Spaces:
Sleeping
Sleeping
Commit History
-
Upload folder using huggingface_hub
Verified signaturee406ddb -
Upload folder using huggingface_hub
Verified signaturebe68e43 -
Upload folder using huggingface_hub
Verified signature74e067a -
Upload folder using huggingface_hub
Verified signature5b54e78
-
Upload folder using huggingface_hub
Verified signature9796e1e
-
Upload folder using huggingface_hub
Verified signature90b2c67
-
Upload folder using huggingface_hub
Verified signaturecf58242 -
Upload folder using huggingface_hub
Verified signature100ddbe
-
chore: bump API version to 0.8.0
06065d7 -
feat: update ML pipeline with preprocessing module and string-based inputs
- Replace integer-encoded fields (genre, heure_supplementaires, frequence_deplacement) with human-readable Literal/String types in schema and DB models - Add departement and niveau_hierarchique_poste fields to schema and DB models - Extract feature engineering from predict route into ml_model/preprocessing.py - Update best_pipeline.pkl to include preprocessing stepsf67b60a
-
chore: bump API version to 0.7.0
4dd5069 -
fix: lazy-initialize DB engine to prevent CI import crash
Move create_engine and SessionLocal inside get_db so the database connection is only created when actually needed. Fixes ImportError during pytest when DB environment variables are not set.36fe10e -
chore: bump API version to 0.5.0
82af15e
-
chore: bump API version to 0.4.0
77fb213 -
feat: log predictions to database after each inference
75c7c56 -
feat: add PostgreSQL database layer with SQLAlchemy ORM
- Add Employee and Prediction SQLAlchemy models (app/db/models.py) - Add database session/engine configuration from env vars (app/db/session.py) - Add script to create database and tables if not exists (scripts/create_db.py) - Add docker-compose with PostgreSQL + pgAdmin services - Update .env.example with DB_* connection variables - Add psycopg2-binary, SQLAlchemy, greenlet to requirements.txt - Document database schema, UML diagram and column constraints in READMEa194bc8 -
chore: bump API version to 0.3.0
88f79b5 -
chore: bump API version to 0.2.0
b6e1abe
-
docs: complete README with usage guide and API documentation
- Improve API key validation: handle missing key with explicit 401 error message - Exclude data/ directory from version control via .gitignore59826c1 -
feat: secure endpoints with API key authentication
a9477b4 -
feat: add FastAPI REST API with HR churn prediction endpoint
- Create app/main.py with FastAPI app and lifespan model loading - Add Pydantic schemas for input validation (24 fields) and structured output (prediction, label, probabilite) - Implement /health, /model/info and /predict endpoints - Add ml_model/loader.py to load pipeline and custom threshold from pkl - Apply feature engineering before inference: derive stabilite_management, flag_surcharge_et_deplacement, ratio_promotion_anciennete, satisfaction_globale, satisfaction_min, log_revenu; drop 3 raw columns - Classify with custom threshold (predict_proba >= threshold) - Update requirements.txt with fastapi, uvicorn, scikit-learn, pandase59e621
-
init: project structure and base configuration
a395150