Personal Video Database

English => Development => Topic started by: gibbon on December 16, 2011, 05:58:54 pm

Title: Broken character encoding
Post by: gibbon on December 16, 2011, 05:58:54 pm

Hello,
thanks for the great app and especially it's Unicode support.
I've started to write an import script for japanese site (EUC-JP encoding, CP:20932) and I've noticed very strange thing.

Raw page saved automatically to file 'page.html' is OK. But when I process the page by the script (even when I output the HTML string right at the script start) it becomes broken at some places. For example:

Code: [Select]

<a href="/digital/videoa/-/detail/=/cid=41djk012/">猥らなほどに悩ましい 古都ひかる</a></p>becomes

Code: [Select]

<a href="/digital/videoa/-/detail/=/cid=41djk012/">猥らなほどに悩ましい 古都ひか・E/a></p>Which besides of changing the text, destroys the whole HTML structure.
Other examples out of many more:

Code: [Select]

～ -> ?
奥さん！ ->  ・E気鵝・
女 2 -> ・E2
For many hours I've been trying different codepages of the script (20932, autodetect, UTF8), but garbled text or errors like this one are the only results.
Has anyone had similar experience? Can't there be a bug in the script parser? Is there any way around?

Thank you very much in advance.

Title: Re: Broken character encoding
Post by: nostra on December 16, 2011, 10:52:51 pm

I think it would be easier if you provide the script itself for test.

Title: Re: Broken character encoding
Post by: gibbon on December 17, 2011, 09:55:25 am

Here is the testing version of the script. It doesn't even generate results, just writes the test files.
I used Adult DVD Empire script as a template.
It's possible that "analyze=V1EBCFcEUAc_" part of the  SEARCH_STR is generated different for each IP/user. So if the search doesn't work, please go to the website and replace it with yours unique one.
Thank you again. :)

Code: [Select]

//This version of script is for use with PVD versions 0.9.9.16 and above!!!
 (*
 Additional types and functions that can be used in scripts:

  //Types
  TWIDEARRAY : array of String

  //Field functions
  procedure AddSearchResult(Title1, Title2, Year, URL, PreviewURL : String)
  procedure AddFieldValue(AField: Integer; AValue : String)
  procedure AddMoviePerson(Name, TransName, Role, URL : String; AType : Byte)
  procedure AddPersonMovie(Title, OrigTitle, Role, Year, URL : String; AType : Byte)
  procedure AddAward(Event, Award, Category, Recipient, Year: String; const Won : Boolean)
  procedure AddConnection(Title, OrigTitle, Category, URL, Year: String)
  procedure AddEpisode(Title, OrigTitle, Description, URL, Year, Season, Episode : String)
 
  //String functions
  function Pos(Substr : String; Str: String): Integer
  function PosFrom(const SubStr, Str : String; FromIndex : Integer) : Integer
  function LastPos(const SubStr, Str : String) : Integer
  function PrevPos(const SubStr, Str : String; APos : Integer) : Integer
  function RemoveTags(AText : String; doLineBreaks : Boolean) : String
  function ExplodeString(AText : String; var Items : TWideArray; Delimiters : String) : Integer
  function Copy(S: String; Index, Count: Integer): String
  procedure Delete(var S: String; Index, Count: Integer)
  procedure Insert(Source: String; var Dest: String; Index: Integer)
  function Length(S: String): Integer
  function Trim(S: String): String
  function CompareText(S1, S2: String): Integer
  function CompareStr(S1, S2: String): Integer
  function UpperCase(S: String): String
  function LowerCase(S: String): String
  function StringReplace(S, OldPattern, NewPattern: String; ReplaceAll : Boolean; IgnoreCase : Boolean; WholeWord: Boolean): String
  function StrToInt(const S: String): Integer
  function IntToStr(const Value: Integer): String
  function StrToFloat(const S: String): Extended
  function FloatToStr(const Value: Extended): String

  function HTMLValues(const HTML : String; ABegin, AEnd, ItemBegin, ItemEnd : String; ValDelim : String; var Pos : Integer) : String
  function HTMLValues2(const HTML : String; ABegin, AEnd, ItemBegin, ItemEnd : String; ValDelim : String; var Pos : Integer) : String
  function TextBetween(const HTML : String; ABegin, AEnd : String; doLineBreaks : Boolean; var Pos : Integer) : String
  function HTMLToText(const HTML : String) : String

  procedure ShowMessage(const Msg, Head : String)
*)

const
 pauseBeforeLoad = 0; // Pause before loading (in millisecond)

//Some useful constants
const
 //Script types
 stMovies = 0;
 stPeople = 1;
 stPoster = 2;
 
 //Script modes
 smSearch = 0;
 smNormal = 1;
 smPoster = 2;
 
