mirror/Jackett
1
0
Fork 0
mirror of https://github.com/Jackett/Jackett synced 2025-03-09 13:52:03 +00:00
Code Issues Releases Wiki Activity

funfile: refactor parsing and multi-category filtering (#13954)

Browse source
This commit is contained in:
Bogdan 2023-02-01 10:00:26 +02:00 • committed by GitHub
parent 24d4198e59
commit 40fcb1e43b
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
1 changed files with 51 additions and 47 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

98
src/Jackett.Common/Indexers/FunFile.cs
View file

@ -80,6 +80,7 @@ namespace Jackett.Common.Indexers
var pairs = new Dictionary<string, string> { var pairs = new Dictionary<string, string> {
{ "username", configData.Username.Value }, { "username", configData.Username.Value },
{ "password", configData.Password.Value }, { "password", configData.Password.Value },
{ "returnto", "" },
{ "login", "Login" } { "login", "Login" }
}; };
@ -107,11 +108,12 @@ namespace Jackett.Common.Indexers
* so ?search=reality+love+s04e19&s_title=1&s_tag=1 * so ?search=reality+love+s04e19&s_title=1&s_tag=1
* will find Love Island S04E19 as well as My Kitchen Rules S12E02 * will find Love Island S04E19 as well as My Kitchen Rules S12E02
* since the former has a match for Love in the title, and the latter has a tag for Reality-TV. * since the former has a match for Love in the title, and the latter has a tag for Reality-TV.
* *
* FunFiles only has genres for movies and tv * FunFiles only has genres for movies and tv
*/ */
var ValidList = new List<string>() { var validList = new List<string>
{
"action", "action",
"adventure", "adventure",
"animation", "animation",
@ -147,25 +149,27 @@ namespace Jackett.Common.Indexers
var qc = new NameValueCollection var qc = new NameValueCollection
{ {
{"incldead", "1"}, { "cat", "0" },
{"showspam", "1"}, { "incldead", "1" },
{"cat", MapTorznabCapsToTrackers(query).FirstIfSingleOrDefault("0")} { "showspam", "1" },
{ "s_title", "1" }
}; };
var queryCats = MapTorznabCapsToTrackers(query);
queryCats.ForEach(cat => qc.Set($"c{cat}", "1"));
if (query.IsImdbQuery) if (query.IsImdbQuery)
{ {
qc.Add("search", query.ImdbID); qc.Set("search", query.ImdbID);
qc.Add("s_desc", "1"); qc.Set("s_desc", "1");
} }
else else if (query.IsGenreQuery)
if (query.IsGenreQuery)
{ {
qc.Add("search", query.Genre + " " + query.GetQueryString()); qc.Set("search", query.Genre + " " + query.GetQueryString());
qc.Add("s_tag", "1"); qc.Set("s_tag", "1");
qc.Add("s_title", "1");
} }
else else
qc.Add("search", query.GetQueryString()); qc.Set("search", query.GetQueryString());
var searchUrl = SearchUrl + "?" + qc.GetQueryString(); var searchUrl = SearchUrl + "?" + qc.GetQueryString();
var results = await RequestWithCookiesAndRetryAsync(searchUrl); var results = await RequestWithCookiesAndRetryAsync(searchUrl);
@ -176,64 +180,64 @@ namespace Jackett.Common.Indexers
results = await RequestWithCookiesAndRetryAsync(searchUrl); results = await RequestWithCookiesAndRetryAsync(searchUrl);
} }
char[] delimiters = { ',', ' ', '/', ')', '(', '.', ';', '[', ']', '"', '|', ':' };
try try
{ {
var parser = new HtmlParser(); var parser = new HtmlParser();
var dom = parser.ParseDocument(results.ContentString); var dom = parser.ParseDocument(results.ContentString);
var rows = dom.QuerySelectorAll("table[cellpadding=2] > tbody > tr:has(td.row3)");
var rows = dom.QuerySelectorAll("table.mainframe table[cellpadding=\"2\"] > tbody > tr:has(td.row3)");
foreach (var row in rows) foreach (var row in rows)
{ {
var qDownloadLink = row.QuerySelector("a[href^=\"download.php\"]"); var qDownloadLink = row.QuerySelector("a[href^=\"download.php\"]");
if (qDownloadLink == null) if (qDownloadLink == null)
throw new Exception("Download links not found. Make sure you can download from the website."); throw new Exception("Download links not found. Make sure you can download from the website.");
