LangChain

Repo for maintaining code implementation related to LangChain
Note:-

Install the dependencies from requirements.txt using the below command:-
pip install -r requirements.txt
To know more about requirements.txt visit pip freeze > requirements.txt

To create venv execute python -m venv .venv

Table of Contents

Release Version	Feature Implementation
1.0.0	Ollama Integration Groq Integration GPT4All Integration Prompt Engineering LinkedIn Scraping RAG Model Pinecone as Vector DB
1.0.3	ReAct Agents Tools Pydantic Output Parser Agent Executors
1.0.4	Flask Application Re-Using ReAct Agent API response from Agent Response Agent Executors
1.0.5	LangSmith Tracing
1.0.6	Build ReAct Agent Executor from Scratch Tools defining for Agent ReAct Prompting LLM Reasoning Engine Output Parsing Tool Execution Handling Agent Action Setting up Agent Finish LLM Callback
1.0.7	Uploading and Reading PDF Data Chunking of PDF Data to list of Documents / Text Implementation of both predefined and custom chat prompts Pinecone Vector Store integration FAISS Local Vector Store Integration FAISS Vector saving local copy Retrieval from saved local FAISS Retrieval Chain implementation Implementation of Runnable Passthrough
1.0.8	Beautiful Soup Integration to manually scrape doc URLs Using OpenAIEmbeddings for text-embedding-3-small model Using RecursiveUrlLoader lib from Langchain to read docs from URL Splitting and chunking data Initialising the vector store using split documents and embeddings Chat Retrieval QA Chain Creation Streamlit Chat Components in UI/li> Chat History / Context Addition to ChatLLM Using Rephrase Prompt with History Aware Retriever to hold Chat History Using FireCrawl to scrape and crawl websites and index them to Pinecone index
1.0.9	Understanding of Langchain Experimental Tools PKG Using the PythonREPL tool from Langchain Using the tool to write Python code Saving the reponse generated from the code Creating a CSV Agent to answer questions related to CSV file Creating a router agent to switch out to different agents based on the input prompt Create a Tool Calling Agent Bind Tools directly to a LLM Create Tool Calling Agent on 2 LLM Providers - OpenAI and Groq
1.1.0	Dynamic record config generation using LLM Ability to change column data type, remove and associate reference value Sanitize and convert raw order data to DataFrame Read raw order data from a string and load it into a pandas DataFrame. Strip leading/trailing whitespace from column names. Ensure data consistency and handle missing/malformed data. Convert column data types to the appropriate format. Generate and save config files in json Read config files and apply them to raw order data Generate and save generated dataframe into CSV file Generate Order Line Item Details Adding option for the service with added validation of each columns

LangSmith Tracing


Agent Workflow