Parse play-by-play data from basketballvalue.com for use in Gephi

parse_pbp.py

#!/usr/bin/python

import argparse
import csv
import re

parser = argparse.ArgumentParser(
    description='Extract plays involving assists from raw play-by-play data')
parser.add_argument('input_file',metavar='FILE')
parser.add_argument('--team', help='Three letter team name to search \
                    for, e.g. SAC, MEM, NYK')
args = parser.parse_args()

assist_pattern = re.compile(r'''
                            \[%s\ # Scoring team
                            (?P<t0_score>\d+)-(?P<t1_score>.*?)\]\ # Team scores
                            (?P<scorer>(?:[A-Z]\.\ )?\w+)\ # Player who scored
                            (?P<score_method>.*?)[:].*? # Method of scoring
                            Assist:\ (?P<assister>.*?)\ \( # Assister
                            ''' % args.team, re.X)

players = {}

def get_player_idx(p):
    if p not in players:
        players[p] = len(players) + 1
    return players[p]

with open(args.input_file, 'rb') as f, open('assists.csv', 'wb') as outfile:
    data = csv.reader(f, delimiter='\t')
    out = csv.writer(outfile, delimiter='\t')
    headers= ['SOURCE', 'TARGET']

    out.writerow(headers)

    for game, line, time_remaining, play in data:
        assisted = re.search(assist_pattern, play)
        if not assisted:
            continue
        p1 = get_player_idx(assisted.group('scorer'))
        p2 = get_player_idx(assisted.group('assister'))
        out.writerow([p1, p2])

with open('players.csv', 'wb') as players_outfile:
    players_out = csv.writer(players_outfile, delimiter='\t')
    players_out.writerow(['ID', 'LABEL'])
    for player, pid in players.items():
        players_out.writerow([pid, player])