var link = new Uri(SiteLink + qDownloadLink.GetAttribute("href")); var link = new Uri(SiteLink + qDownloadLink.GetAttribute("href"));
var qDetailsLink = row.QuerySelector("a[href^=\"details.php?id=\"]"); var qDetailsLink = row.QuerySelector("a[href^=\"details.php?id=\"]");
var title = qDetailsLink.GetAttribute("title").Trim(); var title = qDetailsLink?.GetAttribute("title")?.Trim();
var details = new Uri(SiteLink + qDetailsLink.GetAttribute("href")); var details = new Uri(SiteLink + qDetailsLink?.GetAttribute("href")?.Replace("&hit=1", ""));
var qCatLink = row.QuerySelector("a[href^=\"browse.php?cat=\"]"); var categoryLink = row.QuerySelector("a[href^=\"browse.php?cat=\"]")?.GetAttribute("href");
var catStr = qCatLink.GetAttribute("href").Split('=')[1].Split('&')[0]; var cat = ParseUtil.GetArgumentFromQueryString(categoryLink, "cat");
var files = ParseUtil.CoerceInt(row.Children[3].TextContent);
var publishDate = DateTimeUtil.FromTimeAgo(row.Children[5].TextContent);
var size = ReleaseInfo.GetBytes(row.Children[7].TextContent);
var grabs = ParseUtil.CoerceInt(row.Children[8].TextContent);
var seeders = ParseUtil.CoerceInt(row.Children[9].TextContent); var seeders = ParseUtil.CoerceInt(row.Children[9].TextContent);
var leechers = ParseUtil.CoerceInt(row.Children[10].TextContent); var leechers = ParseUtil.CoerceInt(row.Children[10].TextContent);
var ka = row.NextElementSibling.QuerySelector("table > tbody > tr:nth-child(3)");
var ulFactor = ParseUtil.CoerceDouble(ka.Children[0].TextContent.Replace("X", ""));
var dlFactor = ParseUtil.CoerceDouble(ka.Children[1].TextContent.Replace("X", ""));
ka = row.NextElementSibling.QuerySelector("span[style=\"float:left\"]");
var description = ka.TextContent;
var qGenres = description.ToLower().Replace(" & ", "_&_").Replace(" and ", "_and_");
char[] delimiters = { ',', ' ', '/', ')', '(', '.', ';', '[', ']', '"', '|', ':' };
var releaseGenres = ValidList.Intersect(qGenres.Split(delimiters, System.StringSplitOptions.RemoveEmptyEntries));
releaseGenres = releaseGenres.Select(x => x.Replace("_", " "));
var release = new ReleaseInfo var release = new ReleaseInfo
{ {
Title = title,
Details = details,
Link = link,
Guid = link, Guid = link,
Category = MapTrackerCatToNewznab(catStr), Link = link,
PublishDate = publishDate, Details = details,
Size = size, Title = title,
Files = files, Category = MapTrackerCatToNewznab(cat),
Grabs = grabs, Size = ReleaseInfo.GetBytes(row.Children[7].TextContent),
Files = ParseUtil.CoerceInt(row.Children[3].TextContent),
Grabs = ParseUtil.CoerceInt(row.Children[8].TextContent),
Seeders = seeders, Seeders = seeders,
Peers = leechers + seeders, Peers = leechers + seeders,
Description = description, PublishDate = DateTimeUtil.FromTimeAgo(row.Children[5].TextContent),
DownloadVolumeFactor = 1,
UploadVolumeFactor = 1,
MinimumRatio = 1, MinimumRatio = 1,
MinimumSeedTime = 172800, // 48 hours MinimumSeedTime = 172800 // 48 hours
DownloadVolumeFactor = dlFactor,
UploadVolumeFactor = ulFactor
}; };
if (release.Genres == null)
release.Genres = new List<string>(); var nextRow = row.NextElementSibling;
release.Genres = releaseGenres.ToList(); if (nextRow != null)
{
var qStats = nextRow.QuerySelector("table > tbody > tr:nth-child(3)");
release.UploadVolumeFactor = ParseUtil.CoerceDouble(qStats?.Children[0].TextContent.Replace("X", ""));
release.DownloadVolumeFactor = ParseUtil.CoerceDouble(qStats?.Children[1].TextContent.Replace("X", ""));
release.Description = nextRow.QuerySelector("span[style=\"float:left\"]")?.TextContent.Trim();
var genres = release.Description.ToLower().Replace(" & ", "_&_").Replace(" and ", "_and_");
var releaseGenres = validList.Intersect(genres.Split(delimiters, StringSplitOptions.RemoveEmptyEntries));
release.Genres = releaseGenres.Select(x => x.Trim().Replace("_", " ")).ToList();
}
releases.Add(release); releases.Add(release);
} }