def scrape_games(loops, count_max, action):
stats_db = SqlDataframes()
games_table = stats_db.read_table('games',['game_id','bref'])
for i in range(loops):
logger.info('Current Loop: {}'.format(i+1))
logger.info('Running game soups...')
cfg = get_dbconfig(section='scraping')
checktables_cfg = cfg['check_tables'].split(',')
id_bref_soup = get_game_soups(games_table, limit=count_max, check_tables=checktables_cfg)
if not id_bref_soup:
break
add_basic_gamestats(id_bref_soup, commit_changes=action)
if i == loops - 1:
stats_db = SqlDataframes()
games_max = stats_db.read_max('games','game_id')
boxs_max = stats_db.read_max('boxscores', 'game_id')
logger.info('FINISHED...Games remaining to scrape: {}'.format(games_max-boxs_max))
import pandas as pd
import numpy as np
import datetime as dt
import re
import time
import matplotlib.pyplot as plt
import seaborn as sns
import mysql.connector as sql
from nba_stats.read_write.config import get_dbconfig
cfg = get_dbconfig()
HOST = cfg['host']
PORT = int(cfg['port'])
USER = cfg['user']
PASSWORD = cfg['password']
DB = cfg['db']
def isfloat(number):
try:
float(number)
return True
except (ValueError, TypeError):
return False
def add_where(original, add):
if original == '':
return 'WHERE ' + add
else: