How dhakshesh-48/se-project works
ReqMind is a local financial requirements-engineering app. Stakeholders provide project context through an adaptive interview and optional source uploads; the API masks, extracts, chunks, and indexes those sources, then retrieves evidence for cited draft requirements. Users inspect evidence, review quality findings, and set approval statuses before requesting an SDLC recommendation or exporting the audit trail. The map distinguishes the API orchestrator from processing and analysis services. README capabilities are included, but some main.py wiring is inferred from documented endpoints because its excerpt omits lines.
Main entry points: main.py API orchestration · app.js ReqMind UI · rag.py Semantic retrieval · llm.py LLM generation
Components 17
User experience
- ReqMind UI
app/static/app.js - App page
app/templates/index.html - API orchestration
app/main.py
Evidence pipeline
- Document processing
app/services/documents.py - Semantic retrieval
app/services/rag.py - ChromaDB
Requirements analysis
- LLM generation
app/services/llm.py - Generation prompt
app/prompts/requirement_generation.txt - Evidence-only drafts
app/services/generator.py - Quality analysis
app/services/quality.py - Requirement review
app/main.py - SDLC recommendation
app/services/sdlc.py
Project records
- Domain schemas
app/schemas.py - Project storage
app/services/storage.py - Audit and export
app/main.py
Other components
- Stakeholder
- Embedding API
Connections 20
- Stakeholder → ReqMind UI: provides context
- App page → ReqMind UI: renders
- ReqMind UI → API orchestration: calls API
- API orchestration → Document processing: extracts, masks, chunks
- API orchestration → Semantic retrieval: indexes sources
- Semantic retrieval → Embedding API: requests embeddings
- Semantic retrieval → ChromaDB: indexes and queries
- API orchestration → Project storage: loads and saves
- API orchestration → Semantic retrieval: retrieves evidence
- API orchestration → LLM generation: generates drafts
- LLM generation → Generation prompt: reads template
- LLM generation → Embedding API: requests generation
- API orchestration → Evidence-only drafts: creates fallback drafts
- API orchestration → Quality analysis: checks requirements
- API orchestration → Requirement review: records decision
- ReqMind UI → Requirement review: submits review
- API orchestration → SDLC recommendation: scores requirements
- API orchestration → Audit and export: requests export
- Audit and export → Project storage: loads project
- API orchestration → Domain schemas: validates project data