From b439fa0942e79913e7b65619af0fc0f83eba88e5 Mon Sep 17 00:00:00 2001
From: Richard Anthony Hein <richard.hein@gmail.com>
Date: Thu, 22 Aug 2024 14:33:43 +0000
Subject: [PATCH] updated readme, fix paths for imports after moving to
 playground demos

---
 playground/demos/chatbot/README.md               | 13 ++++++++++++-
 playground/demos/chatbot/server/__init__.py      |  2 +-
 playground/demos/chatbot/server/requirements.txt |  3 +++
 playground/demos/chatbot/server/server.py        |  6 +++---
 playground/demos/chatbot/server/vector_store.py  |  2 +-
 5 files changed, 20 insertions(+), 6 deletions(-)
 create mode 100644 playground/demos/chatbot/server/requirements.txt

diff --git a/playground/demos/chatbot/README.md b/playground/demos/chatbot/README.md
index 01f8611d..af25f803 100644
--- a/playground/demos/chatbot/README.md
+++ b/playground/demos/chatbot/README.md
@@ -18,7 +18,18 @@ In theory, any OpenAI compatible LLM endpoint is supported via the OpenAIChatLLM
 
 ### Quickstart
 
-* Start vLLM using Docker container by running the [dockerRunVllm](./server/dockerRunVllm.sh).  Adjust the script to select your desired model and set the HUGGING_FACE_HUB_TOKEN.
+* Start vLLM using GPUs using Docker container by running the [dockerRunVllm](./server/dockerRunVllm.sh).  Adjust the script to select your desired model and set the HUGGING_FACE_HUB_TOKEN.  
+
+** For CPU support (not recommended for vLLM), build and run it in docker using this [Dockerfile](./Dockerfile).
+```bash
+cd <root>/swarms/playground/demos/chatbot
+docker build -t llm-serving:vllm-cpu -f ~/vllm/Dockerfile.cpu .
+docker run --rm --env "HF_TOKEN=<your hugging face token>" \
+  --ipc=host \
+  -p 8000:8000 \
+  llm-serving:vllm-cpu \
+  --model NousResearch/Hermes-3-Llama-3.1-8B
+```
 
 * Start the Chatbot API Server with the following shell command:
 
diff --git a/playground/demos/chatbot/server/__init__.py b/playground/demos/chatbot/server/__init__.py
index 2b7423ff..3d7154bd 100644
--- a/playground/demos/chatbot/server/__init__.py
+++ b/playground/demos/chatbot/server/__init__.py
@@ -1,4 +1,4 @@
-from swarms.server.vector_store import VectorStorage
+from playground.demos.chatbot.server.vector_store import VectorStorage
 
 __all__ = [
     "VectorStorage",
diff --git a/playground/demos/chatbot/server/requirements.txt b/playground/demos/chatbot/server/requirements.txt
new file mode 100644
index 00000000..e96ae42f
--- /dev/null
+++ b/playground/demos/chatbot/server/requirements.txt
@@ -0,0 +1,3 @@
+vllm
+msgspc
+librosa
\ No newline at end of file
diff --git a/playground/demos/chatbot/server/server.py b/playground/demos/chatbot/server/server.py
index dcf7f9c3..7773fbf7 100644
--- a/playground/demos/chatbot/server/server.py
+++ b/playground/demos/chatbot/server/server.py
@@ -27,9 +27,9 @@ from swarms.prompts.conversational_RAG import (
     E_SYS,
     QA_PROMPT_TEMPLATE_STR,
 )
-from swarms.server.responses import StreamingResponse
-from swarms.server.server_models import ChatRequest
-from swarms.server.vector_store import VectorStorage
+from playground.demos.chatbot.server.responses import StreamingResponse
+from playground.demos.chatbot.server.server_models import ChatRequest
+from playground.demos.chatbot.server.vector_store import VectorStorage
 from swarms.models.popular_llms import OpenAIChatLLM
 
 # Explicitly specify the path to the .env file
diff --git a/playground/demos/chatbot/server/vector_store.py b/playground/demos/chatbot/server/vector_store.py
index b8fa853d..a3fcde55 100644
--- a/playground/demos/chatbot/server/vector_store.py
+++ b/playground/demos/chatbot/server/vector_store.py
@@ -15,7 +15,7 @@ from langchain.storage import LocalFileStore
 from langchain.text_splitter import RecursiveCharacterTextSplitter
 from langchain_chroma import Chroma
 
-from swarms.server.async_parent_document_retriever import \
+from playground.demos.chatbot.server.async_parent_document_retriever import \
     AsyncParentDocumentRetriever
 
 STORE_TYPE = "local"  # "redis" or "local"