How debadrita-mukhopadhyay5/visionguardai works
VisionGuard AI is a Streamlit civic-safety inspection app. Its main workflow accepts one or more images, detects hazards, checks image reliability, computes a transparent risk score, and presents prioritized results; single inspections can also generate and download a report. The graph preserves the separate detection, quality, scoring, AI guidance, report formatting, and batch-prioritization stages. Edges are limited to calls visible in the sampled source or the README; external AI services are shown as optional integrations.
Main entry points: app.py Streamlit App · detector.py Hazard Detection · genai.py AI Report Generation · config.py Runtime Config
Components 11
Inspection UI
- Streamlit App
app.py - Batch Prioritization
app.py
Image Analysis
- Hazard Detection
detector.py - Image Reliability
quality.py - Risk Scoring
risk_engine.py - Runtime Config
config.py
Guidance & Reports
- AI Report Generation
genai.py - Report Formatting
report.py
Other components
- Inspector
- Roboflow API
- Anthropic API
Connections 11
- Inspector → Streamlit App: uploads images
app.py - Streamlit App → Hazard Detection: detects hazards
app.py - Streamlit App → Image Reliability: assesses quality
app.py - Streamlit App → Risk Scoring: calculates risk
app.py - Streamlit App → AI Report Generation: requests report
app.py - Streamlit App → Report Formatting: formats download
app.py - Streamlit App → Batch Prioritization: prioritizes batch
app.py - Hazard Detection → Runtime Config: reads settings
detector.py - Hazard Detection → Roboflow API: posts image
detector.py - AI Report Generation → Runtime Config: reads settings
genai.py - AI Report Generation → Anthropic API: requests generation
genai.py