How santiesleo/ti2-2025-1-e7probabilibets works
Project type and purpose
• From the file tree and docs, this is a predictive‐analytics application for sports matches. It ingests historical datasets (matches, seasons, team stats), applies Bayesian and rule‐based inference to generate predictions, and exposes those predictions via a Telegram bot UI.
File structure
• Top-level:
– datasets/ (raw CSV data)
– src/
· data/ (DataManager: loading, cleaning, caching CSVs)
· inference/ (BayesianModel, BayesianAnalyzer, MatchAnalyzer, RuleEngine)
· ui/ (TelegramBot adaptor)
– main.py (entry point/orchestrator)
– tests/ (unit and stress tests)
– docs/ (requirements, analysis, report)
• Config/build: requirements.txt, .python-version, .gitignore
• The pattern is a layered architecture: Data → Inference → Presentation.
README
• The README itself contains only a classroom badge, so we rely on docs/*.md for design insights. The docs likely describe data flows, model training, and inference steps.