//Parse IMDb page data
private static void ParseIMDbPage(string imdbUrl, bool GetExtraInfo, ImdbMovie mov)
{
string html = GetUrlData(imdbUrl + "combined");
mov.Id = match(@"<link rel=""canonical"" href=""http://www.imdb.com/title/(tt\d{7})/combined"" />", html);
if (!string.IsNullOrEmpty(mov.Id))
{
mov.Status = true;
mov.Title = match(@"<title>(IMDb \- )*(.*?) \(.*?</title>", html, 2);
mov.OriginalTitle = match(@"title-extra"">(.*?)<", html);
mov.Year = match(@"<title>.*?\(.*?(\d{4}).*?\).*?</title>", match(@"(<title>.*?</title>)", html));
mov.Rating = match(@"<b>(\d.\d)/10</b>", html);
mov.Genres = MatchAll(@"<a.*?>(.*?)</a>", match(@"Genre.?:(.*?)(</div>|See more)", html)).Cast <string>().ToList();
mov.Plot = match(@"Plot:</h5>.*?<div class=""info-content"">(.*?)(<a|</div)", html);
//mov.Directors = matchAll(@"<td valign=""top""><a.*?href=""/name/.*?/"">(.*?)</a>", match(@"Directed by</a></h5>(.*?)</table>", html));
//mov.Writers = matchAll(@"<td valign=""top""><a.*?href=""/name/.*?/"">(.*?)</a>", match(@"Writing credits</a></h5>(.*?)</table>", html));
//mov.Producers = matchAll(@"<td valign=""top""><a.*?href=""/name/.*?/"">(.*?)</a>", match(@"Produced by</a></h5>(.*?)</table>", html));
//mov.Musicians = matchAll(@"<td valign=""top""><a.*?href=""/name/.*?/"">(.*?)</a>", match(@"Original Music by</a></h5>(.*?)</table>", html));
//mov.Cinematographers = matchAll(@"<td valign=""top""><a.*?href=""/name/.*?/"">(.*?)</a>", match(@"Cinematography by</a></h5>(.*?)</table>", html));
//mov.Editors = matchAll(@"<td valign=""top""><a.*?href=""/name/.*?/"">(.*?)</a>", match(@"Film Editing by</a></h5>(.*?)</table>", html));
//mov.Cast = matchAll(@"<td class=""nm""><a.*?href=""/name/.*?/"".*?>(.*?)</a>", match(@"<h3>Cast</h3>(.*?)</table>", html));
//mov.PlotKeywords = matchAll(@"<a.*?>(.*?)</a>", match(@"Plot Keywords:</h5>.*?<div class=""info-content"">(.*?)</div", html));
//mov.ReleaseDate = match(@"Release Date:</h5>.*?<div class=""info-content"">.*?(\d{1,2} (January|February|March|April|May|June|July|August|September|October|November|December) (19|20)\d{2})", html);
//mov.Runtime = match(@"Runtime:</h5><div class=""info-content"">(\d{1,4}) min[\s]*.*?</div>", html);
//mov.Top250 = match(@"Top 250: #(\d{1,3})<", html);
//mov.Oscars = match(@"Won (\d+) Oscars?\.", html);
//if (string.IsNullOrEmpty(mov.Oscars) && "Won Oscar.".Equals(match(@"(Won Oscar\.)", html))) mov.Oscars = "1";
//mov.Awards = match(@"(\d{1,4}) wins", html);
//mov.Nominations = match(@"(\d{1,4}) nominations", html);
//mov.Tagline = match(@"Tagline:</h5>.*?<div class=""info-content"">(.*?)(<a|</div)", html);
//mov.MpaaRating = match(@"MPAA</a>:</h5><div class=""info-content"">Rated (G|PG|PG-13|PG-14|R|NC-17|X) ", html);
//mov.Votes = match(@">(\d+,?\d*) votes<", html);
//mov.Languages = matchAll(@"<a.*?>(.*?)</a>", match(@"Language.?:(.*?)(</div>|>.?and )", html));
//mov.Countries = matchAll(@"<a.*?>(.*?)</a>", match(@"Country:(.*?)(</div>|>.?and )", html));
mov.Poster = match(@"<div class=""photo"">.*?<a name=""poster"".*?><img.*?src=""(.*?)"".*?</div>", html);
if (!string.IsNullOrEmpty(mov.Poster) && mov.Poster.IndexOf("media-imdb.com") > 0)
{
mov.Poster = Regex.Replace(mov.Poster, @"_V1.*?.jpg", "_V1._SY200.jpg");
//mov.PosterLarge = Regex.Replace(mov.Poster, @"_V1.*?.jpg", "_V1._SY500.jpg");
//mov.PosterFull = Regex.Replace(mov.Poster, @"_V1.*?.jpg", "_V1._SY0.jpg");
