@ -7,24 +7,36 @@ The `Conversation` class is a powerful tool for managing and structuring convers
## Table of Contents
## Table of Contents
1. **Class Definition**
1. **Class Definition**
- Overview
- Attributes
- Overview
2. **Methods**
- Attributes
- `__init__(self, time_enabled: bool = False, *args, **kwargs)`
- `add(self, role: str, content: str, *args, **kwargs)`
- Initialization Parameters
- `delete(self, index: str)`
- `update(self, index: str, role, content)`
2. **Core Methods**
- `query(self, index: str)`
- `search(self, keyword: str)`
- Message Management
- `display_conversation(self, detailed: bool = False)`
- `export_conversation(self, filename: str)`
- History Operations
- `import_conversation(self, filename: str)`
- `count_messages_by_role(self)`
- Export/Import
- `return_history_as_string(self)`
- `save_as_json(self, filename: str)`
- Search and Query
- `load_from_json(self, filename: str)`
- `search_keyword_in_conversation(self, keyword: str)`
- Cache Management
- Memory Management
3. **Advanced Features**
- Token Counting
- Memory Providers
- Caching System
- Batch Operations
---
---
@ -36,217 +48,303 @@ The `Conversation` class is designed to manage conversations by keeping track of
#### Attributes
#### Attributes
- `time_enabled (bool)` : A flag indicating whether to enable timestamp recording for messages.
- `id (str)` : Unique identifier for the conversation
- `conversation_history (list)` : A list that stores messages in the conversation.
### 2. Methods - `name (str)` : Name of the conversation
#### `__init__(self, time_enabled: bool = False, *args, **kwargs)` - `system_prompt (Optional[str])` : System prompt for the conversation
- **Description** : Initializes a new Conversation object.
- `time_enabled (bool)` : Flag to enable time tracking for messages
- **Parameters** :
- `time_enabled (bool)` : If `True` , timestamps will be recorded for each message. Default is `False` .
#### `add(self, role: str, content: str, *args, **kwargs)` - `autosave (bool)` : Flag to enable automatic saving
- **Description** : Adds a message to the conversation history.
- `save_filepath (str)` : File path for saving conversation history
- **Parameters** :
- `role (str)` : The role of the speaker (e.g., "user," "assistant").
- `content (str)` : The content of the message.
#### `delete(self, index: str)` - `conversation_history (list)` : List storing conversation messages
- **Description** : Deletes a message from the conversation history.
- `tokenizer (Any)` : Tokenizer for counting tokens
- **Parameters** :
- `index (str)` : The index of the message to delete.
#### `update(self, index: str, role, content)` - `context_length (int)` : Maximum number of tokens allowed
- **Description** : Updates a message in the conversation history.
- `rules (str)` : Rules for the conversation
- **Parameters** :
- `index (str)` : The index of the message to update.
- `role (_type_)` : The new role of the speaker.
- `content (_type_)` : The new content of the message.
#### `query(self, index: str)` - `custom_rules_prompt (str)` : Custom prompt for rules
- **Description** : Retrieves a message from the conversation history.
- `user (str)` : User identifier for messages
- **Parameters** :
- `index (str)` : The index of the message to query.
- **Returns** : The message as a string.
#### `search(self, keyword: str)` - `auto_save (bool)` : Flag for auto-saving
- **Description** : Searches for messages containing a specific keyword in the conversation history.
- `save_as_yaml (bool)` : Flag to save as YAML
- **Parameters** :
- `keyword (str)` : The keyword to search for.
- **Returns** : A list of messages that contain the keyword.
#### `display_conversation(self, detailed: bool = False)`
- **Description** : Displays the conversation history.
- `save_as_json_bool (bool)` : Flag to save as JSON
- **Parameters** :
- `detailed (bool, optional)` : If `True` , provides detailed information about each message. Default is `False` .
#### `export_conversation(self, filename: str)` - `token_count (bool)` : Flag to enable token counting
- **Description** : Exports the conversation history to a text file.
- `cache_enabled (bool)` : Flag to enable prompt caching
- **Parameters** :
- `filename (str)` : The name of the file to export to.
#### `import_conversation(self, filename: str)` - `cache_stats (dict)` : Statistics about cache usage
- **Description** : Imports a conversation history from a text file.
- `provider (Literal["mem0", "in-memory"])` : Memory provider type
- **Parameters** :
- `filename (str)` : The name of the file to import from.
