ai-mentorship/debate-match-rag

How ai-mentorship/debate-match-rag works

DebateMatch is an AI-assisted debate transcript explorer and question-answering system. Its principal workflow is a user submitting a question to the backend, which retrieves relevant debate passages and returns a grounded answer; a separate claim-checking route evaluates claims against external evidence. Transcript ingestion and indexing support these features, while the React frontend also provides transcript browsing and informational pages. The excerpts establish the backend’s retrieval and fact-checking calls, but do not show frontend-to-backend request wiring or the full ingestion/indexing call sequence, so those relationships are omitted or marked unverified.

Main entry points: main.py Flask application · App.jsx Page router · retriever.py Passage retrieval · AI_FactChecker.py Claim fact checker

Components 21

User interface

Page router src/frontend/src/App.jsx
Analyzer src/frontend/src/pages/Analyzer.jsx
Transcript browser src/frontend/src/pages/Transcripts.jsx
Home page src/frontend/src/pages/Home.jsx
Mission page src/frontend/src/pages/Missions.jsx
Team page src/frontend/src/pages/Team.jsx
Navigation header src/frontend/src/components/Header.jsx

Backend routes

Flask application src/main.py

Transcript ingestion

Transcript cleaner src/backend/preprocessing/preprocess_script.py
MongoDB collections src/backend/database/connection.py
Transcript loader src/backend/database/insert.py
FAISS index builder src/backend/embeddings_faiss/build_index.py
Incremental indexing src/backend/embeddings_faiss/incremental_index.py

Retrieval and analysis

Passage retrieval src/backend/retriever/retriever.py
RAG answer pipeline src/backend/qa_pipeline/QA_pipeline.py
Claim fact checker src/backend/fact_checker_prototype/AI_FactChecker.py
GPT client src/backend/core_llm/gpt5_nano.py

Other components

User
OpenAI services
Wikipedia API
News API

Connections 24

User → Page router: uses
Page router → Navigation header: renders src/frontend/src/App.jsx
Page router → Home page: routes src/frontend/src/App.jsx
Page router → Analyzer: routes src/frontend/src/App.jsx
Page router → Transcript browser: routes src/frontend/src/App.jsx
Page router → Mission page: routes src/frontend/src/App.jsx
Page router → Team page: routes src/frontend/src/App.jsx
Transcript browser → User: shows transcripts src/frontend/src/pages/Transcripts.jsx
Analyzer → Flask application: submits queries
Flask application → Passage retrieval: retrieves passages src/main.py
Flask application → Claim fact checker: checks claims src/main.py
Flask application → RAG answer pipeline: generates answers src/main.py
Flask application → Transcript cleaner: processes uploads src/main.py
Flask application → Transcript loader: loads transcripts src/main.py
Flask application → FAISS index builder: builds index src/main.py
Flask application → Incremental indexing: updates index src/main.py
Transcript loader → MongoDB collections: writes records src/backend/database/insert.py
MongoDB collections → Transcript loader: connects through src/backend/database/insert.py
Passage retrieval → OpenAI services: embeds queries src/backend/retriever/retriever.py
Passage retrieval → FAISS index builder: searches index
RAG answer pipeline → OpenAI services: embeds and generates src/backend/qa_pipeline/QA_pipeline.py
Flask application → GPT client: uses client src/main.py
Claim fact checker → Wikipedia API: searches evidence src/backend/fact_checker_prototype/AI_FactChecker.py
Claim fact checker → News API: searches evidence

More JavaScript diagrams: adrianhajdin/project_mern_memories · chalarangelo/30-seconds-of-react · scottschiller/soundmanager2 · libnoname/noname · gnmyt/nexterm

Drawn by GitDiagram on October 1, 2026 from the repository's file tree, README and source files. Read as Markdown ·