private static string[] ExtractTeams(HtmlNode game, ScrapingInfo info)
{
var separators = new string[] { " vs. ", " vs ", " v " };
string[] teams = null;
if (info.HomeTeamNameExpression != info.AwayTeamNameExpression ||
info.HomeTeamAttribute != info.AwayTeamAttribute)
{
teams = new string[2];
teams[0] = ReadFromNode(game, info.HomeTeamNameExpression, info.HomeTeamAttribute);
teams[1] = ReadFromNode(game, info.AwayTeamNameExpression, info.AwayTeamAttribute);
}
else
{
var teamsOneLine = ReadFromNode(game, info.HomeTeamNameExpression, info.HomeTeamAttribute).Replace(" ", " ");
foreach (var sep in separators)
{
if (teamsOneLine.Contains(sep))
{
teams = teamsOneLine.Split(new[] { sep }, StringSplitOptions.RemoveEmptyEntries);
break;
}
}
}
if (teams != null)
{
return(teams);
}
else
{
throw new FormatException();
}
}
private static DateTime GetDateTime(HtmlNode game, ScrapingInfo info)
{
var date = ParseDateTime(ReadFromNode(game, info.DateExpression, info.DateAttribute), info.DateFormat);
var time = ParseDateTime(ReadFromNode(game, info.TimeExpression, info.TimeAttribute), info.TimeFormat);
return(new DateTime(date.Year, date.Month, date.Day, time.Hour, time.Minute, time.Second));
}
static bool TryParsePage(string html, int siteId, ScrapingInfo info)
{
var page = new HtmlDocument();
page.LoadHtml(html);
var games = page.DocumentNode.SelectNodes(info.GameListExpression);
if (games == null)
{
return(false);
}
foreach (var game in games)
{
var datetime = GetDateTime(game, info);
var teams = ExtractTeams(game, info);
var team1 = teams[0].Trim();
var team2 = teams[1].Trim();
var r1 = ParseOdd(ReadFromNode(game, info.HomeRatioExpression, info.HomeRatioAttribute));
var rX = ParseOdd(ReadFromNode(game, info.RatioXExpression, info.RatioXAttribute));
var r2 = ParseOdd(ReadFromNode(game, info.AwayRatioExpression, info.AwayRatioAttribute));
AddGame(siteId, datetime, team1, team2, r1, rX, r2);
//Console.WriteLine("{0}: {1} vs {2}\nHome: {3}\tDraw: {4}\tAway: {5}\n\n", datetime, team1, team2, r1, rX, r2);
}
return(true);
}