From the uDemy course on LLM engineering.
https://www.udemy.com/course/llm-engineering-master-ai-and-large-language-models
You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
224 lines
6.6 KiB
224 lines
6.6 KiB
{ |
|
"cells": [ |
|
{ |
|
"cell_type": "code", |
|
"execution_count": null, |
|
"metadata": {}, |
|
"outputs": [], |
|
"source": [ |
|
"# # Document loading, retrieval methods and text splitting\n", |
|
"# !pip install -qU langchain langchain_community\n", |
|
"\n", |
|
"# # Local vector store via Chroma\n", |
|
"# !pip install -qU langchain_chroma\n", |
|
"\n", |
|
"# # Local inference and embeddings via Ollama\n", |
|
"# !pip install -qU langchain_ollama\n", |
|
"\n", |
|
"# # Web Loader\n", |
|
"# !pip install -qU beautifulsoup4\n", |
|
"\n", |
|
"# # Pull the model first\n", |
|
"# !ollama pull nomic-embed-text\n", |
|
"\n", |
|
"# !pip install -qU pypdf" |
|
] |
|
}, |
|
{ |
|
"cell_type": "code", |
|
"execution_count": null, |
|
"metadata": {}, |
|
"outputs": [], |
|
"source": [ |
|
"#Imports\n", |
|
"import os\n", |
|
"import glob\n", |
|
"from dotenv import load_dotenv\n", |
|
"import gradio as gr\n", |
|
"from langchain_community.document_loaders import PyPDFLoader, DirectoryLoader\n", |
|
"from langchain_text_splitters import CharacterTextSplitter, RecursiveCharacterTextSplitter\n", |
|
"from langchain_chroma import Chroma\n", |
|
"from langchain_ollama import OllamaEmbeddings\n", |
|
"from langchain_ollama import ChatOllama\n", |
|
"from langchain_core.output_parsers import StrOutputParser\n", |
|
"from langchain_core.prompts import ChatPromptTemplate\n", |
|
"from langchain_core.runnables import RunnablePassthrough" |
|
] |
|
}, |
|
{ |
|
"cell_type": "code", |
|
"execution_count": null, |
|
"metadata": {}, |
|
"outputs": [], |
|
"source": [ |
|
"# Read in documents using LangChain's loaders\n", |
|
"# Take everything in all the sub-folders of our knowledgebase\n", |
|
"\n", |
|
"folders = glob.glob(\"Manuals/*\")\n", |
|
"\n", |
|
"def add_metadata(doc, doc_type):\n", |
|
" doc.metadata[\"doc_type\"] = doc_type\n", |
|
" return doc\n", |
|
"\n", |
|
"documents = []\n", |
|
"for folder in folders:\n", |
|
" doc_type = os.path.basename(folder)\n", |
|
" loader = DirectoryLoader(folder, glob=\"**/*.pdf\", loader_cls=PyPDFLoader)\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)}\")" |
|
] |
|
}, |
|
{ |
|
"cell_type": "code", |
|
"execution_count": null, |
|
"metadata": {}, |
|
"outputs": [], |
|
"source": [ |
|
"# Put the chunks of data into a Vector Store that associates a Vector Embedding with each chunk\n", |
|
"# Chroma is a popular open source Vector Database based on SQLLite\n", |
|
"DB_NAME = \"vector_db\"\n", |
|
"\n", |
|
"embeddings = OllamaEmbeddings(model=\"nomic-embed-text\")\n", |
|
"\n", |
|
"# Delete if already exists\n", |
|
"\n", |
|
"if os.path.exists(DB_NAME):\n", |
|
" Chroma(persist_directory=DB_NAME, embedding_function=embeddings).delete_collection()\n", |
|
"\n", |
|
"# Create vectorstore\n", |
|
"\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, |
|
"metadata": {}, |
|
"outputs": [], |
|
"source": [ |
|
"#run a quick test - should return a list of documents = 4\n", |
|
"question = \"What kind of grill is the Spirt II?\"\n", |
|
"docs = vectorstore.similarity_search(question)\n", |
|
"len(docs)" |
|
] |
|
}, |
|
{ |
|
"cell_type": "code", |
|
"execution_count": null, |
|
"metadata": {}, |
|
"outputs": [], |
|
"source": [ |
|
"docs[0]" |
|
] |
|
}, |
|
{ |
|
"cell_type": "code", |
|
"execution_count": null, |
|
"metadata": {}, |
|
"outputs": [], |
|
"source": [ |
|
"# create a new Chat with Ollama\n", |
|
"from langchain.memory import ConversationBufferMemory\n", |
|
"from langchain.chains import ConversationalRetrievalChain\n", |
|
"MODEL = \"llama3.2:latest\"\n", |
|
"llm = ChatOllama(temperature=0.7, model=MODEL)\n", |
|
"\n", |
|
"# set up the conversation memory for the chat\n", |
|
"memory = ConversationBufferMemory(memory_key='chat_history', return_messages=True)\n", |
|
"\n", |
|
"# the retriever is an abstraction over the VectorStore that will be used during RAG\n", |
|
"retriever = vectorstore.as_retriever()\n", |
|
"\n", |
|
"# putting it together: set up the conversation chain with the GPT 3.5 LLM, the vector store and memory\n", |
|
"conversation_chain = ConversationalRetrievalChain.from_llm(llm=llm, retriever=retriever, memory=memory)" |
|
] |
|
}, |
|
{ |
|
"cell_type": "code", |
|
"execution_count": null, |
|
"metadata": {}, |
|
"outputs": [], |
|
"source": [ |
|
"# Let's try a simple question\n", |
|
"\n", |
|
"query = \"How do I change the water bottle ?\"\n", |
|
"result = conversation_chain.invoke({\"question\": query})\n", |
|
"print(result[\"answer\"])" |
|
] |
|
}, |
|
{ |
|
"cell_type": "code", |
|
"execution_count": 15, |
|
"metadata": {}, |
|
"outputs": [], |
|
"source": [ |
|
"# set up a new conversation memory for the chat\n", |
|
"memory = ConversationBufferMemory(memory_key='chat_history', return_messages=True)\n", |
|
"\n", |
|
"# putting it together: set up the conversation chain with the LLM, the vector store and memory\n", |
|
"conversation_chain = ConversationalRetrievalChain.from_llm(llm=llm, retriever=retriever, memory=memory)" |
|
] |
|
}, |
|
{ |
|
"cell_type": "code", |
|
"execution_count": 16, |
|
"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": "markdown", |
|
"metadata": {}, |
|
"source": [ |
|
"## Now we will bring this up in Gradio using the Chat interface -\n", |
|
"\n", |
|
"A quick and easy way to prototype a chat with an LLM" |
|
] |
|
}, |
|
{ |
|
"cell_type": "code", |
|
"execution_count": null, |
|
"metadata": {}, |
|
"outputs": [], |
|
"source": [ |
|
"# And in Gradio:\n", |
|
"\n", |
|
"view = gr.ChatInterface(chat, type=\"messages\").launch(inbrowser=True)" |
|
] |
|
} |
|
], |
|
"metadata": { |
|
"kernelspec": { |
|
"display_name": "Python 3 (ipykernel)", |
|
"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": "3.11.11" |
|
} |
|
}, |
|
"nbformat": 4, |
|
"nbformat_minor": 4 |
|
}
|
|
|