alt/ahl15-16/getahlgames.py

   1 #!/usr/bin/env python
   2 # -*- coding: utf-8 -*-
   3
   4 import re
   5 import os
   6 import sys
   7 import urllib
   8 import urllib2
   9 import cookielib
  10 import StringIO
  11 import gzip
  12 import time
  13 import datetime
  14 import argparse
  15 import urlparse
  16 import sqlite3
  17 import unicodedata
  18
  19 sqlcon = sqlite3.connect("../hockey15-16.db3")
  20 sqlcur = sqlcon.cursor()
  21
  22 leaguename = "AHL"
  23 getforday = "9"
  24 getformonth = "10"
  25 getforyear = "2015"
  26 useragent = "Mozilla/5.0 (Windows NT 6.1; rv:41.0) Gecko/20100101 Firefox/41.0"
  27
  28 if (len(sys.argv) > 1):
  29     getforday = sys.argv[1]
  30 if (len(sys.argv) > 2):
  31     getformonth = sys.argv[2]
  32 if (len(sys.argv) > 3):
  33     getforyear = sys.argv[3]
  34
  35
  36 def GetFull2Team(sqldatacon, TeamName):
  37     global leaguename
  38     return str(sqldatacon[0].execute("SELECT TeamName FROM "+leaguename+"Teams WHERE FullName=\""+str(TeamName)+"\"").fetchone()[0])
  39
  40
  41 def GetTeam2Full(sqldatacon, TeamName):
  42     global leaguename
  43     return str(sqldatacon[0].execute("SELECT FullName FROM "+leaguename+"Teams WHERE TeamName=\""+str(TeamName)+"\"").fetchone()[0])
  44
  45
  46 geturls_cj = cookielib.CookieJar()
  47 geturls_opener = urllib2.build_opener(urllib2.HTTPCookieProcessor(geturls_cj))
  48 geturls_opener.addheaders = [("Referer", "http://www.nhl.com/"), ("User-Agent", useragent), ("Accept-Encoding", "gzip, deflate"), ("Accept-Language", "en-US,en;q=0.8,en-CA,en-GB;q=0.6"),
  49                              ("Accept-Charset", "ISO-8859-1,ISO-8859-15,utf-8;q=0.7,*;q=0.7"), ("Accept", "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8"), ("Connection", "close")]
  50 urllib2.install_opener(geturls_opener)
  51 geturls_text = geturls_opener.open(
  52     "http://theahl.com/stats/schedule.php?date="+getforyear+"-"+getformonth+"-"+getforday)
  53 if (geturls_text.info().get("Content-Encoding") == "gzip" or geturls_text.info().get("Content-Encoding") == "deflate"):
  54     strbuf = StringIO.StringIO(geturls_text.read())
  55     gzstrbuf = gzip.GzipFile(fileobj=strbuf)
  56     prehockey_text = gzstrbuf.read()[:]
  57 if (geturls_text.info().get("Content-Encoding") != "gzip" and geturls_text.info().get("Content-Encoding") != "deflate"):
  58     prehockey_text = geturls_text.read()[:]
  59 pre_get_todays_games = re.escape(
  60     "<a href=\"game-summary.php?game_id=")+"([0-9]+)"+re.escape("\">Game Summary</a>")
  61 get_todays_games = re.findall(pre_get_todays_games, prehockey_text)
  62 num_todays_games = len(get_todays_games)
  63 cur_todays_games = 0
  64 if (num_todays_games > 0):
  65     print("print(\"Inserting \"+leaguename+\" Game Data From " +
  66           getformonth.lstrip('0')+"/"+getforday.lstrip('0')+"/"+getforyear+".\\n\");")
  67 while (cur_todays_games < num_todays_games):
  68     newgetforday = getforday
  69     if (len(getforday) == 1):
  70         newgetforday = "0"+getforday
  71     newgetformonth = getformonth
  72     if (len(getformonth) == 1):
  73         newgetformonth = "0"+getformonth
  74     geturls_opener = urllib2.build_opener(
  75         urllib2.HTTPCookieProcessor(geturls_cj))
  76     geturls_opener.addheaders = [("Referer", "http://theahl.com/stats/schedule.php?date="+getforyear+"-"+getformonth+"-"+getforday), ("User-Agent", useragent), ("Accept-Encoding", "gzip, deflate"), ("Accept-Language",
  77                                                                                                                                                                                                        "en-US,en;q=0.8,en-CA,en-GB;q=0.6"), ("Accept-Charset", "ISO-8859-1,ISO-8859-15,utf-8;q=0.7,*;q=0.7"), ("Accept", "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8"), ("Connection", "close")]
  78     urllib2.install_opener(geturls_opener)
  79     geturls_text = geturls_opener.open(
  80         "http://theahl.com/stats/game-summary.php?game_id="+get_todays_games[cur_todays_games])
  81     if (geturls_text.info().get("Content-Encoding") == "gzip" or geturls_text.info().get("Content-Encoding") == "deflate"):
  82         strbuf = StringIO.StringIO(geturls_text.read())
  83         gzstrbuf = gzip.GzipFile(fileobj=strbuf)
  84         curgame_text = gzstrbuf.read()[:]
  85     if (geturls_text.info().get("Content-Encoding") != "gzip" and geturls_text.info().get("Content-Encoding") != "deflate"):
  86         curgame_text = geturls_text.read()[:]
  87     numpers = 5
  88     pre_curgame_score_text = re.escape("<td align=\"center\" class=\"content\">")+"([0-9]+)"+re.escape("</td><td align=\"center\" class=\"content\">")+"([0-9]+)"+re.escape("</td><td align=\"center\" class=\"content\">")+"([0-9]+)"+re.escape(
  89         "</td><td align=\"center\" class=\"content\">")+"([0-9]+)"+re.escape("</td><td align=\"center\" class=\"content\">")+"([0-9]+)"+re.escape("</td><td align=\"center\" class=\"content\"><b>")+"([0-9]+)"+re.escape("</b></td>")
  90     curgame_score_text = re.findall(pre_curgame_score_text, curgame_text)
  91     if (len(curgame_score_text) == 0):
  92         numpers = 4
  93         pre_curgame_score_text = re.escape("<td align=\"center\" class=\"content\">")+"([0-9]+)"+re.escape("</td><td align=\"center\" class=\"content\">")+"([0-9]+)"+re.escape(
  94             "</td><td align=\"center\" class=\"content\">")+"([0-9]+)"+re.escape("</td><td align=\"center\" class=\"content\">")+"([0-9]+)"+re.escape("</td><td align=\"center\" class=\"content\"><b>")+"([0-9]+)"+re.escape("</b></td>")
  95         curgame_score_text = re.findall(pre_curgame_score_text, curgame_text)
  96     if (len(curgame_score_text) == 0):
  97         numpers = 3
  98         pre_curgame_score_text = re.escape("<td align=\"center\" class=\"content\">")+"([0-9]+)"+re.escape("</td><td align=\"center\" class=\"content\">")+"([0-9]+)"+re.escape(
  99             "</td><td align=\"center\" class=\"content\">")+"([0-9]+)"+re.escape("</td><td align=\"center\" class=\"content\"><b>")+"([0-9]+)"+re.escape("</b></td>")
 100         curgame_score_text = re.findall(pre_curgame_score_text, curgame_text)
 101     pre_team_text = re.escape(
 102         "<tr class=\"light\"><td class=\"content\" nowrap>")+"(.*)"+re.escape("</td>")
 103     team_text = re.findall(pre_team_text, unicodedata.normalize(
 104         'NFKD', curgame_text.decode("utf-8")).encode('ASCII', 'ignore'))
 105     pre_team_stats = re.escape(
 106         "<td class=\"light\" align=\"center\">")+"(.*)"+re.escape("</td>")
 107     get_team_stats = re.findall(pre_team_stats, curgame_text)
 108     if (numpers == 3):
 109         print("MakeHockeyGame((sqlcur, sqlcon), "+getforyear+newgetformonth+newgetforday+", \""+team_text[3]+"\", \""+team_text[2]+"\", \""+curgame_score_text[1][0]+":"+curgame_score_text[0][0]+","+curgame_score_text[1][1]+":"+curgame_score_text[0][1]+"," +
 110               curgame_score_text[1][2]+":"+curgame_score_text[0][2]+"\", \""+curgame_score_text[3][0]+":"+curgame_score_text[2][0]+","+curgame_score_text[3][1]+":"+curgame_score_text[2][1]+","+curgame_score_text[3][2]+":"+curgame_score_text[2][2]+"\", 0, False);")
 111     if (numpers == 4):
 112         print("MakeHockeyGame((sqlcur, sqlcon), "+getforyear+newgetformonth+newgetforday+", \""+team_text[3]+"\", \""+team_text[2]+"\", \""+curgame_score_text[1][0]+":"+curgame_score_text[0][0]+","+curgame_score_text[1][1]+":"+curgame_score_text[0][1]+","+curgame_score_text[1][2]+":"+curgame_score_text[0][2]+"," +
 113               curgame_score_text[1][3]+":"+curgame_score_text[0][3]+"\", \""+curgame_score_text[3][0]+":"+curgame_score_text[2][0]+","+curgame_score_text[3][1]+":"+curgame_score_text[2][1]+","+curgame_score_text[3][2]+":"+curgame_score_text[2][2]+","+curgame_score_text[3][3]+":"+curgame_score_text[2][3]+"\", 0, False);")
 114     if (numpers == 5):
 115         print("MakeHockeyGame((sqlcur, sqlcon), "+getforyear+newgetformonth+newgetforday+", \""+team_text[3]+"\", \""+team_text[2]+"\", \""+curgame_score_text[1][0]+":"+curgame_score_text[0][0]+","+curgame_score_text[1][1]+":"+curgame_score_text[0][1]+","+curgame_score_text[1][2]+":"+curgame_score_text[0][2]+","+curgame_score_text[1][3]+":"+curgame_score_text[0][3]+"," +
 116               curgame_score_text[1][4]+":"+curgame_score_text[0][4]+"\", \""+curgame_score_text[3][0]+":"+curgame_score_text[2][0]+","+curgame_score_text[3][1]+":"+curgame_score_text[2][1]+","+curgame_score_text[3][2]+":"+curgame_score_text[2][2]+","+curgame_score_text[3][3]+":"+curgame_score_text[2][3]+","+curgame_score_text[3][4]+":"+curgame_score_text[2][4]+"\", 0, False);")
 117     cur_todays_games = cur_todays_games + 1
 118
 119 sqlcon.close()