How dagbolade/all_leagues-_prediction works
Matchday is a multi-sport analysis website and CLI, with sportslab/ as the active implementation. The principal workflow is provider data refresh into SQLite, chronological feature construction and model training, then website predictions and a private, settleable forecast journal. The graph follows the README’s documented active architecture; source excerpts for sportslab/web.py, features.py, model.py, and the CLI are unavailable, so those internal links are based on README statements rather than inspected call sites.
Main entry points: web.py Flask website · ingest.py Provider importer · __main__.py Sportslab CLI · model.py Training and benchmark
Components 17
User interfaces
- Sportslab CLI
sportslab/__main__.py - Flask website
sportslab/web.py - Website interface
sportslab/static/app.js
Data operations
- Provider importer
sportslab/ingest.py - Provider catalog
sportslab/catalog.py - Team aliases
sportslab/aliases.py - Writer lock
sportslab/locking.py
Prediction and evaluation
- Chronological features
sportslab/features.py - Sport models
sportslab/model.py - Training and benchmark
sportslab/model.py
State and artifacts
- SQLite store
sportslab/store.py - Raw downloads
sportslab/ingest.py - Model artifacts
sportslab/model.py - Forecast journal
sportslab/store.py
Other components
- Analyst
- Browser
- Sports data providers
Connections 22
- Analyst → Sportslab CLI: runs commands
README.md - Analyst → Browser: uses website
README.md - Browser → Website interface: uses interface
README.md - Website interface → Flask website: requests views
README.md - Sportslab CLI → Provider importer: refreshes data
README.md - Sports data providers → Provider importer: provides results
README.md - Provider importer → Provider catalog: selects sources
README.md - Provider importer → Team aliases: resolves identities
README.md - Provider importer → Raw downloads: archives responses
README.md - Provider importer → SQLite store: upserts matches
README.md - Sportslab CLI → Writer lock: locks writers
README.md - Sportslab CLI → Training and benchmark: trains models
README.md - Training and benchmark → Chronological features: builds features
README.md - Training and benchmark → SQLite store: reads matches
README.md - Training and benchmark → Model artifacts: writes versions
README.md - Flask website → Chronological features: builds features
README.md - Flask website → Sport models: predicts matches
README.md - Flask website → SQLite store: reads match data
README.md - Sport models → Model artifacts: loads serving models
README.md - Flask website → Forecast journal: records forecasts
README.md - Flask website → Forecast journal: settles forecasts
README.md - Flask website → SQLite store: shows coverage
README.md