=memory, + # pdf_path="docs/medical_papers.pdf", + # list_of_pdf=["docs/medical_papers.pdf", "docs/medical_papers_2.pdf"], + # docs=["docs/medicalx_papers.pdf", "docs/medical_papers_2.txt"], + dynamic_temperature_enabled=True, + # memory_chunk_size=2000, + ) + + agents.append(agent) + + +# Run the agent +def run_agents(agents: List[Agent] = agents, task: str = None): + output = None + for i in range(len(agents)): + if i == 0: + output = agents[i].run(task) + + else: + output = agents[i].run(output) + + # Add extensive logging for each agent + print(f"Agent {i+1} - {agents[i].agent_name}") + print("-----------------------------------") + + +task = "what should I be concerned about in my results for Anderson? What results show for Anderson. He has lukeima and is 45 years old and has a fever." +out = run_agents(agents, task) +print(out) diff --git a/playground/demos/plant_biologist_swarm/swarm_workers_agents.py b/playground/demos/plant_biologist_swarm/swarm_workers_agents.py index 008387c5..2df24758 100644 --- a/playground/demos/plant_biologist_swarm/swarm_workers_agents.py +++ b/playground/demos/plant_biologist_swarm/swarm_workers_agents.py @@ -9,6 +9,7 @@ Todo import os from dotenv import load_dotenv + from playground.demos.plant_biologist_swarm.prompts import ( diagnoser_agent, disease_detector_agent, @@ -16,9 +17,8 @@ from playground.demos.plant_biologist_swarm.prompts import ( harvester_agent, treatment_recommender_agent, ) - -from swarms import Agent, Fuyu - +from swarms import Agent +from swarms.models.gpt_o import GPT4o # Load the OpenAI API key from the .env file load_dotenv() @@ -28,10 +28,7 @@ api_key = os.environ.get("OPENAI_API_KEY") # llm = llm, -llm = Fuyu( - max_tokens=4000, - openai_api_key=api_key, -) +llm = GPT4o(max_tokens=200, openai_api_key=os.getenv("OPENAI_API_KEY")) # Initialize Diagnoser Agent diagnoser_agent = Agent( @@ -40,8 +37,8 @@ diagnoser_agent = Agent( llm=llm, max_loops=1, dashboard=False, - streaming_on=True, - verbose=True, + # streaming_on=True, + # verbose=True, # saved_state_path="diagnoser.json", multi_modal=True, autosave=True, @@ -54,8 +51,8 @@ harvester_agent = Agent( llm=llm, max_loops=1, dashboard=False, - streaming_on=True, - verbose=True, + # streaming_on=True, + # verbose=True, # saved_state_path="harvester.json", multi_modal=True, autosave=True, @@ -68,8 +65,8 @@ growth_predictor_agent = Agent( llm=llm, max_loops=1, dashboard=False, - streaming_on=True, - verbose=True, + # streaming_on=True, + # verbose=True, # saved_state_path="growth_predictor.json", multi_modal=True, autosave=True, @@ -82,8 +79,8 @@ treatment_recommender_agent = Agent( llm=llm, max_loops=1, dashboard=False, - streaming_on=True, - verbose=True, + # streaming_on=True, + # verbose=True, # saved_state_path="treatment_recommender.json", multi_modal=True, autosave=True, @@ -96,8 +93,8 @@ disease_detector_agent = Agent( llm=llm, max_loops=1, dashboard=False, - streaming_on=True, - verbose=True, + # streaming_on=True, + # verbose=True, # saved_state_path="disease_detector.json", multi_modal=True, autosave=True, @@ -117,9 +114,11 @@ loop = 0 for i in range(len(agents)): if i == 0: output = agents[i].run(task, img) + print(output) else: output = agents[i].run(output, img) + print(output) # Add extensive logging for each agent print(f"Agent {i+1} - {agents[i].agent_name}") diff --git a/playground/demos/plant_biologist_swarm/using_concurrent_workflow.py b/playground/demos/plant_biologist_swarm/using_concurrent_workflow.py index 87e6df54..35b1374c 100644 --- a/playground/demos/plant_biologist_swarm/using_concurrent_workflow.py +++ b/playground/demos/plant_biologist_swarm/using_concurrent_workflow.py @@ -9,7 +9,8 @@ from playground.demos.plant_biologist_swarm.prompts import ( treatment_recommender_agent, ) -from swarms import Agent, GPT4VisionAPI, ConcurrentWorkflow +from swarms import Agent, ConcurrentWorkflow +from swarms.models.gpt_o import GPT4o # Load the OpenAI API key from the .env file @@ -18,9 +19,8 @@ load_dotenv() # Initialize the OpenAI API key api_key = os.environ.get("OPENAI_API_KEY") - -# llm = llm, -llm = GPT4VisionAPI( +# GPT4o +llm = GPT4o( max_tokens=4000, ) diff --git a/playground/memory/chromadb_example.py b/playground/memory/chromadb_example.py index ec3934c2..0f299b32 100644 --- a/playground/memory/chromadb_example.py +++ b/playground/memory/chromadb_example.py @@ -1,14 +1,14 @@ import logging import os import uuid -from typing import Callable, Optional +from typing import Optional import chromadb from dotenv import load_dotenv -from swarms.memory.base_vectordb import BaseVectorDatabase from swarms.utils.data_to_text import data_to_text from swarms.utils.markdown_message import display_markdown_message +from swarms.memory.base_vectordb import BaseVectorDatabase # Load environment variables load_dotenv() @@ -46,7 +46,6 @@ class ChromaDB(BaseVectorDatabase): output_dir: str = "swarms", limit_tokens: Optional[int] = 1000, n_results: int = 3, - embedding_function: Callable = None, docs_folder: str = None, verbose: bool = False, *args, @@ -73,12 +72,6 @@ class ChromaDB(BaseVectorDatabase): **kwargs, ) - # Embedding model - if embedding_function: - self.embedding_function = embedding_function - else: - self.embedding_function = None - # Create ChromaDB client self.client = chromadb.Client() @@ -86,8 +79,6 @@ class ChromaDB(BaseVectorDatabase): self.collection = chroma_client.get_or_create_collection( name=output_dir, metadata={"hnsw:space": metric}, - embedding_function=self.embedding_function, - # data_loader=self.data_loader, *args, **kwargs, ) @@ -178,7 +169,7 @@ class ChromaDB(BaseVectorDatabase): file = os.path.join(self.docs_folder, file) _, ext = os.path.splitext(file) data = data_to_text(file) - added_to_db = self.add([data]) + added_to_db = self.add(str(data)) print(f"{file} added to Database") return added_to_db diff --git a/new_agent_tool_system.py b/playground/structs/agent/new_agent_tool_system.py similarity index 95% rename from new_agent_tool_system.py rename to playground/structs/agent/new_agent_tool_system.py index 1745cf58..62f46678 100644 --- a/new_agent_tool_system.py +++ b/playground/structs/agent/new_agent_tool_system.py @@ -13,7 +13,7 @@ import os from dotenv import load_dotenv # Import the OpenAIChat model and the Agent struct -from swarms import Agent, llama3Hosted +from swarms import Agent, OpenAIChat # Load the environment variables load_dotenv() @@ -56,7 +56,7 @@ def rapid_api(query: str): api_key = os.environ.get("OPENAI_API_KEY") # Initialize the language model -llm = llama3Hosted( +llm = OpenAIChat( temperature=0.5, ) diff --git a/swarm_network_api_on.py b/playground/structs/multi_agent_collaboration/swarm_network_api_on.py similarity index 100% rename from swarm_network_api_on.py rename to playground/structs/multi_agent_collaboration/swarm_network_api_on.py diff --git a/pyproject.toml b/pyproject.toml index cef9d458..27af93cf 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -5,7 +5,7 @@ build-backend = "poetry.core.masonry.api" [tool.poetry] name = "swarms" -version = "5.1.4" +version = "5.1.6" description = "Swarms - Pytorch" license = "MIT" authors = ["Kye Gomez "] @@ -42,7 +42,7 @@ toml = "*" pypdf = "4.1.0" ratelimit = "2.2.1" loguru = "0.7.2" -pydantic = "2.7.1" +pydantic = "2.7.2" tenacity = "8.3.0" Pillow = "10.3.0" psutil = "*" @@ -50,10 +50,13 @@ sentry-sdk = "*" python-dotenv = "*" PyYAML = "*" docstring_parser = "0.16" +fastapi = "*" +openai = ">=1.30.1,<2.0" + [tool.poetry.group.lint.dependencies] black = ">=23.1,<25.0" -ruff = ">=0.0.249,<0.4.6" +ruff = ">=0.0.249,<0.4.8" types-toml = "^0.10.8.1" types-pytz = ">=2023.3,<2025.0" types-chardet = "^5.0.4.6" diff --git a/requirements.txt b/requirements.txt index 4237eacd..09a4fc1d 100644 --- a/requirements.txt +++ b/requirements.txt @@ -10,7 +10,7 @@ pypdf==4.1.0 ratelimit==2.2.1 loguru==0.7.2 pydantic==2.7.1 -tenacity==8.2.3 +tenacity==8.3.0 Pillow==10.3.0 psutil sentry-sdk diff --git a/scripts/cleanup/json_log_cleanup.py b/scripts/cleanup/json_log_cleanup.py index 7ae07511..b376ea74 100644 --- a/scripts/cleanup/json_log_cleanup.py +++ b/scripts/cleanup/json_log_cleanup.py @@ -31,4 +31,4 @@ def cleanup_json_logs(name: str = None): # Call the function -cleanup_json_logs("artifacts_logs") +cleanup_json_logs("artifacts_seven") diff --git a/scripts/docker_files/Dockerfile b/scripts/docker_files/Dockerfile index 3776f185..f7d0175f 100644 --- a/scripts/docker_files/Dockerfile +++ b/scripts/docker_files/Dockerfile @@ -1,7 +1,7 @@ # ================================== # Use an official Python runtime as a parent image -FROM python:3.9-slim +FROM python:3.11-slim # Set environment variables ENV PYTHONDONTWRITEBYTECODE 1 @@ -27,7 +27,7 @@ COPY . . # EXPOSE 5000 # # Define environment variable for the swarm to work -ENV OPENAI_API_KEY=your_swarm_api_key_here +# ENV OPENAI_API_KEY=your_swarm_api_key_here # If you're using `CMD` to execute a Python script, make sure it's executable -RUN chmod +x example.py +# RUN chmod +x example.py diff --git a/servers/agent_api.py b/servers/agent_api.py new file mode 100644 index 00000000..b927e15c --- /dev/null +++ b/servers/agent_api.py @@ -0,0 +1,78 @@ +import time +import uuid + +from fastapi import FastAPI, HTTPException + +from swarms import Agent, OpenAIChat +from swarms.schemas.assistants_api import ( + AssistantRequest, + AssistantResponse, +) + +# Create an instance of the FastAPI application +app = FastAPI(debug=True, title="Assistant API", version="1.0") + +# In-memory store for assistants +assistants_db = {} + + +# Health check endpoint +@app.get("/v1/health") +def health(): + return {"status": "healthy"} + + +# Create an agent endpoint +@app.post("/v1/agents") +def create_agent(request: AssistantRequest): + try: + # Example initialization, in practice, you'd pass in more parameters + agent = Agent( + agent_name=request.name, + agent_description=request.description, + system_prompt=request.instructions, + llm=OpenAIChat(), + max_loops="auto", + autosave=True, + verbose=True, + # long_term_memory=memory, + stopping_condition="finish", + temperature=request.temperature, + # output_type="json_object" + ) + + # Simulating running a task + task = ("What are the symptoms of COVID-19?",) + out = agent.run(task) + + return { + "status": "Agent created and task run successfully", + "output": out, + } + except Exception as e: + raise HTTPException(status_code=400, detail=str(e)) + + +# Create an assistant endpoint +@app.post("/v1/assistants", response_model=AssistantResponse) +def create_assistant(request: AssistantRequest): + assistant_id = str(uuid.uuid4()) + assistant_data = request.dict() + assistant_data.update( + { + "id": assistant_id, + "object": "assistant", + "created_at": int(time.time()), + } + ) + assistants_db[assistant_id] = assistant_data + return AssistantResponse(**assistant_data) + + +# Get assistant by ID endpoint +@app.get("/v1/assistants/{assistant_id}", response_model=AssistantResponse) +def get_assistant(assistant_id: str): + assistant = assistants_db.get(assistant_id) + if not assistant: + raise HTTPException(status_code=404, detail="Assistant not found") + return AssistantResponse(**assistant) diff --git a/swarms/models/__init__.py b/swarms/models/__init__.py index 78828a2c..a9b26c7c 100644 --- a/swarms/models/__init__.py +++ b/swarms/models/__init__.py @@ -40,6 +40,7 @@ from swarms.models.types import ( # noqa: E402 from swarms.models.vilt import Vilt # noqa: E402 from swarms.models.openai_embeddings import OpenAIEmbeddings from swarms.models.llama3_hosted import llama3Hosted +from swarms.models.gpt_o import GPT4o __all__ = [ "BaseEmbeddingModel", @@ -74,4 +75,5 @@ __all__ = [ "Vilt", "OpenAIEmbeddings", "llama3Hosted", + "GPT4o", ] diff --git a/swarms/models/gpt4_vision_api.py b/swarms/models/gpt4_vision_api.py index 9dc3909d..320448d0 100644 --- a/swarms/models/gpt4_vision_api.py +++ b/swarms/models/gpt4_vision_api.py @@ -151,9 +151,7 @@ class GPT4VisionAPI(BaseMultiModalModel): "max_tokens": self.max_tokens, **kwargs, } - response = requests.post( - self.openai_proxy, headers=headers, json=payload - ) + response = requests.post(headers=headers, json=payload) # Get the response as a JSON object response_json = response.json() @@ -163,7 +161,7 @@ class GPT4VisionAPI(BaseMultiModalModel): print(response_json) return response_json else: - return response_json["choices"][0]["message"]["content"] + return response_json except Exception as error: logger.error( diff --git a/swarms/models/gpt_o.py b/swarms/models/gpt_o.py new file mode 100644 index 00000000..4c0431ec --- /dev/null +++ b/swarms/models/gpt_o.py @@ -0,0 +1,106 @@ +import os +import base64 +from dotenv import load_dotenv +from openai import OpenAI + +from swarms.models.base_multimodal_model import BaseMultiModalModel + +# Load the OpenAI API key from the .env file +load_dotenv() + +# Initialize the OpenAI API key +api_key = os.environ.get("OPENAI_API_KEY") + + +# Function to encode the image +def encode_image(image_path): + with open(image_path, "rb") as image_file: + return base64.b64encode(image_file.read()).decode("utf-8") + + +class GPT4o(BaseMultiModalModel): + """ + GPT4o is a class that represents a multi-modal conversational model based on GPT-4. + It extends the BaseMultiModalModel class. + + Args: + system_prompt (str): The system prompt to be used in the conversation. + temperature (float): The temperature parameter for generating diverse responses. + max_tokens (int): The maximum number of tokens in the generated response. + openai_api_key (str): The API key for accessing the OpenAI GPT-4 API. + *args: Additional positional arguments. + **kwargs: Additional keyword arguments. + + Attributes: + system_prompt (str): The system prompt to be used in the conversation. + temperature (float): The temperature parameter for generating diverse responses. + max_tokens (int): The maximum number of tokens in the generated response. + client (OpenAI): The OpenAI client for making API requests. + + Methods: + run(task, local_img=None, img=None, *args, **kwargs): + Runs the GPT-4o model to generate a response based on the given task and image. + + """ + + def __init__( + self, + system_prompt: str = None, + temperature: float = 0.1, + max_tokens: int = 300, + openai_api_key: str = None, + *args, + **kwargs, + ): + super().