How sattyajeetmaharana/mind_tone works
MindTone screens for mental-health risks by combining questionnaire predictions with conversational-audio classification, then produces a fused result and PDF report. The web app provides survey, account, dashboard, history, report, and administrative experiences; README documents the web-to-FastAPI assessment and PDF-download flow. The Python API and prediction path are sampled; some web wiring and unsampled model/report details are represented only where documented, so uncertain internal links are omitted.
Main entry points: server.py FastAPI assessment API · tabular/predict.py Questionnaire prediction · route.ts Assessment endpoint · audio/predict.py Audio prediction
Components 20
Web experience
- Survey and audio UI
web/src/app/(dashboard)/survey/page.tsx - Assessment endpoint
web/src/app/api/assessment/submit/route.ts - Dashboard and history
web/src/app/(dashboard)/dashboard/page.tsx - Report viewer and download
web/src/app/api/assessment/pdf/[pdfName]/route.ts - Account and admin
web/src/app/api/auth/login/route.ts
Assessment engine
- FastAPI assessment API
src/api/server.py - Questionnaire prediction
src/tabular/predict.py - Risk flag rules
src/tabular/risk_engine.py - Tabular preprocessing
src/tabular/preprocess.py - Tabular model scoring
src/tabular/model_manager.py - Audio prediction
src/audio/predict.py - Acoustic feature extraction
src/audio/extract_features.py - Audio model scoring
src/audio/model_manager.py - Multimodal fusion
src/fusion/fusion.py
Explainability and reports
- SHAP explainability
src/tabular/shap_analysis.py - Assessment graphs
src/report/graphs.py - PDF report generation
src/report/generate_report.py
Persistence
- CLI history and cleanup
src/history/save_history.py - Web relational database
web/prisma
Other components
- Assessment user
Connections 20
- Assessment user → Survey and audio UI: submits assessment
README.md - Survey and audio UI → Assessment endpoint: sends answers
README.md - Assessment endpoint → FastAPI assessment API: forwards payload
README.md - FastAPI assessment API → Questionnaire prediction: runs questionnaire
src/api/server.py - FastAPI assessment API → Audio prediction: runs audio analysis
src/api/server.py - Questionnaire prediction → Risk flag rules: generates flags
src/tabular/predict.py - Questionnaire prediction → Tabular preprocessing: preprocesses inputs
src/tabular/predict.py - Questionnaire prediction → Tabular model scoring: scores features
src/tabular/predict.py - Audio prediction → Acoustic feature extraction: extracts features
src/audio/predict.py - Audio prediction → Audio model scoring: scores features
src/audio/predict.py - FastAPI assessment API → Multimodal fusion: fuses predictions
src/api/server.py - FastAPI assessment API → Assessment graphs: generates plots
src/api/server.py - FastAPI assessment API → PDF report generation: compiles report
src/api/server.py - Assessment user → Account and admin: signs in
README.md - Account and admin → Web relational database: stores account data
README.md - Dashboard and history → Web relational database: uses check-in records
- Assessment endpoint → Web relational database: stores assessment records
- FastAPI assessment API → Report viewer and download: returns report name
README.md - Report viewer and download → PDF report generation: streams generated PDF
README.md - FastAPI assessment API → SHAP explainability: explainability capability