}
else
{
mov.Poster = string.Empty;
//mov.PosterLarge = string.Empty;
//mov.PosterFull = string.Empty;
}
mov.ImdbURL = "http://www.imdb.com/title/" + mov.Id + "/";
if (GetExtraInfo)
{
string plotHtml = GetUrlData(imdbUrl + "plotsummary");
//mov.Storyline = match(@"<p class=""plotpar"">(.*?)(<i>|</p>)", plotHtml);
GetReleaseDatesAndAka(mov);
//mov.MediaImages = getMediaImages(mov);
//mov.RecommendedTitles = getRecommendedTitles(mov);
}
}
}
public static ImdbMovie ImdbScrapeFromId(string imdbId, bool GetExtraInfo = true)
{
ImdbMovie mov = new ImdbMovie();
string imdbUrl = "http://www.imdb.com/title/" + imdbId + "/";
mov.Status = false;
if (!string.IsNullOrEmpty(imdbUrl))
{
ParseIMDbPage(imdbUrl, GetExtraInfo, mov);
}
return(mov);
}
//Constructor
public static ImdbMovie ImdbScrape(string MovieName, bool GetExtraInfo = true)
{
ImdbMovie mov = new ImdbMovie();
string imdbUrl = GetIMDbUrl(System.Uri.EscapeUriString(MovieName));
mov.Status = false;
if (!string.IsNullOrEmpty(imdbUrl))
{
ParseIMDbPage(imdbUrl, GetExtraInfo, mov);
}
return(mov);
}
//Get Recommended Titles
private static ArrayList GetRecommendedTitles(ImdbMovie mov)
{
ArrayList list = new ArrayList();
string recUrl = "http://www.imdb.com/widget/recommendations/_ajax/get_more_recs?specs=p13nsims%3A" + mov.Id;
string json = GetUrlData(recUrl);
list = MatchAll(@"title=\\""(.*?)\\""", json);
HashSet <String> set = new HashSet <string>();
foreach (String rec in list)
{
set.Add(rec);
}
return(new ArrayList(set.ToList()));
}
//Get all media images
private static ArrayList GetMediaImages(ImdbMovie mov)
{
ArrayList list = new ArrayList();
string mediaurl = "http://www.imdb.com/title/" + mov.Id + "/mediaindex";
string mediahtml = GetUrlData(mediaurl);
int pagecount = MatchAll(@"<a href=""\?page=(.*?)"">", match(@"<span style=""padding: 0 1em;"">(.*?)</span>", mediahtml)).Count;
for (int p = 1; p <= pagecount + 1; p++)
{
mediahtml = GetUrlData(mediaurl + "?page=" + p);
foreach (Match m in new Regex(@"src=""(.*?)""", RegexOptions.Multiline).Matches(match(@"<div class=""thumb_list"" style=""font-size: 0px;"">(.*?)</div>", mediahtml)))
{
String image = m.Groups[1].Value;
list.Add(Regex.Replace(image, @"_V1\..*?.jpg", "_V1._SY0.jpg"));
}
}
return(list);
}
//Get all release dates and aka-s
private static void GetReleaseDatesAndAka(ImdbMovie mov)
{
Dictionary <string, string> release = new Dictionary <string, string>();
string releasehtml = GetUrlData("http://www.imdb.com/title/" + mov.Id + "/releaseinfo");
foreach (string r in MatchAll(@"<tr class="".*?"">(.*?)</tr>", match(@"<table id=""release_dates"" class=""subpage_data spFirst"">\n*?(.*?)</table>", releasehtml)))
{
Match rd = new Regex(@"<td>(.*?)</td>\n*?.*?<td class=.*?>(.*?)</td>", RegexOptions.Multiline).Match(r);
release[StripHTML(rd.Groups[1].Value.Trim())] = StripHTML(rd.Groups[2].Value.Trim());
}
//mov.ReleaseDates = release;
Dictionary <string, string> aka = new Dictionary <string, string>();
ArrayList list = MatchAll(@".*?<tr class="".*?"">(.*?)</tr>", match(@"<table id=""akas"" class=.*?>\n*?(.*?)</table>", releasehtml));
foreach (string r in list)
{
Match rd = new Regex(@"\n*?.*?<td>(.*?)</td>\n*?.*?<td>(.*?)</td>", RegexOptions.Multiline).Match(r);
aka[StripHTML(rd.Groups[1].Value.Trim())] = StripHTML(rd.Groups[2].Value.Trim());
}
mov.Aka = aka;
}
