Week5 GenAi Andela bootcamp project

week5/community-contributions/kachaje-andela-genai-bootcamp/week5-genai-bootcamp.ipynb

@@ -0,0 +1,204 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "b4992675",
+   "metadata": {},
+   "source": [
+    "# Personal Knowledge Worker Using RAG\n",
+    "\n",
+    "Tool for querying personal file storage using RAG. Working with local Llama instance"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "2680f9c0",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "! pip -q install langchain langchain-community sentence-transformers ollama"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "47b6caab",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# imports\n",
+    "\n",
+    "import os\n",
+    "from langchain_community.document_loaders import DirectoryLoader\n",
+    "from langchain.text_splitter import CharacterTextSplitter\n",
+    "from langchain.document_loaders import TextLoader\n",
+    "import glob\n",
+    "from langchain_community.llms import Ollama\n",
+    "from langchain.embeddings import OllamaEmbeddings\n",
+    "from langchain_chroma import Chroma\n",
+    "from langchain.memory import ConversationBufferMemory\n",
+    "from langchain.chains import ConversationalRetrievalChain\n",
+    "import gradio as gr\n",
+    "from langchain_core.callbacks import StdOutCallbackHandler\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "d42c7523",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Using mistral model to run locally\n",
+    "\n",
+    "MODEL = \"mistral\"\n",
+    "db_name = \"vector_db\""
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "a3ac38f7",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "def add_metadata(doc, doc_type):\n",
+    "    doc.metadata[\"doc_type\"] = doc_type\n",
+    "    return doc\n",
+    "\n",
+    "\n",
+    "# Using reference data as default to avoid repeating the same steps. This can be substituted with any other data later dynamically\n",
+    "def load_documents(folders = glob.glob(\"../../knowledge-base/*\")):\n",
+    "    text_loader_kwargs = {'encoding': 'utf-8'}\n",
+    "\n",
+    "    documents = []\n",
+    "    for folder in folders:\n",
+    "        doc_type = os.path.basename(folder)\n",
+    "        loader = DirectoryLoader(folder, glob=\"**/*.md\", loader_cls=TextLoader, loader_kwargs=text_loader_kwargs)\n",
+    "        folder_docs = loader.load()\n",
+    "        documents.extend([add_metadata(doc, doc_type) for doc in folder_docs])\n",
+    "\n",
+    "    text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=200)\n",
+    "    chunks = text_splitter.split_documents(documents)\n",
+    "\n",
+    "    print(f\"Total number of chunks: {len(chunks)}\")\n",
+    "    print(f\"Document types found: {set(doc.metadata['doc_type'] for doc in documents)}\")\n",
+    "\n",
+    "    return documents, chunks\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "75a53301",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "embeddings = OllamaEmbeddings(model=MODEL)\n",
+    "\n",
+    "documents, chunks = load_documents()\n",
+    "\n",
+    "# Use existing vectorstore if it exists\n",
+    "if os.path.exists(db_name):\n",
+    "    vectorstore = Chroma(persist_directory=db_name, embedding_function=embeddings)\n",
+    "else:\n",
+    "    # Create a new vectorstore\n",
+    "    vectorstore = Chroma.from_documents(documents=chunks, embedding=embeddings, persist_directory=db_name)\n",
+    "print(f\"Vectorstore created with {vectorstore._collection.count()} documents\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "2e906f8a",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "llm = Ollama(model=MODEL)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "7f1ca473",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "retriever = vectorstore.as_retriever()\n",
+    "\n",
+    "memory = ConversationBufferMemory(memory_key='chat_history', return_messages=True)\n",
+    "\n",
+    "conversation_chain = ConversationalRetrievalChain.from_llm(llm=llm, retriever=retriever, memory=memory)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "34d5a20a",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "question = \"Please explain what Insurellm is in a couple of sentences\"\n",
+    "result = conversation_chain.invoke(question)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "e0ad3702",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "print(result[\"answer\"])"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "652d8420",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Wrapping that in a function\n",
+    "\n",
+    "def chat(question, history):\n",
+    "    result = conversation_chain.invoke({\"question\": question})\n",
+    "    return result[\"answer\"]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "73bd1d46",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# And in Gradio:\n",
+    "\n",
+    "view = gr.ChatInterface(chat, type=\"messages\").launch(inbrowser=True)"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": ".venv",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.12.10"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}