document-parser
An end-to-end RAG pipeline for .docx files — chunk, embed, store, and query
document-parser is a small end-to-end RAG pipeline built around .docx files.
- Splits a
.docxdocument into chunks using regex + LangChain - Embeds chunks and stores them in Pinecone
- On query: embeds the query, runs similarity search in Pinecone, then feeds retrieved chunks to an OpenAI LLM to generate an answer
- Run via CLI (
python doc_parser.py --filename yourfile.docx) or an included walkthrough notebook
Early/exploratory project — handles .docx input specifically, not arbitrary document formats.