#### `count_messages_by_role(self)`
#### Initialization Parameters
- **Description** : Counts the number of messages by role in the conversation.
```python
- **Returns** : A dictionary containing the count of messages for each role.
conversation = Conversation(
id="unique_id", # Optional: Unique identifier
name="conversation_name", # Optional: Name of conversation
system_prompt="System message", # Optional: Initial system prompt
time_enabled=True, # Optional: Enable timestamps
autosave=True, # Optional: Enable auto-saving
save_filepath="path/to/save.json", # Optional: Save location
tokenizer=your_tokenizer, # Optional: Token counter
context_length=8192, # Optional: Max tokens
rules="conversation rules", # Optional: Rules
custom_rules_prompt="custom", # Optional: Custom rules
user="User:", # Optional: User identifier
auto_save=True, # Optional: Auto-save
save_as_yaml=True, # Optional: Save as YAML
save_as_json_bool=False, # Optional: Save as JSON
token_count=True, # Optional: Count tokens
cache_enabled=True, # Optional: Enable caching
conversations_dir="path/to/dir", # Optional: Cache directory
provider="in-memory" # Optional: Memory provider
)
```
### 2. Core Methods
#### `return_history_as_string(self)`
#### Message Management
- **Description** : Returns the entire conversation history as a single string.
##### `add(role: str, content: Union[str, dict, list], metadata: Optional[dict] = None)`
- **Returns** : The conversation history as a string.
#### `save_as_json(self, filename: str)`
Adds a message to the conversation history.
- **Description** : Saves the conversation history as a JSON file.
```python
- **Parameters** :
# Add a simple text message
- `filename (str)` : The name of the JSON file to save.
conversation.add("user", "Hello, how are you?")
#### `load_from_json(self, filename: str)`
# Add a structured message
conversation.add("assistant", {
"type": "response",
"content": "I'm doing well!"
})
- **Description** : Loads a conversation history from a JSON file.
# Add with metadata
- **Parameters** :
conversation.add("user", "Hello", metadata={"timestamp": "2024-03-20"})
- `filename (str)` : The name of the JSON file to load.
```
#### `search_keyword_in_conversation(self, keyword: str)`
##### `add_multiple_messages(roles: List[str], contents: List[Union[str, dict, list]] )`
- **Description** : Searches for a keyword in the conversation history and returns matching messages.
Adds multiple messages at once.
- **Parameters** :
- `keyword (str)` : The keyword to search for.
- **Returns** : A list of messages containing the keyword.
## Examples
```python
conversation.add_multiple_messages(
roles=["user", "assistant"],
contents=["Hello!", "Hi there!"]
)
```
Here are some usage examples of the `Conversation` class:
##### `add_tool_output_to_agent(role: str, tool_output: dict)`
### Creating a Conversation
Adds a tool output to the conversation.
```python
```python
from swarms.structs import Conversation
conversation.add_tool_output_to_agent(
"tool",
{"name": "calculator", "result": "42"}
)
```
#### History Operations
conv = Conversation()
##### `get_last_message_as_string() -> str`
Returns the last message as a string.
```python
last_message = conversation.get_last_message_as_string()
# Returns: "assistant: Hello there!"
```
```
### Adding Messages
##### `get_final_message() -> str`
Returns the final message from the conversation.
```python
```python
conv.add("user", "Hello, world!")
final_message = conversation.get_final_message( )
conv.add("assistant", "Hello, user!") # Returns: "assistant: Goodbye!"
```
```
### Displaying the Conversation
##### `get_final_message_content() -> str`
Returns just the content of the final message.
```python
```python
conv.display_conversation()
final_content = conversation.get_final_message_content()
# Returns: "Goodbye!"
```
```
### Searching for Messages
##### `return_all_except_first() -> list`
Returns all messages except the first one.
```python
```python
result = conv.search("Hello" )messages = conversation.return_all_except_first( )
```
```
### Exporting and Importing Conversations
##### `return_all_except_first_string() -> str`
Returns all messages except the first one as a string.
```python
```python
conv.export_conversation("conversation.txt")
messages_str = conversation.return_all_except_first_string()
conv.import_conversation("conversation.txt")
```
```
### Counting Messages by Role
#### Export/Import
##### `to_json() -> str`
Converts conversation to JSON string.
```python
json_str = conversation.to_json()
```
##### `to_dict() -> list`
Converts conversation to dictionary.
```python
dict_data = conversation.to_dict()
```
##### `to_yaml() -> str`
Converts conversation to YAML string.