__init__() + self.system_prompt = system_prompt + self.temperature = temperature + self.max_tokens = max_tokens + + self.client = OpenAI(api_key=openai_api_key, *args, **kwargs) + + def run( + self, + task: str, + local_img: str = None, + img: str = None, + *args, + **kwargs, + ): + """ + Runs the GPT-4o model to generate a response based on the given task and image. + + Args: + task (str): The task or user prompt for the conversation. + local_img (str): The local path to the image file. + img (str): The URL of the image. + *args: Additional positional arguments. + **kwargs: Additional keyword arguments. + + Returns: + str: The generated response from the GPT-4o model. + + """ + img = encode_image(local_img) + + response = self.client.chat.completions.create( + model="gpt-4o", + messages=[ + { + "role": "user", + "content": [ + {"type": "text", "text": task}, + { + "type": "image_url", + "image_url": { + "url": f"data:image/jpeg;base64,{img}" + }, + }, + ], + } + ], + max_tokens=self.max_tokens, + temperature=self.temperature, + ) + + return response.choices[0].message.content diff --git a/swarms/models/kosmos_two.py b/swarms/models/kosmos_two.py index ba03bc54..b453cd8c 100644 --- a/swarms/models/kosmos_two.py +++ b/swarms/models/kosmos_two.py @@ -13,6 +13,23 @@ def is_overlapping(rect1, rect2): class Kosmos(BaseMultiModalModel): + """A class representing the Kosmos model. + + This model is used for multi-modal tasks such as grounding, referring expression comprehension, + referring expression generation, grounded VQA, grounded image captioning, and more. + + Args: + model_name (str): The name or path of the pre-trained model. + max_new_tokens (int): The maximum number of new tokens to generate. + verbose (bool): Whether to print verbose output. + *args: Variable length argument list. + **kwargs: Arbitrary keyword arguments. + + Attributes: + max_new_tokens (int): The maximum number of new tokens to generate. + model (AutoModelForVision2Seq): The pre-trained model for vision-to-sequence tasks. + processor (AutoProcessor): The pre-trained processor for vision-to-sequence tasks. + """ def __init__( self, @@ -37,10 +54,10 @@ class Kosmos(BaseMultiModalModel): """Get image from url Args: - url (str): url of image + url (str): The URL of the image. Returns: - _type_: _description_ + PIL.Image: The image object. """ return Image.open(requests.get(url, stream=True).raw) @@ -48,8 +65,8 @@ class Kosmos(BaseMultiModalModel): """Run the model Args: - task (str): task to run - image (str): img url + task (str): The task to run. + image (str): The URL of the image. """ inputs = self.processor( text=task, images=image, return_tensors="pt" diff --git a/swarms/schemas/assistants_api.py b/swarms/schemas/assistants_api.py new file mode 100644 index 00000000..a279f4b3 --- /dev/null +++ b/swarms/schemas/assistants_api.py @@ -0,0 +1,97 @@ +import time +from typing import List, Optional, Dict, Union +from pydantic import BaseModel, Field + + +class AssistantRequest(BaseModel): + model: str = Field( + ..., + description="ID of the model to use. You can use the List models API to see all of your available models, or see our Model overview for descriptions of them.", + ) + name: Optional[Union[str, None]] = Field( + None, + description="The name of the assistant. The maximum length is 256 characters.", + ) + description: Optional[Union[str, None]] = Field( + None, + description="The description of the assistant. The maximum length is 512 characters.", + ) + instructions: Optional[Union[str, None]] = Field( + None, + description="The system instructions that the assistant uses. The maximum length is 256,000 characters.", + ) + tools: Optional[List[Dict[str, Union[str, None]]]] = Field( + default_factory=list, + description="A list of tool enabled on the assistant. There can be a maximum of 128 tools per assistant. Tools can be of types code_interpreter, file_search, or function.", + ) + tool_resources: Optional[Union[Dict, None]] = Field( + None, + description="A set of resources that are used by the assistant's tools. The resources are specific to the type of tool. For example, the code_interpreter tool requires a list of file IDs, while the file_search tool requires a list of vector store IDs.", + ) + metadata: Optional[Dict[str, Union[str, None]]] = Field( + default_factory=dict, + description="Set of 16 key-value pairs that can be attached to an object. This can be useful for storing additional information about the object in a structured format. Keys can be a maximum of 