Perf Fixes: Speed up rewards calculation

SouthridgeCorp / player-outcome-predictor

Predictive modeling to forecast player outcomes over a T20 tournament

MIT License

0 stars 0 forks source link

Closed swati0308 closed 1 year ago

swati0308 commented 1 year ago

Speeds up rewards calculation:
- Test case: IPL 2009 test case with scenario count = 50
  - benchmark 1: Granularity = 'tournament', 1m40s (main) vs 32s (this branch)
  - benchmark 2: Granularity = 'tournament_stage', 1m54s (main) vs 40s (this branch)
Gets rid of inefficient for loops and replaces them by recommended data frame operations