```python
yaml_str = conversation.to_yaml()
```
##### `return_json() -> str`
Returns conversation as formatted JSON string.
```python
```python
counts = conv.count_messages_by_role()
json_str = conversation.return_json ()
```
```
### Loading and Saving as JSON
#### Search and Query
##### `get_visible_messages(agent: "Agent", turn: int) -> List[Dict]`
Gets visible messages for a specific agent and turn.
```python
```python
conv.save_as_json("conversation.json")
visible_msgs = conversation.get_visible_messages(agent, turn=1)
conv.load_from_json("conversation.json")
```
```
Certainly! Let's continue with more examples and additional information about the `Conversation` class.
#### Cache Management
### Querying a Specific Message
##### `get_cache_stats() -> Dict[str, int]`
You can retrieve a specific message from the conversation by its index:
Gets statistics about cache usage.
```python
```python
message = conv.query(0) # Retrieves the first message
stats = conversation.get_cache_stats()
# Returns: {
# "hits": 10,
# "misses": 5,
# "cached_tokens": 1000,
# "total_tokens": 2000,
# "hit_rate": 0.67
# }
```
```
### Updating a Message
#### Memory Management
You can update a message's content or role within the conversation:
##### `clear_memory()`
Clears the conversation memory.
```python
```python
conv.update(0, "user", "Hi there!") # Updates the first message
conversation.clear_memory()
```
```
### Deleting a Message
##### `clear()`
If you want to remove a message from the conversation, you can use the `delete` method: Clears the conversation history.
```python
```python
conv.delete(0) # Deletes the first message
conversation.clear()
```
```
### Counting Messages by Role
### 3. Advanced Features
#### Token Counting
You can count the number of messages by role in the conversation:
The class supports automatic token counting when enabled :
```python
```python
counts = conv.count_messages_by_role()
conversation = Conversation(token_count=True)
# Example result: {'user': 2, 'assistant': 2}
conversation.add("user", "Hello world")
# Token count will be automatically calculated and stored
```
```
### Exporting and Importing as Text
#### Memory Providers
You can export the conversation to a text file and later import it :The class supports different memory providers :
```python
```python
conv.export_conversation("conversation.txt") # Export
# In-memory provider (default)
conv.import_conversation("conversation.txt") # Import
conversation = Conversation(provider="in-memory")
# Mem0 provider
conversation = Conversation(provider="mem0")
```
```
### Exporting and Importing as JSON
#### Caching System
Conversations can also be saved and loaded as JSON files :The caching system can be enabled to improve performance :
```python
```python
conv.save_as_json("conversation.json") # Save as JSON
conversation = Conversation(cache_enabled=True)
conv.load_from_json("conversation.json") # Load from JSON # Messages will be cached for faster retrieval
```
```
### Searching for a Keyword
#### Batch Operations
You can search for messages containing a specific keyword within the conversation :The class supports batch operations for efficiency :
```python
```python
results = conv.search_keyword_in_conversation("Hello")
# Batch add messages
conversation.batch_add([
{"role": "user", "content": "Hello"},
{"role": "assistant", "content": "Hi"}
])
```
```
### Class Methods
#### `load_conversation(name: str, conversations_dir: Optional[str] = None) -> "Conversation"`
Loads a conversation from cache.
These examples demonstrate the versatility of the `Conversation` class in managing and interacting with conversation data. Whether you're building a chatbot, conducting analysis, or simply organizing dialogues, this class offers a robust set of tools to help you accomplish your goals.
```python
conversation = Conversation.load_conversation("my_conversation")
```
#### `list_cached_conversations(conversations_dir: Optional[str] = None) -> List[str]`
Lists all cached conversations.
```python
conversations = Conversation.list_cached_conversations()
```
## Conclusion
## Conclusion
The `Conversation` class is a valuable utility for handling conversation data in Python. With its ability to add, update, delete, search, export, and import messages, you have the flexibility to work with conversations in various ways. Feel free to explore its features and adapt them to your specific projects and applications.
The `Conversation` class provides a comprehensive set of tools for managing conversations in Python applications. With support for multiple memory providers, caching, token counting, and various export formats, it's suitable for a wide range of use cases from simple chat applications to complex AI system s.
If you have any further questions or need additional assistance, please don't hesitate to ask!
For more information or specific use cases, please refer to the examples above or consult the source code.
Kye Gomez
3 weeks ago