Skip to content

README.md

Latest commit

 

History

History
42 lines (29 loc) · 966 Bytes

README.md

File metadata and controls

42 lines (29 loc) · 966 Bytes
Raw

LongTracer Demo Application

This directory contains a complete RAG pipeline demo using LongTracer with ChromaDB + Ollama. It is NOT part of the published longtracer PyPI package.

Prerequisites

  • Python 3.10+
  • Ollama running locally with llama3.1 model
  • PDF documents to ingest

Setup

# From the repo root
pip install -e ".[all]"
pip install pymupdf langchain-ollama langchain-text-splitters langchain-chroma langchain-huggingface

# Copy and edit the env file
cp examples/.env.example examples/.env

# Pull the Ollama model
ollama pull llama3.1

Usage

cd examples

# Ingest PDFs
python demo_pipeline.py --ingest './your-pdfs/'

# Query with verification
python demo_pipeline.py --query 'What is the main contribution?'

What this demo shows

  1. PDF ingestion → ChromaDB vector store
  2. RAG retrieval + Ollama LLM generation
  3. LongTracer parallel verification pipeline
  4. Full trace report with claim-level results