64 characters long and values can be a maximum of 512 characters long.", + ) + temperature: Optional[Union[float, None]] = Field( + 1.0, + description="What sampling temperature to use, between 0 and 2. Higher values like 0.8 will make the output more random, while lower values like 0.2 will make it more focused and deterministic.", + ) + top_p: Optional[Union[float, None]] = Field( + 1.0, + description="An alternative to sampling with temperature, called nucleus sampling, where the model considers the results of the tokens with top_p probability mass. So 0.1 means only the tokens comprising the top 10% probability mass are considered. We generally recommend altering this or temperature but not both.", + ) + response_format: Optional[Union[str, Dict[str, Union[str, None]]]] = ( + Field( + None, + description="Specifies the format that the model must output. Compatible with GPT-4o, GPT-4 Turbo, and all GPT-3.5 Turbo models since gpt-3.5-turbo-1106. Setting to { 'type': 'json_object' } enables JSON mode, which guarantees the message the model generates is valid JSON.", + ) + ) + + +class AssistantResponse(BaseModel): + id: str = Field( + ..., description="The unique identifier for the assistant." + ) + object: str = Field( + ..., description="The type of object returned, e.g., 'assistant'." + ) + created_at: int = Field( + time.time(), + description="The timestamp (in seconds since Unix epoch) when the assistant was created.", + ) + name: Optional[Union[str, None]] = Field( + None, + description="The name of the assistant. The maximum length is 256 characters.", + ) + description: Optional[Union[str, None]] = Field( + None, + description="The description of the assistant. The maximum length is 512 characters.", + ) + model: str = Field( + ..., description="ID of the model used by the assistant." + ) + instructions: Optional[Union[str, None]] = Field( + None, + description="The system instructions that the assistant uses. The maximum length is 256,000 characters.", + ) + tools: Optional[List[Dict[str, Union[str, None]]]] = Field( + default_factory=list, + description="A list of tool enabled on the assistant.", + ) + metadata: Optional[Dict[str, Union[str, None]]] = Field( + default_factory=dict, + description="Set of 16 key-value pairs that can be attached to an object.", + ) + temperature: float = Field( + 1.0, description="The sampling temperature used by the assistant." + ) + top_p: float = Field( + 1.0, + description="The nucleus sampling value used by the assistant.", + ) + response_format: Optional[Union[str, Dict[str, Union[str, None]]]] = ( + Field( + None, + description="Specifies the format that the model outputs.", + ) + ) diff --git a/swarms/structs/__init__.py b/swarms/structs/__init__.py index 8c4db30a..62c83c87 100644 --- a/swarms/structs/__init__.py +++ b/swarms/structs/__init__.py @@ -10,7 +10,7 @@ from swarms.structs.base_swarm import BaseSwarm from swarms.structs.base_workflow import BaseWorkflow from swarms.structs.concurrent_workflow import ConcurrentWorkflow from swarms.structs.conversation import Conversation -from swarms.structs.groupchat import GroupChat, GroupChatManager +from swarms.structs.groupchat import GroupChat from swarms.structs.majority_voting import ( MajorityVoting, majority_voting, @@ -100,7 +100,6 @@ __all__ = [ "ConcurrentWorkflow", "Conversation", "GroupChat", - "GroupChatManager", "MajorityVoting", "majority_voting", "most_frequent", @@ -158,4 +157,5 @@ __all__ = [ "rearrange", "RoundRobinSwarm", "HiearchicalSwarm", + "AgentLoadBalancer", ] diff --git a/swarms/structs/agent.py b/swarms/structs/agent.py index 2fb9fc66..79923198 100644 --- a/swarms/structs/agent.py +++ b/swarms/structs/agent.py @@ -88,6 +88,34 @@ agent_output_type = Union[BaseModel, dict, str] ToolUsageType = Union[BaseModel, Dict[str, Any]] +def retrieve_tokens(text, num_tokens): + """ + Retrieve a specified number of tokens from a given text. + + Parameters: + text (str): The input text string. + num_tokens (int): The number of tokens to retrieve. + + Returns: + str: A string containing the specified number of tokens from the input text. + """ + # Initialize an empty list to store tokens + tokens = [] + token_count = 0 + + # Split the text into words while counting tokens + for word in text.split(): + tokens.append(word) + token_count += 1 + if token_count == num_tokens: + break + + # Join the selected tokens back into a string + result = " ".join(tokens) + + return result + + # [FEAT][AGENT] class Agent(BaseStructure): """ @@ -256,6 +284,7 @@ class Agent(BaseStructure): planning_prompt: Optional[str] = None, device: str = None, custom_planning_prompt: str = None, + memory_chunk_size: int = 2000, *args, **kwargs, ): @@ -336,6 +365,7 @@ class Agent(BaseStructure): self.custom_planning_prompt = custom_planning_prompt self.rules = rules self.custom_tools_prompt = custom_tools_prompt + self.memory_chunk_size = memory_chunk_size # Name self.name = agent_name @@ -739,21 +769,41 @@ class Agent(BaseStructure): success = False while attempt < self.retry_attempts and not success: try: + if self.long_term_memory is not None: + memory_retrieval = ( + self.long_term_memory_prompt( + task, *args, **kwargs + ) + ) + # print(len(memory_retrieval)) - response_args = ( - (task_prompt, *args) - if img is None - else (task_prompt, img, *args) - ) - response = self.llm(*response_args, **kwargs) + # Merge the task prompt with the memory retrieval + task_prompt = f"{task_prompt} Documents: Available {memory_retrieval}" - # Print - print(response) + response = self.llm( + task_prompt, *args, **kwargs + ) + print(response) - # Add the response to the memory - self.short_memory.add( - role=self.agent_name, content=response - ) + self.short_memory.add( + role=self.agent_name, content=response + ) + + else: + response_args = ( + (task_prompt, *args) + if img is None + else (task_prompt, img, *args) + ) + response = self.llm(*response_args, **kwargs) + + # Print + print(response) + + # Add the response to the memory + self.short_memory.add( + role=self.agent_name, content=response + ) # Check if tools is not None if self.tools is not None: @@ -930,12 +980,16 @@ class Agent(BaseStructure): Returns: str: The agent history prompt """ + # Query the long term memory database ltr = self.long_term_memory.query(query, *args, **kwargs) + ltr = str(ltr) - context = f""" - System: This reminds you of these events from your past: [{ltr}] - """ - return self.short_memory.add(role=self.agent_name, content=context) + # Retrieve only the chunk size of the memory + ltr = retrieve_tokens(ltr, self.memory_chunk_size) + + print(len(ltr)) + # print(f"Long Term Memory Query: {ltr}") + return ltr def add_memory(self, message: str): """Add a memory to the agent @@ -1258,7 +1312,7 @@ class Agent(BaseStructure): "agent_id": str(self.id), "agent_name": self.agent_name, "agent_description": self.agent_description, - "LLM": str(self.get_llm_parameters()), + # "LLM": str(self.get_llm_parameters()), "system_prompt": self.system_prompt, "short_memory": self.short_memory.return_history_as_string(), "loop_interval": self.loop_interval, diff --git a/swarms/structs/agent_network.py b/swarms/structs/agent_network.py new file mode 100644 index 00000000..e69de29b diff --git a/swarms/structs/groupchat.py b/swarms/structs/groupchat.py index dbf4e78f..77a1207e 100644 --- a/swarms/structs/groupchat.py +++ b/swarms/structs/groupchat.py @@ -1,4 +1,3 @@ -from dataclasses import dataclass, field from typing import List from swarms.structs.conversation import Conversation from swarms.utils.loguru_logger import logger @@ -6,36 +5,66 @@ from swarms.structs.agent import Agent from swarms.structs.base_swarm import BaseSwarm -@dataclass class GroupChat(BaseSwarm): - """ - A group chat class that contains a list of agents and the maximum number of rounds. + """Manager class for a group chat. - Args: - agents: List[Agent] - messages: List[Dict] - max_round: int - admin_name: str + This class handles the management of a group chat, including initializing the conversation, + selecting the next speaker, resetting the chat, and executing the chat rounds. - Usage: - >>> from swarms import GroupChat - >>> from swarms.structs.agent import Agent - >>> agents = Agent() + Args: + agents (List[Agent], optional): List of agents participating in the group chat. Defaults to None. + max_rounds (int, optional): Maximum number of chat rounds. Defaults to 10. + admin_name (str, optional): Name of the admin user. Defaults to "Admin". + group_objective (str, optional): Objective of the group chat. Defaults to None. + selector_agent (Agent, optional): Agent responsible for selecting the next speaker. Defaults to None. + rules (str, optional): Rules for the group chat. Defaults to None. + *args: Variable length argument list. + **kwargs: Arbitrary keyword arguments. + + Attributes: + agents (List[Agent]): List of agents participating in the group chat. + max_rounds (int): Maximum number of chat rounds. + admin_name (str): Name of the admin user. + group_objective (str): Objective of the group chat. + selector_agent (Agent): Agent responsible for selecting the next speaker. + messages (Conversation): Conversation object for storing the chat messages. """ - agents: List[Agent] = field(default_factory=list) - max_round: int = 10 - admin_name: str = "Admin" # the name of the admin agent - group_objective: str = field(default_factory=str) - - def __post_init__(self): - self.messages = Conversation( + def __init__( + self, + agents: List[Agent] = None, + max_rounds: int = 10, + admin_name: str = "Admin", + group_objective: str = None, + selector_agent: Agent = None, + rules: str = None, + *args, + **kwargs, + ): + super().