Project guide · Source · Issues · Bala's work
Current stage: Research implementation · live pipeline validation pending. See the evidence and next release checklist.
A Python sports prediction research system connecting data collection, feature engineering, probabilistic models, backtesting and a web interface.
| Area | Source |
|---|---|
| Sports data collectors | Collectors |
| Feature engineering, models and evaluation | ML pipeline |
| Staking, portfolio limits and circuit breakers | Execution research |
| HTTP API and Telegram interface | API · Bot |
| Prediction interface | Web application |
| Automated checks | Tests |
This repository includes a Betfair data collector. It is a separate project from the owner's local Betfair Bot. Source code and backtest utilities do not establish profitability or validate a live trading system.
Python 3.11 or later is declared in pyproject.toml. PostgreSQL and the relevant provider configuration are required for database-backed workflows.
git clone https://github.com/BalaShankar9/Sharp-Edge.git
cd Sharp-Edge
python3 -m venv .venv
source .venv/bin/activate
python -m pip install -e '.[dev,ml,api]'
cp .env.example .env
# Configure a development database and only the providers you intend to use.
alembic upgrade head
python -m sharpedge.apiThe API defaults to port 8000. The commands above reflect the declared package extras and API entry point; a clean installation and all optional model dependencies still need release validation. Do not use production credentials for local experiments.
For the interface, see web/README.md, web/package.json and web/.env.example.
python -m pytestThe test suite covers collectors, models, execution helpers, API routes and pipelines. Live data jobs are configured separately in .github/workflows. A scheduled data job is not a substitute for CI.
Keep training and evaluation periods separate. Report calibration, baseline comparisons, sample sizes and source freshness before drawing conclusions. See the project guide for the proposed release evidence.