From 1404afdfa38234b24314f5b14c14ef798a7adcda Mon Sep 17 00:00:00 2001 From: Andy <88590076+AAndyProgram@users.noreply.github.com> Date: Mon, 1 Sep 2025 16:41:44 +0300 Subject: [PATCH] 2025.9.1.0 API.PornHub: update regex and data parsing --- Changelog.md | 7 +++++++ SCrawler/API/PornHub/Declarations.vb | 2 +- SCrawler/API/PornHub/UserData.vb | 29 +++++++++++++++++++++------- SCrawler/My Project/AssemblyInfo.vb | 4 ++-- 4 files changed, 32 insertions(+), 10 deletions(-) diff --git a/Changelog.md b/Changelog.md index cd0a5bd..d8667bb 100644 --- a/Changelog.md +++ b/Changelog.md @@ -1,3 +1,10 @@ +# 2025.9.1.0 + +*2025-09-01* + +- Fixed + - PornHub: data is not downloading + # 2025.8.30.0 *2025-08-30* diff --git a/SCrawler/API/PornHub/Declarations.vb b/SCrawler/API/PornHub/Declarations.vb index d1993a6..7068542 100644 --- a/SCrawler/API/PornHub/Declarations.vb +++ b/SCrawler/API/PornHub/Declarations.vb @@ -20,7 +20,7 @@ Namespace API.PornHub Friend ReadOnly RegexVideo_FlashVars_Vars As RParams = RParams.DM("var ([\w\d]{10,})=("".+?)(?=(;|\Z))", 0, RegexReturn.List) Friend ReadOnly RegexVideo_FlashVars_Compiler As RParams = RParams.DM("(?<=\*/)([\w\d\S]{10,})", 0, RegexReturn.List) Friend ReadOnly RegexVideo_FlashVars_UrlResolution As RParams = RParams.DMS("/(\d+)[^/]+\.mp4", 1, EDP.ReturnValue) - Friend ReadOnly RegexUserVideos As RParams = RParams.DM("(\
  • ).)*?)(\).)*?)|)(\).)+?)(\
    ).)*?)(\)", + Friend ReadOnly RegexUserVideos As RParams = RParams.DM("(\
  • ).)*?)(\).)*?)|)(\).)+?)(\
    ).)*?)(\)", 0, RegexOptions.Singleline, RegexReturn.List, EDP.ReturnValue, UnicodeHexConverter) Friend ReadOnly RegexVideo_Video_VideoKey As RParams = RParams.DMS("viewkey=([\w\d]+)", 1, EDP.ReturnValue) Friend ReadOnly RegexVideoPageTitle As RParams = RParams.DMS("meta (property|name)=""[^:]+?:title"" content=""([^""]+)""", 2, EDP.ReturnValue) diff --git a/SCrawler/API/PornHub/UserData.vb b/SCrawler/API/PornHub/UserData.vb index 82709f8..1848a38 100644 --- a/SCrawler/API/PornHub/UserData.vb +++ b/SCrawler/API/PornHub/UserData.vb @@ -50,6 +50,7 @@ Namespace API.PornHub Friend URL As String Friend ID As String Friend Title As String + Friend UserRef As String Friend Type As VideoTypes Friend Function ToUserMedia(Optional ByVal SpecialFolder As String = Nothing) As UserMedia Return New UserMedia(URL, UTypes.VideoPre) With { @@ -66,14 +67,16 @@ Namespace API.PornHub URL = String.Empty Else URL = String.Format(UrlPattern, URL.TrimStart("/")) + Title = TitleHtmlConverter(ParamsArray(1)) If Not ParamsArray(2).IsEmptyString Then Type = VideoTypes.Private - ElseIf Not ParamsArray(3).IsEmptyString Then - Type = VideoTypes.Tagged + 'ElseIf Not ParamsArray(3).IsEmptyString Then + ' Type = VideoTypes.Tagged Else Type = VideoTypes.Uploaded End If + If Not ParamsArray(3).IsEmptyString Then UserRef = ParamsArray(3).StringTrim End If End If Return Me @@ -338,10 +341,13 @@ Namespace API.PornHub Dim tryNextPage As Boolean = False Dim limit% = If(DownloadTopCount, -1) Dim cBefore% = _TempMediaList.Count + Dim usrRef$ = String.Empty + Dim npd$ = "?" If IsUser Then URL = $"https://www.pornhub.com/{PersonType}/{NameTrue}" + usrRef = $"/{PersonType}/{NameTrue}" If Type = VideoTypes.Uploaded Then - URL &= "/videos/upload" + If Not PersonType = PersonTypeCannel Then URL &= "/videos/upload?o=mr" : npd = "&" ElseIf Type = VideoTypes.Tagged Then If Not SecondMode Then URL &= "/videos" specFolder = "Tagged" @@ -354,7 +360,7 @@ Namespace API.PornHub Else Throw New ArgumentException($"Type '{Type}' is not implemented in the video download function", "Type") End If - If Page > 1 Then URL &= $"?page={Page}" + If Page > 1 Then URL &= $"{npd}page={Page}" ElseIf SiteMode = SiteModes.Playlists Then If PlaylistToken.IsEmptyString Then Throw New ArgumentNullException("PlaylistToken", "Unable to get 'PlaylistToken'") URL = String.Format(PlayListUrlPattern, NameTrue, PlaylistToken, Page) @@ -367,12 +373,21 @@ Namespace API.PornHub 'Debug.WriteLine(URL) Dim r$ = Responser.GetResponse(URL) If Not r.IsEmptyString Then - Dim l As List(Of UserVideo) = RegexFields(Of UserVideo)(r, {RegexUserVideos}, {6, 7, 3, 10}) + Dim l As List(Of UserVideo) = RegexFields(Of UserVideo)(r, {RegexUserVideos}, {6, 7, 3, 11}) 'If l.ListExists And Not SiteMode = SiteModes.Playlists Then l = l.ListTake(3, l.Count).ToList - If l.ListExists And Not SiteMode = SiteModes.Playlists Then l = l.ListTake(1, l.Count).ToList + If l.ListExists And Not SiteMode = SiteModes.Playlists And Not IsUser Then l = l.ListTake(1, l.Count).ToList If l.ListExists Then If IsUser Then - If Type = VideoTypes.Favorite Then + If Type = VideoTypes.Tagged Then + l = l.ListTake(4, l.Count) + If l.ListExists Then l.RemoveAll(Function(uv) Not uv.UserRef.IsEmptyString AndAlso uv.UserRef = usrRef) + ElseIf Type = VideoTypes.Uploaded Then + If PersonType = PersonTypeCannel Then + l = l.ListTake(4, l.Count) + Else + l.RemoveAll(Function(uv) uv.UserRef.IsEmptyString OrElse Not uv.UserRef = usrRef) + End If + ElseIf Type = VideoTypes.Favorite Then l.RemoveAll(Function(uv) uv.Type = VideoTypes.Private) ElseIf Not PersonType = PersonTypeCannel Then l.RemoveAll(Function(uv) Not uv.Type = Type) diff --git a/SCrawler/My Project/AssemblyInfo.vb b/SCrawler/My Project/AssemblyInfo.vb index 8684323..485afcc 100644 --- a/SCrawler/My Project/AssemblyInfo.vb +++ b/SCrawler/My Project/AssemblyInfo.vb @@ -32,6 +32,6 @@ Imports System.Runtime.InteropServices ' by using the '*' as shown below: ' - - + +