How hippomasakib1/aegnyxzero works
AegnyxZero turns curated NASA combustion experiments into a decision dashboard: planners set conditions, inspect ranked material risks and experimental coverage, and can ask evidence-grounded questions. The backend serves the dashboard’s ranking and coverage requests and routes questions through retrieval, synthesis, and deterministic validation before returning an answer. Evidence details and comparisons are separate UI views. The README documents ingestion and offline snapshot capabilities, but those data-pipeline relationships are shown only at the documented level; some frontend-to-API and provider wiring is likewise not fully excerpted.
Main entry points: App.tsx Dashboard app · main.py FastAPI routes · pipeline_v2.py QA pipeline · models.py Experiment database
Components 24
Decision dashboard
- Dashboard app
frontend/src/App.tsx - Scenario controls
frontend/src/components/ConditionsBar.tsx - Risk ranking
frontend/src/components/RankedTable.tsx - Coverage map
frontend/src/components/CoverageMap.tsx - Evidence drawer
frontend/src/components/EvidenceDrawer.tsx - Comparison charts
frontend/src/components/ComparisonCharts.tsx - Ask AI drawer
frontend/src/components/AskAIDrawer.tsx - Review and info
frontend/src/components/Modals.tsx - Offline snapshot
frontend/src/data/snapshot.json
API and analysis
- FastAPI routes
backend/app/main.py - Risk scoring
backend/app/scoring.py - Coverage computation
backend/app/coverage.py
Grounded answers
- Ask handler
backend/app/ask_ai.py - QA pipeline
backend/app/pipeline_v2.py - Evidence retrieval
backend/app/retrieval.py - Answer validators
backend/app/validators.py - LLM gateway
backend/app/llm_gateway.py - Completion cache
backend/app/llm_gateway.py
Experiment evidence
- Experiment database
backend/app/models.py - Database sessions
backend/app/database.py - Vector index
backend/app/retrieval.py - Data preparation
data_pipeline
Other components
- Mission planner
- LLM providers
Connections 29
- Mission planner → Dashboard app: uses
README.md - Dashboard app → Scenario controls: renders
frontend/src/App.tsx - Dashboard app → Risk ranking: renders
frontend/src/App.tsx - Dashboard app → Coverage map: renders
frontend/src/App.tsx - Dashboard app → Evidence drawer: renders
frontend/src/App.tsx - Dashboard app → Comparison charts: renders
frontend/src/App.tsx - Dashboard app → Ask AI drawer: renders
frontend/src/App.tsx - Dashboard app → Review and info: renders
frontend/src/App.tsx - Dashboard app → Offline snapshot: loads
frontend/src/App.tsx - Dashboard app → FastAPI routes: requests ranking and coverage
frontend/src/App.tsx - Dashboard app → FastAPI routes: requests answers
- FastAPI routes → Experiment database: queries experiments
backend/app/main.py - FastAPI routes → Risk scoring: computes ranking
backend/app/main.py - FastAPI routes → Coverage computation: computes coverage
backend/app/main.py - FastAPI routes → Ask handler: routes questions
backend/app/main.py - FastAPI routes → QA pipeline: routes questions
backend/app/main.py - FastAPI routes → Database sessions: opens sessions
backend/app/main.py - Experiment database → Database sessions: uses base
backend/app/models.py - Ask handler → Answer validators: validates answers
backend/app/ask_ai.py - QA pipeline → Evidence retrieval: retrieves evidence
backend/app/pipeline_v2.py - QA pipeline → LLM gateway: synthesizes answers
backend/app/pipeline_v2.py - QA pipeline → Answer validators: validates drafts
backend/app/pipeline_v2.py - Evidence retrieval → Experiment database: loads cited rows
backend/app/retrieval.py - Evidence retrieval → Vector index: searches chunks
backend/app/retrieval.py - LLM gateway → Completion cache: reads and writes
backend/app/llm_gateway.py - LLM gateway → LLM providers: calls with fallback
backend/app/llm_gateway.py - Data preparation → Experiment database: loads curated rows
README.md - Data preparation → Vector index: builds index
README.md - Data preparation → Offline snapshot: exports snapshot
README.md