saicharan0811/schemeiq-plus

How saicharan0811/schemeiq-plus works

SchemeIQ+ helps people in Telangana discover welfare schemes, assess eligibility, and ask questions grounded in official sources. The main journey runs from a citizen through the web interface and REST API into profile validation, deterministic rule assessment, and ranked recommendations; a separate question-answering route retrieves source material and produces grounded answers. Document ingestion and RAG indexing support that knowledge workflow. The README describes multilingual support and official-source grounding. Some routes and internal wiring are unsampled, so the graph omits uncertain links rather than inferring them.

Main entry points: ingest.py RAG indexing · App.jsx Page router · service.py Eligibility service · rag_service.py RAG orchestrator

Components 27

Web interface

Page router frontend/src/App.jsx
Recommendations frontend/src/pages/RecommendPage.jsx
Scheme details frontend/src/pages/SchemeDetailPage.jsx
Ask a question frontend/src/pages/AskPage.jsx
API client frontend/src/api.js

API layer

REST API src/api/app.py

Eligibility and recommendations

Eligibility service src/eligibility/service.py
Profile handling src/eligibility/profile.py
Scheme ranking src/eligibility/recommender.py
Eligibility evaluator src/eligibility/evaluator.py
Rule engine src/eligibility/rule_engine.py
Explanations src/eligibility/explanation.py

Grounded Q&A

RAG orchestrator src/rag/rag_service.py
Scheme detection src/rag/query_detector.py
Hybrid retrieval src/rag/hybrid_retriever.py
Keyword retrieval src/rag/keyword_retriever.py
Vector index src/rag/vector_store.py
Answer generation src/rag/answer_generator.py

Source ingestion

Document downloads src/ingestion/pipeline.py
Source manifest src/ingestion/manifest_loader.py
Official downloader src/ingestion/downloader.py
RAG indexing src/rag/ingest.py
Document loader src/rag/document_loader.py
Structure chunker src/rag/chunker.py
Embedding provider src/rag/embeddings.py

Other components

Citizen
Official sources

Connections 27

Citizen → Page router: uses
Page router → Recommendations: routes to
Page router → Scheme details: routes to
Page router → Ask a question: routes to
Recommendations → API client: requests
Scheme details → API client: requests
Ask a question → API client: requests
API client → REST API: calls
REST API → Eligibility service: assesses
REST API → RAG orchestrator: answers
Eligibility service → Profile handling: normalizes
Eligibility service → Scheme ranking: ranks
Scheme ranking → Eligibility evaluator: evaluates
Eligibility evaluator → Rule engine: checks rules
Eligibility service → Explanations: formats
RAG orchestrator → Scheme detection: detects scheme
RAG orchestrator → Hybrid retrieval: retrieves
RAG orchestrator → Answer generation: generates answer
Hybrid retrieval → Keyword retrieval: queries
Hybrid retrieval → Vector index: queries
Official sources → Document downloads: provides documents
Document downloads → Source manifest: loads entries
Document downloads → Official downloader: downloads
RAG indexing → Document loader: loads documents
RAG indexing → Structure chunker: chunks documents
RAG indexing → Embedding provider: embeds chunks
RAG indexing → Vector index: indexes chunks

More HTML diagrams: ossu/computer-science · ripienaar/free-for-dev · pkuflyingpig/cs-self-learning · shanraisshan/claude-code-best-practice · cathrynlavery/diagram-design

Drawn by GitDiagram on September 28, 2026 from the repository's file tree, README and source files. Read as Markdown ·