How pooja-314/enterprise-intelligence-platform works
Enterprise Intelligence Platform provides a Streamlit dashboard over a FastAPI document-intelligence backend. Users authenticate, upload documents for parsing, chunking and indexing, then ask questions through conversational routing, hybrid retrieval and grounded answer generation; multi-step agent tools and claim verification extend that workflow. PostgreSQL/pgvector stores users, documents, chunks and conversations. The graph preserves documented evaluation and telemetry capabilities; unsampled route wiring is included only where the README or source index supports it.
Main entry points: api_client.py HTTP API Client · generator.py RAG Answer Generator · retrieval.py Hybrid Retrieval · pipeline.py Ingestion Pipeline
Components 25
Dashboard
- Streamlit Dashboard
frontend/app.py - Product Views
frontend/pages_views - HTTP API Client
frontend/api_client.py
Backend API
- Authentication API
app/api/auth.py - Document API
app/api/documents.py - RAG API
app/api/rag.py - Conversation API
app/api/conversations.py - Agent API
app/api/agent.py
Document Intelligence
- Ingestion Pipeline
app/ingestion/pipeline.py - Document Parser
app/ingestion/parser.py - Recursive Chunker
app/ingestion/chunker.py - Embedding Service
app/rag/embeddings.py - Hybrid Retrieval
app/rag/retrieval.py - RAG Answer Generator
app/rag/generator.py - Claim Verifier
app/rag/verifier.py
Agents and Memory
- Intent Router
app/agent/query_router.py - Agent Orchestrator
app/agent/orchestrator.py - Agent Tools
app/agent/tools - Conversation Memory
app/memory/manager.py
Platform Services
- LLM Fallback Router
app/llm/fallback_router.py - LLM Providers
app/llm - Relational and Vector Store
app/db/models.py - RAG Evaluation
app/evaluation/runner.py - Usage Telemetry
app/core/telemetry.py
Other components
- Enterprise User
Connections 36
- Enterprise User → Streamlit Dashboard: uses
README.md - Streamlit Dashboard → Product Views: renders
frontend/app.py - Product Views → HTTP API Client: requests data
frontend/pages_views/chat_view.py - HTTP API Client → Authentication API: authenticates
frontend/api_client.py - HTTP API Client → Document API: manages documents
README.md - HTTP API Client → RAG API: submits queries
frontend/api_client.py - HTTP API Client → Conversation API: uses conversations
- HTTP API Client → Agent API: uses agent
- Document API → Ingestion Pipeline: processes uploads
README.md - Ingestion Pipeline → Document Parser: parses files
app/ingestion/pipeline.py - Ingestion Pipeline → Recursive Chunker: chunks text
app/ingestion/pipeline.py - Ingestion Pipeline → Embedding Service: embeds chunks
app/ingestion/pipeline.py - Ingestion Pipeline → Relational and Vector Store: persists documents
README.md - RAG API → Hybrid Retrieval: retrieves evidence
app/api/rag.py - RAG API → RAG Answer Generator: generates answers
app/api/rag.py - RAG API → Claim Verifier: verifies claims
app/api/rag.py - Hybrid Retrieval → Embedding Service: embeds queries
app/rag/retrieval.py - Hybrid Retrieval → Relational and Vector Store: searches chunks
app/rag/retrieval.py - RAG Answer Generator → Hybrid Retrieval: gets evidence
app/rag/generator.py - RAG Answer Generator → Claim Verifier: checks response
app/rag/generator.py - RAG Answer Generator → LLM Fallback Router: generates text
app/rag/generator.py - Conversation API → Intent Router: routes messages
app/api/conversations.py - Intent Router → Agent Orchestrator: dispatches agent work
app/agent/query_router.py - Intent Router → RAG Answer Generator: answers factual queries
app/agent/query_router.py - Agent Orchestrator → Agent Tools: runs selected tool
app/agent/orchestrator.py - Agent Orchestrator → LLM Fallback Router: reasons with
app/agent/orchestrator.py - Agent Orchestrator → Conversation Memory: uses context
app/agent/orchestrator.py - Agent Tools → Hybrid Retrieval: searches knowledge
app/agent/tools/kb_search.py - Agent Tools → LLM Fallback Router: analyzes and reports
app/agent/tools/doc_analysis.py - Conversation API → Conversation Memory: persists dialogue
app/api/conversations.py - Conversation Memory → Relational and Vector Store: stores conversations
app/memory/manager.py - LLM Fallback Router → LLM Providers: fails over
app/llm/fallback_router.py - RAG Evaluation → Hybrid Retrieval: evaluates retrieval
app/evaluation/runner.py - RAG Evaluation → RAG Answer Generator: evaluates answers
app/evaluation/runner.py - RAG Evaluation → Claim Verifier: evaluates grounding
app/evaluation/runner.py - LLM Providers → Usage Telemetry: records usage
app/llm/groq_provider.py