 //Parse results
 prError     = 0;
 prFinished  = 1;
 prList      = 2;
 prListImage = 3;
 prDownload  = 4;
 
 //Movie fields
 mfURL         = 0;
 mfTitle       = 1;
 mfOrigTitle   = 2;
 mfAka         = 3;
 mfYear        = 4;
 mfGenre       = 5;
 mfCategory    = 6;
 mfCountry     = 7;
 mfStudio      = 8;
 mfMPAA        = 9;
 mfRating      = 10;
 mfTags        = 11;
 mfTagline     = 12;
 mfDescription = 13;
 mfDuration    = 14;
 mfFeatures    = 15;
 
 //People fields
 pfURL        = 0;
 pfName       = 1;
 pfTransName  = 2;
 pfAltNames   = 3;
 pfBirthday   = 4;
 pfBirthplace = 5;
 pfGenre      = 6;
 pfBio        = 7;
 pfDeathDate  = 8;
 
 //Credits types
 ctActors    = 0;
 ctDirectors = 1;
 ctWriters   = 2;
 ctComposers = 3;
 ctProducers = 4;
 
 //Script data
 SCRIPT_VERSION = '0.0.0.4';
 SCRIPT_NAME    = 'DMM.co.jp';
 SCRIPT_DESC    = '[EN] Get movie information DMM.co.jp';
 SCRIPT_LANG    = $11; //Japanese  //Tested both English & Japanese
 SCRIPT_TYPE    = stMovies;
 
 BASE_URL    = 'http://www.dmm.co.jp';
 RATING_NAME = 'DMM';
 
 SEARCH_STR  = 'http://www.dmm.co.jp/search/=/searchstr=%s/analyze=V1EBCFcEUAc_/limit=30/sort=rank_asc/view=text/num=1/';
 
 CODE_PAGE   = 20932; // Tested: 0, 33722, 51932, 20932, 65001

 
//Global variables
var
 Mode : Byte;
 PosterURL : String;

//Functions
function GetScriptVersion : String;
begin
 Result := SCRIPT_VERSION;
end;
 
function GetScriptName : String;
begin
 Result :=  SCRIPT_NAME;
end;

function GetScriptDesc : String;
begin
 Result := SCRIPT_DESC;
end;

function GetRatingName : String;
begin
 Result := RATING_NAME;
end;

function GetScriptLang: Cardinal;
begin
 Result := SCRIPT_LANG;
end;

function GetCodePage : Cardinal;
begin
 Result := CODE_PAGE;
end;

function GetBaseURL : AnsiString;
begin
 Result := BASE_URL;
end;

function GetDownloadURL : AnsiString;
begin
 if PosterURL = '' then
  Result := SEARCH_STR
 else
  Result := PosterURL;
end;

function GetScriptType : Byte;
begin
 Result := SCRIPT_TYPE;
end;

function GetCurrentMode : Byte;
begin
 Result := Mode;
end;

procedure FindPoster(HTML : String);
var
 curPos, EndPos : Integer;
begin
// not yet implemented
end;
 
//  ******************  ParseMovie
procedure ParseMovie(MovieURL : String; HTML : String);
var
 curPos, EndPos, P, P2, L : Integer;
 Tmp, URL, Name : String;
begin
 AddFieldValue(mfURL, MovieURL);
// not yet implemented
end;

//  ******************  ParseSearchResults
procedure ParseSearchResults(HTML : String);
var
 curPos, EndPos, P : Integer;
 Title, URL, Tabulka : String;
begin
 curPos := PosFrom('<table summary="', HTML, 1); //beginning of results table
 EndPos := PosFrom('</table>', HTML, curPos); //end of results table
 Tabulka := Copy(HTML, curPos, EndPos - curPos);
 
 StringToFile('wholepage.htm', HTML, false ,false); //debug - output test
 StringToFile('tabulka.htm', Tabulka, false ,false); //debug - output test
// LogMessage('CodePage:' + IntToStr(GetCodePage)); 
 
end;

//  ******************  ParsePage
function ParsePage(HTML : String; URL : AnsiString) : Cardinal;
begin
 Wait (pauseBeforeLoad);
 //HTML := ConvertEncoding(HTML, 20932);
 StringToFile('start.htm', HTML, false ,false); //debug - output test
 
 if Pos('<div class="othertxt">', HTML) > 0 then begin
  LogMessage('Nothing Found!'); //debug
  Result := prError
 end else
 if Pos('<th scope="col">', HTML) > 0 then begin
  LogMessage('Calling ParseSearchResults!'); //debug
  ParseSearchResults(HTML);
  Result := prList;
 end else
 if Pos('<h1 id="title"', HTML) > 0 then begin
  LogMessage('Product page!'); //debug
  FindPoster(HTML);
  Mode := smPoster;
  if PosterURL <> '' then
   Result := prDownload
  else
   Result := prFinished; 
 end else begin
  ParseMovie(URL, HTML);
  Mode := smNormal;
  if PosterURL = '' then
   Result := prFinished
  else
   Result := prDownload; 
 end;
end;

begin
 Mode := smSearch;
end.


