private static void CrawlGame(string gid)
{
string gameId = gid.ToString();
string url = @"http://rivals.yahoo.com/ncaa/basketball/boxscore?gid=" + gameId;
if (long.Parse(gameId) > 201103100000)
{
url += "&old_bs=1";
}
Uri boxscoreUrl = new Uri(url);
string html = DownloadAndCache(boxscoreUrl);
HtmlDocument document = new HtmlDocument();
document.LoadHtml(html);
// the date is the first 8 integers of the game id
string date = gid.Substring(0, 8);
// get the team code values from the summarized table
// /html[1]/head[1]/body[1]/div[1]/table[1]/tr[1]/td[1]/table[2]/tr[1]/td[1]/div[1]/table[3]
// path to the top team code URL
HtmlNode awayUrlNode =
document.DocumentNode.SelectSingleNode(@"/html[1]/head[1]/body[1]/div[1]/table[1]/tr[1]/td[1]/table[2]/tr[1]/td[1]/div[1]/table[3]/tr[1]/td[2]/table[1]/tr[1]/td[1]/table[1]/tr[1]/td[1]/table[1]/tr[4]/td[2]");
HtmlNode homeUrlNode =
document.DocumentNode.SelectSingleNode(@"/html[1]/head[1]/body[1]/div[1]/table[1]/tr[1]/td[1]/table[2]/tr[1]/td[1]/div[1]/table[3]/tr[1]/td[2]/table[1]/tr[1]/td[1]/table[1]/tr[1]/td[1]/table[1]/tr[6]/td[2]");
// get the teamcode from the node
string awayTeamCode = GetTeamCodeFromNode(awayUrlNode);
string homeTeamCode = GetTeamCodeFromNode(homeUrlNode);
HtmlNode awayTable =
document.DocumentNode.SelectSingleNode(@"/html[1]/head[1]/body[1]/div[1]/table[1]/tr[1]/td[1]/table[2]/tr[1]/td[1]/div[1]/table[5]");
HtmlNode homeTable =
document.DocumentNode.SelectSingleNode(@"/html[1]/head[1]/body[1]/div[1]/table[1]/tr[1]/td[1]/table[2]/tr[1]/td[1]/div[1]/table[7]");
// we have both tables and both teams set
if (!string.IsNullOrEmpty(awayTeamCode) && !string.IsNullOrEmpty(homeTeamCode) && awayTable != null && homeTable != null)
{
// create the game object and write to disk
GameRow row = new GameRow();
row.date = date;
row.gid = gid;
row.awayTeam = awayTeamCode;
row.homeTeam = homeTeamCode;
WriteGameRow(row);
// parse the top (away) table
ParseScoreTable(awayTable, gid, awayTeamCode);
// parse the bottom (home) table
ParseScoreTable(homeTable, gid, homeTeamCode);
}
else
{
// could not parse this game, log it
ErrorLog.WriteLine("Could not parse " + gid);
ErrorLog.Flush();
}
}
private static void WriteGameRow(GameRow row)
{
if (GameStream == null)
{
GameStream = new StreamWriter(Path.Combine(OUTPUT_PATH, "Games.tsv"));
}
GameStream.WriteLine("{0}\t{1}\t{2}\t{3}", row.gid, row.date, row.homeTeam, row.awayTeam);
}
private static void WriteGameRow(GameRow row)
{
if (GameStream == null)
{
GameStream = new StreamWriter(Path.Combine(OUTPUT_PATH, "Games.tsv"));
}
GameStream.WriteLine("{0}\t{1}\t{2}\t{3}", row.gid, row.date, row.homeTeam, row.awayTeam);
}
private static void CrawlGame(string gid)
{
string gameId = gid.ToString();
string url = @"http://rivals.yahoo.com/ncaa/basketball/boxscore?gid=" + gameId;
if (long.Parse(gameId) > 201103100000)
{
url += "&old_bs=1";
}
Uri boxscoreUrl = new Uri(url);
string html = DownloadAndCache(boxscoreUrl);
HtmlDocument document = new HtmlDocument();
document.LoadHtml(html);
// the date is the first 8 integers of the game id
string date = gid.Substring(0, 8);
// get the team code values from the summarized table
// /html[1]/head[1]/body[1]/div[1]/table[1]/tr[1]/td[1]/table[2]/tr[1]/td[1]/div[1]/table[3]
// path to the top team code URL
HtmlNode awayUrlNode =
document.DocumentNode.SelectSingleNode(@"/html[1]/head[1]/body[1]/div[1]/table[1]/tr[1]/td[1]/table[2]/tr[1]/td[1]/div[1]/table[3]/tr[1]/td[2]/table[1]/tr[1]/td[1]/table[1]/tr[1]/td[1]/table[1]/tr[4]/td[2]");
HtmlNode homeUrlNode =
document.DocumentNode.SelectSingleNode(@"/html[1]/head[1]/body[1]/div[1]/table[1]/tr[1]/td[1]/table[2]/tr[1]/td[1]/div[1]/table[3]/tr[1]/td[2]/table[1]/tr[1]/td[1]/table[1]/tr[1]/td[1]/table[1]/tr[6]/td[2]");
// get the teamcode from the node
string awayTeamCode = GetTeamCodeFromNode(awayUrlNode);
string homeTeamCode = GetTeamCodeFromNode(homeUrlNode);
HtmlNode awayTable =
document.DocumentNode.SelectSingleNode(@"/html[1]/head[1]/body[1]/div[1]/table[1]/tr[1]/td[1]/table[2]/tr[1]/td[1]/div[1]/table[5]");
HtmlNode homeTable =
document.DocumentNode.SelectSingleNode(@"/html[1]/head[1]/body[1]/div[1]/table[1]/tr[1]/td[1]/table[2]/tr[1]/td[1]/div[1]/table[7]");
// we have both tables and both teams set
if (!string.IsNullOrEmpty(awayTeamCode) && !string.IsNullOrEmpty(homeTeamCode) && awayTable != null && homeTable != null)
{
// create the game object and write to disk
GameRow row = new GameRow();
row.date = date;
row.gid = gid;
row.awayTeam = awayTeamCode;
row.homeTeam = homeTeamCode;
WriteGameRow(row);
// parse the top (away) table
ParseScoreTable(awayTable, gid, awayTeamCode);
// parse the bottom (home) table
ParseScoreTable(homeTable, gid, homeTeamCode);
}
else
{
// could not parse this game, log it
ErrorLog.WriteLine("Could not parse " + gid);
ErrorLog.Flush();
}
}
