Implement InferentialModels.predict_ball_by_ball_outcome()

[swati0308]

Input:

• A dataframe indexed by the match key & scenario # representing the current state of the match so far

Output: A dataframe indexed by the match key & scenario # which represents the outcome for the current status, and must include the following fields:

• batter_runs: Number of runs attributable to the batter
• extras: Number of extras scored
• is_wicket: 1 if there was a wicket in this ball, else 0. Total_runs, batter_runs and extras may or may not be 0.
• dismissal_kind: If is_wicket == 1, should be one of: 'bowled', 'caught', 'run out', 'stumped', 'caught and bowled', else 'nan'
• is_direct_runout: if is_wicket == 1 and dismissal == 'run out', populated with 1 if this was a direct runout, else 0
• noballs: 1 if this ball was a no ball, else 0. If this ball was a noball, extras must be >= 1 and batter_runs may or may not be 0.
• wides: 1 if this ball was a wide, else 0. If this ball was a wide, extras must be >= 1 and batter_runs = 0
• player_dismissed: If is_wicket == 1, indicates who has been dismissed: 'striker' if the striker has been dismissed, 'non_striker' if the non striker has been dismissed. 'nan' if is_wicket == 0.

Fields out of scope from this model (to be predicted using the Predictive Model):

• fielder - pick a random player from the bowling playing_xi if the dismissal_kind is one of 'caught', 'run out' or 'stumped'

[swati0308]

This has been completed for batter runs in issue 28. The rest will be done as part of later SOWs