Welcome to the LangChain Guide! This repository serves as a comprehensive resource for anyone looking to dive into the world of LangChain, offering a step-by-step journey from understanding the basics to crafting impactful projects.
This Python script demonstrates how to use LangChain for chat completion. By leveraging the OpenAI API, it allows users to input a section of text and receive a completed version based on the context and language patterns.
Usage:
python chat_completion.py
The review_sentiment_analysis.py file contains code for sentiment analysis on dataframes from a CSV file. It utilizes LangChain to evaluate and classify the sentiment of each review as either negative or positive.
Usage:
python review_sentiment_analysis.py
The pdf_embeddings.py script demonstrates how to use embeddings in LangChain with PDF files. It first loads the PDF using PyMuPDFLoader, uses CharacterTextSplitter to make chunk sizes, and employs the embedding model intfloat provided by HuggingFaceEmbeddings. It uses FAISS as a vectordb and finally performs a similarity search based on the query given by the user.
Usage:
python pdf_embeddings.py
The txt_embeddings.py script performs the same functionality as pdf_embeddings.py but is tailored for text (txt) files.
Usage:
python txt_embeddings.py
This Python script demonstrates how to process PDF files for Question Answering (QA) using LangChain. It loads a PDF document using PyMuPDFLoader, splits the text into chunks using CharacterTextSplitter, utilizes embeddings from HuggingFaceEmbeddings (intfloat model), creates vectors with FAISS, and executes a similarity search to retrieve answers based on the user query.
Usage:
python qa_pdf.py
The qa_txt.py script performs similar functionality as qa_pdf.py, but it's tailored for text (txt) files. It loads a text file using TextLoader, splits the text into chunks using CharacterTextSplitter, applies embeddings from HuggingFaceEmbeddings (intfloat model), generates vectors using FAISS, and executes a similarity search to retrieve answers based on the user query.
Usage:
python qa_txt.py
The qd_pdf.py script illustrates the utilization of Qdrant, a cloud-based vector store platform, for processing PDF documents. It employs LangChain to load a PDF file through PyMuPDFLoader, segment the text into chunks using CharacterTextSplitter, employ embeddings from HuggingFaceEmbeddings (sentence-transformers/all-mpnet-base-v2 model), create vectors in Qdrant using Qdrant's client, and execute a similarity search in the collection based on the provided query.
Usage:
python qd_pdf.py
These comments help users understand the functionalities and usage instructions for each script, enhancing the clarity and usability of the provided code files.
Requirements:
- Dependencies listed in
requirements.txt
- Clone this repository to your local machine:
git clone https://github.com/your-username/langchain-guide.git
cd langchain-guide- Install the required dependencies:
pip install -r requirements.txt- Run the scripts and explore LangChain functionalities:
python chat_completion.py
python review_sentiment_analysis.pyFeel free to explore, experiment, and contribute to the LangChain Guide! If you have any questions or suggestions, please open an issue or submit a pull request.
Happy coding!
Make sure to replace "your-username" with your actual GitHub username. Additionally, if there are more specific instructions or details for using these scripts, you can add them to the README file.