How uditkumar08/final_assessment_gen works
This is a PDF-grounded question-answering application. A user uploads a PDF through the Streamlit UI; the backend extracts page text, chunks it, embeds and indexes it in Qdrant, then answers document-specific questions with retrieved passages and page references. The graph also preserves the separately exposed extraction, chunking, and indexing endpoints and the documented Prometheus metrics capability. Some service implementations and application wiring were not sampled, so documented workflow edges cite the README; uncertain links are omitted.
Main entry points: indexing_service.py Indexing service · rag_service.py RAG answering · ui_app.py Streamlit UI · documents.py Upload endpoint
Components 17
User interface
- Streamlit UI
enterprise-ai-knowledge-assistant/frontend/ui_app.py
Backend API
- FastAPI app
enterprise-ai-knowledge-assistant/backend/app/main.py - Upload endpoint
enterprise-ai-knowledge-assistant/backend/app/api/documents.py - Extraction endpoint
enterprise-ai-knowledge-assistant/backend/app/api/extraction.py - Chunking endpoint
enterprise-ai-knowledge-assistant/backend/app/api/chunking.py - Indexing endpoint
enterprise-ai-knowledge-assistant/backend/app/api/indexing.py - Chat endpoint
enterprise-ai-knowledge-assistant/backend/app/api/chat.py
Document preparation
- PDF extraction
enterprise-ai-knowledge-assistant/backend/app/services/pdf_service.py - Document chunking
enterprise-ai-knowledge-assistant/backend/app/services/document_chunking.py - Text splitter
enterprise-ai-knowledge-assistant/backend/app/services/chunk_service.py - Indexing service
enterprise-ai-knowledge-assistant/backend/app/services/indexing_service.py - Embedding model
enterprise-ai-knowledge-assistant/backend/app/services/embedding_service.py
Retrieval and operations
- RAG answering
enterprise-ai-knowledge-assistant/backend/app/services/rag_service.py - Qdrant vectors
enterprise-ai-knowledge-assistant/backend/app/services/qdrant_service.py - Prometheus metrics
enterprise-ai-knowledge-assistant/monitoring/prometheus.yml
Other components
- PDF user
- Groq language model
Connections 17
- PDF user → Streamlit UI: uploads and asks
README.md - Streamlit UI → Upload endpoint: uploads PDF
enterprise-ai-knowledge-assistant/frontend/ui_app.py - Streamlit UI → Chat endpoint: sends question
enterprise-ai-knowledge-assistant/frontend/ui_app.py - Upload endpoint → PDF extraction: extracts pages
enterprise-ai-knowledge-assistant/backend/app/api/documents.py - Upload endpoint → Indexing service: starts indexing
enterprise-ai-knowledge-assistant/backend/app/api/documents.py - Extraction endpoint → PDF extraction: extracts PDF
enterprise-ai-knowledge-assistant/backend/app/api/extraction.py - Chunking endpoint → Document chunking: chunks document
enterprise-ai-knowledge-assistant/backend/app/api/chunking.py - Document chunking → Text splitter: splits text
enterprise-ai-knowledge-assistant/backend/app/services/document_chunking.py - Indexing service → Document chunking: creates chunks
README.md - Indexing service → Embedding model: embeds chunks
README.md - Indexing service → Qdrant vectors: stores vectors
README.md - Indexing endpoint → Indexing service: indexes document
- Chat endpoint → RAG answering: answers question
enterprise-ai-knowledge-assistant/backend/app/api/chat.py - RAG answering → Embedding model: embeds query
README.md - RAG answering → Qdrant vectors: retrieves passages
README.md - RAG answering → Groq language model: generates answer
README.md - Prometheus metrics → FastAPI app: scrapes metrics
README.md