Fire Compliance and Risk Analysis Platform
A semantic web application for building fire safety compliance checking and risk assessment, powered by an OWL ontology (FiCR), SPARQL-based competency queries, and LLM-assisted analysis.
Live demo: https://RainGo111.github.io/FiCR/
Ontology Documentation — Interactive browser for FiCR TBox classes and properties
Compliance & Risk Report — Pre-generated static report from ficr_demo.ttl (Duplex A), structured data tables + AI analysis
FiCR Agent — LLM-powered pipeline: NL/JSON input → RDF knowledge graph → SPARQL compliance queries → narrative report
QueryLab — SPARQL editor with 15 preset competency queries, live execution against GraphDB, deterministic report generation
flowchart TD
A["NL description / JSON upload"] --> B["LLM #1 — Survey Extractor"]
B --> C["survey JSON (ficr-survey-v1)"]
C --> D["JSON → RDF (rdflib, deterministic)"]
D --> E["ABox (.ttl)"]
F["TBox + Regulatory Config"] --> G
E --> G["SPARQL Runner — 15 CQs"]
G --> H["Deterministic Report<br/>(Python templates)"]
G --> I["LLM #3 — Report Narrator<br/>(SSE streaming)"]
J["User NL question"] --> K["LLM #2 — Query Selector"]
K --> G
style B fill:#e8f4fd,stroke:#4a90d9
style K fill:#e8f4fd,stroke:#4a90d9
style I fill:#e8f4fd,stroke:#4a90d9
style D fill:#f0f0f0,stroke:#999
style G fill:#f0f0f0,stroke:#999
style H fill:#f0f0f0,stroke:#999
Three LLM prompts under backend/prompts/:
| File | Role | Input / Output |
|---|---|---|
1_survey_extractor.md |
LLM #1 | NL building description → ficr-survey-v1 JSON |
2_query_selector.md |
LLM #2 | User NL question → CQ ID + optional FILTER |
3_report_narrator.md |
LLM #3 | SPARQL results → diagnostic analysis + recommendations |
ontology/ # Canonical ontology source
ficr_tbox.ttl # TBox (classes, properties, OWL axioms)
ficr_demo.ttl # ABox — Duplex A instance data
ficr_regulatory_config.ttl # REI thresholds (PG 1b, ADB)
ficr_risk_discovery_queries.sparql # 15 SPARQL competency queries
VERSION
backend/
server.py # FastAPI + SSE streaming
pipeline.py # 4-stage pipeline orchestrator
ficr_json_to_rdf.py # JSON → RDF (deterministic)
ficr_sparql_runner.py # SPARQL execution engine
report_generator.py # Deterministic Markdown report
prompts/
1_survey_extractor.md # LLM #1 — NL → JSON
2_query_selector.md # LLM #2 — question → CQ
3_report_narrator.md # LLM #3 — results → narrative
schemas/survey_schema.json # ficr-survey-v1 JSON Schema
references/ # Synced ontology copies for backend
sessions/ # Pipeline intermediate outputs
src/
pages/ # React pages (Home, Documentation, QueryLab, Chatbot, Report, Roadmap)
components/
report/ # Shared report components (ReportDataView, DataTable, HealthScoreCard, etc.)
chatbot/ # Chat UI (ChatMessage, ChatInput, PipelineProgress)
documentation/ # Ontology browser components
content/queries.ts # 15 preset SPARQL queries
data/demoReport.ts # Static demo data for Demo Report page
hooks/useSparqlQuery.ts # GraphDB query hook
scripts/sync_ontology.py # Sync ontology/ → public/ + backend/references/
- Node.js (v16+), Python (3.10+)
- At least one LLM API key (Claude, OpenAI, Gemini, DeepSeek, or Zhipu GLM) for FiCR Agent
- GraphDB instance for QueryLab (optional — Demo Report and Documentation work without it)
npm install
pip install -r backend/requirements.txt
# Configure environment
cp .env.example .env
# Edit .env — set GraphDB connection (optional, for QueryLab)
cp backend/.env.example backend/.env
# Edit backend/.env — add at least one LLM API key# Terminal 1 — Frontend
npm run dev
# Terminal 2 — Backend
cd backend && uvicorn server:app --port 8000 --reloadpython scripts/sync_ontology.pyPortal implementation: MIT. FiCR ontology retains its original license.