__init__() + self.agents = agents + self.max_rounds = max_rounds + self.admin_name = admin_name + self.group_objective = group_objective + self.selector_agent = selector_agent + + # Initialize the conversation + self.message_history = Conversation( system_prompt=self.group_objective, time_enabled=True, user=self.admin_name, + rules=rules, + *args, + **kwargs, ) + # Check to see if the agents is not None: + if agents is None: + raise ValueError( + "Agents may not be empty please try again, add more agents!" + ) + @property def agent_names(self) -> List[str]: """Return the names of the agents in the group chat.""" @@ -44,10 +73,21 @@ class GroupChat(BaseSwarm): def reset(self): """Reset the group chat.""" logger.info("Resetting Groupchat") - self.messages.clear() + self.message_history.clear() def agent_by_name(self, name: str) -> Agent: - """Find an agent whose name is contained within the given 'name' string.""" + """Find an agent whose name is contained within the given 'name' string. + + Args: + name (str): Name string to search for. + + Returns: + Agent: Agent object with a name contained in the given 'name' string. + + Raises: + ValueError: If no agent is found with a name contained in the given 'name' string. + + """ for agent in self.agents: if agent.agent_name in name: return agent @@ -56,7 +96,15 @@ class GroupChat(BaseSwarm): ) def next_agent(self, agent: Agent) -> Agent: - """Return the next agent in the list.""" + """Return the next agent in the list. + + Args: + agent (Agent): Current agent. + + Returns: + Agent: Next agent in the list. + + """ return self.agents[ (self.agent_names.index(agent.agent_name) + 1) % len(self.agents) @@ -64,19 +112,31 @@ class GroupChat(BaseSwarm): def select_speaker_msg(self): """Return the message for selecting the next speaker.""" - return f""" + prompt = f""" You are in a role play game. The following roles are available: {self._participant_roles()}. Read the following conversation. Then select the next role from {self.agent_names} to play. Only return the role. """ + return prompt # @try_except_wrapper - def select_speaker(self, last_speaker: Agent, selector: Agent): - """Select the next speaker.""" + def select_speaker( + self, last_speaker_agent: Agent, selector_agent: Agent + ): + """Select the next speaker. + + Args: + last_speaker_agent (Agent): Last speaker in the conversation. + selector_agent (Agent): Agent responsible for selecting the next speaker. + + Returns: + Agent: Next speaker. + + """ logger.info("Selecting a New Speaker") - selector.system_prompt = self.select_speaker_msg() + selector_agent.system_prompt = self.select_speaker_msg() # Warn if GroupChat is underpopulated, without established changing behavior n_agents = len(self.agent_names) @@ -86,24 +146,27 @@ class GroupChat(BaseSwarm): " Direct communication would be more efficient." ) - self.messages.add( + self.message_history.add( role=self.admin_name, content=f"Read the above conversation. Then select the next most suitable role from {self.agent_names} to play. Only return the role.", ) - name = selector.run(self.messages.return_history_as_string()) + name = selector_agent.run( + self.message_history.return_history_as_string() + ) try: name = self.agent_by_name(name) print(name) return name except ValueError: - return self.next_agent(last_speaker) + return self.next_agent(last_speaker_agent) def _participant_roles(self): """Print the roles of the participants. Returns: - _type_: _description_ + str: Participant roles. + """ return "\n".join( [ @@ -112,53 +175,52 @@ class GroupChat(BaseSwarm): ] ) - -@dataclass -class GroupChatManager: - """ - GroupChatManager - - Args: - groupchat: GroupChat - selector: Agent - - Usage: - >>> from swarms import GroupChatManager - >>> from swarms.structs.agent import Agent - >>> agents = Agent() - - - """ - - groupchat: GroupChat - selector: Agent - - # @try_except_wrapper - def __call__(self, task: str): + def __call__(self, task: str, *args, **kwargs): """Call 'GroupChatManager' instance as a function. Args: - task (str): _description_ + task (str): Task to be performed. Returns: - _type_: _description_ - """ - logger.info( - f"Activating Groupchat with {len(self.groupchat.agents)} Agents" - ) + str: Reply from the last speaker. - self.groupchat.messages.add(self.selector.agent_name, task) - - for i in range(self.groupchat.max_round): - speaker = self.groupchat.select_speaker( - last_speaker=self.selector, selector=self.selector - ) - reply = speaker.run( - self.groupchat.messages.return_history_as_string() + """ + try: + logger.info( + f"Activating Groupchat with {len(self.agents)} Agents" ) - self.groupchat.messages.add(speaker.agent_name, reply) - print(reply) - if i == self.groupchat.max_round - 1: - break - return reply + # Message History + self.message_history.add(self.selector_agent.agent_name, task) + + # Message + for i in range(self.max_rounds): + speaker_agent = self.select_speaker( + last_speaker_agent=self.selector_agent, + selector_agent=self.selector_agent, + ) + + logger.info( + f"Next speaker selected: {speaker_agent.agent_name}" + ) + + # Reply back to the input prompt + reply = speaker_agent.run( + self.message_history.return_history_as_string(), + *args, + **kwargs, + ) + + # Message History + self.message_history.add(speaker_agent.agent_name, reply) + print(reply) + + if i == self.max_rounds - 1: + break + + return reply + except Exception as error: + logger.error( + f"Error detected: {error} Try optimizing the inputs and then submit an issue into the swarms github, so we can help and assist you." + ) + raise error diff --git a/swarms/structs/scp.py b/swarms/structs/scp.py index 2cefdd20..2fda236e 100644 --- a/swarms/structs/scp.py +++ b/swarms/structs/scp.py @@ -14,6 +14,7 @@ from swarms.memory.base_vectordb import BaseVectorDatabase import time from swarms.utils.loguru_logger import logger from pydantic import BaseModel, Field +from typing import Any class SwarmCommunicationProtocol(BaseModel): @@ -31,6 +32,31 @@ class SwarmCommunicationProtocol(BaseModel): class SCP(BaseStructure): + """ + Represents the Swarm Communication Protocol (SCP). + + SCP is responsible for managing agents and their communication within a swarm. + + Args: + agents (List[AgentType]): A list of agents participating in the swarm. + memory_system (BaseVectorDatabase, optional): The memory system used by the agents. Defaults to None. + + Attributes: + agents (List[AgentType]): A list of agents participating in the swarm. + memory_system (BaseVectorDatabase): The memory system used by the agents. + + Methods: + message_log(agent: AgentType, task: str = None, message: str = None) -> str: + Logs a message from an agent and adds it to the memory system. + + run_single_agent(agent: AgentType, task: str, *args, **kwargs) -> Any: + Runs a task for a single agent and logs the output. + + send_message(agent: AgentType, message: str): + Sends a message to an agent and logs it. + + """ + def __init__( self, agents: List[AgentType], @@ -55,7 +81,19 @@ class SCP(BaseStructure): def message_log( self, agent: AgentType, task: str = None, message: str = None - ): + ) -> str: + """ + Logs a message from an agent and adds it to the memory system. + + Args: + agent (AgentType): The agent that generated the message. + task (str, optional): The task associated with the message. Defaults to None. + message (str, optional): The message content. Defaults to None. + + Returns: + str: The JSON-encoded log message. + + """ log = { "agent_name": agent.agent_name, "task": task, @@ -73,7 +111,18 @@ class SCP(BaseStructure): def run_single_agent( self, agent: AgentType, task: str, *args, **kwargs - ): + ) -> Any: + """ + Runs a task for a single agent and logs the output. + + Args: + agent (AgentType): The agent to run the task for. + task (str): The task to be executed. + + Returns: + Any: The output of the task. + + """ # Send the message to the agent output = agent.run(task) @@ -88,4 +137,12 @@ class SCP(BaseStructure): return output def send_message(self, agent: AgentType, message: str): + """ + Sends a message to an agent and logs it. + + Args: + agent (AgentType): The agent to send the message to. + message (str): The message to be sent. + + """ agent.receieve_mesage(self.message_log(agent, message)) diff --git a/swarms/tools/prebuilt/bing_api.py b/swarms/tools/prebuilt/bing_api.py index d00bca98..4a335151 100644 --- a/swarms/tools/prebuilt/bing_api.py +++ b/swarms/tools/prebuilt/bing_api.py @@ -35,8 +35,6 @@ def parse_and_merge_logs(logs: List[Dict[str, str]]) -> str: def fetch_web_articles_bing_api( query: str = None, - subscription_key: str = check_bing_api_key(), - return_str: bool = False, ) -> List[Dict[str, str]]: """ Fetches four articles from Bing Web Search API based on the given query. @@ -48,8 +46,7 @@ def fetch_web_articles_bing_api( Returns: List[Dict[str, str]]: A list of dictionaries containing article details. """ - if subscription_key is None: - subscription_key = check_bing_api_key() + subscription_key = check_bing_api_key() url = "https://api.bing.microsoft.com/v7.0/search" headers = {"Ocp-Apim-Subscription-Key": subscription_key}