Title: Re: Broken character encoding
Post by: nostra on December 22, 2011, 08:43:01 pm

Are you sure that the actual data in memory is like this. I have investigated the problem and it looks correct in memory, but gets scrambled when saving to file (probably because the encoding information is lost).
Try to parse a page and return some data as result to PVD. Would the fields be filled with correct strings?

Title: Re: Broken character encoding
Post by: gibbon on December 27, 2011, 05:19:40 pm

Quote from: nostra on December 22, 2011, 08:43:01 pm

Are you sure that the actual data in memory is like this. I have investigated the problem and it looks correct in memory, but gets scrambled when saving to file (probably because the encoding information is lost).
Try to parse a page and return some data as result to PVD. Would the fields be filled with correct strings?

Unfortunately they're broken as well. That's how I noticed. File output was just for debugging purpose.
Please find the attached picture with PVD 'Select Movie' window. The checked lines have some broken characters.

This is the log. You can see the encoding errors broke the parsing logic:

Code: [Select]

(12/27/2011 5:07:47 PM) UpdateToolbar: 6
(12/27/2011 5:07:47 PM) UpdateToolbar: 7
(12/27/2011 5:07:56 PM) Compiling script: dmm_0_0_4.psf
(12/27/2011 5:07:56 PM) Script compiled successfully: dmm_0_0_4.psf
[Hint] (381:2): Variable 'P' never used
(12/27/2011 5:07:56 PM) Executing script binary
(12/27/2011 5:07:56 PM) Logging in...
(12/27/2011 5:07:56 PM) Searching movie information for: djk
(12/27/2011 5:07:56 PM) GET: http://www.dmm.co.jp/search/=/searchstr=djk/analyze=V1EBCFcEUAc_/limit=30/sort=rank_asc/view=text/num=1/
(12/27/2011 5:07:58 PM) start!
(12/27/2011 5:07:58 PM) ParseSearchResults! !
(12/27/2011 5:07:58 PM) CodePage:0
(12/27/2011 5:07:58 PM) URL: http://www.dmm.co.jp/digital/m_full/-/detail/=/cid=djk11/
(12/27/2011 5:07:58 PM) Title: 少女の道・E11
(12/27/2011 5:07:58 PM) URL: http://www.dmm.co.jp/digital/m_full/-/detail/=/cid=djk18/
(12/27/2011 5:07:58 PM) Title: 少女の道・E18
(12/27/2011 5:07:58 PM) URL: http://www.dmm.co.jp/digital/m_full/-/detail/=/cid=djk19/
(12/27/2011 5:07:58 PM) Title: 少女の道・E19
(12/27/2011 5:07:58 PM) URL: http://www.dmm.co.jp/digital/m_full/-/detail/=/cid=djk07/
(12/27/2011 5:07:58 PM) Title: 少女の道・E7
(12/27/2011 5:07:58 PM) URL: http://www.dmm.co.jp/digital/m_full/-/detail/=/cid=djk14/
(12/27/2011 5:07:58 PM) Title: 少女の道・E14
(12/27/2011 5:07:58 PM) URL: http://www.dmm.co.jp/digital/m_full/-/detail/=/cid=djk10/
(12/27/2011 5:07:58 PM) Title: 少女の道・E10
(12/27/2011 5:07:58 PM) URL: http://www.dmm.co.jp/digital/m_full/-/detail/=/cid=djk09/
(12/27/2011 5:07:58 PM) Title: 少女の道・E9
(12/27/2011 5:07:58 PM) URL: http://www.dmm.co.jp/digital/videoa/-/detail/=/cid=51djk009/
(12/27/2011 5:07:58 PM) Title: 猥E熟・E8 岡・EE/a></p>
<p class="status">
<span class="ico-st-monopoly"><span>独・E/span></span>
<!--/status--></p>
</td>

<td><a href="/search/=/searchstr=djk/limit=30/n1=FgRCTw9VBA4GAVhfWkIHWw__/n2=Aw1fVhQKX1ZRAlhMUlo5QQgBU1lR/sort=rank_asc/view=text/num=1/">ビデオ
(12/27/2011 5:07:58 PM) URL: http://www.dmm.co.jp/monthly/mania/-/detail/=/cid=51djk009/
(12/27/2011 5:07:58 PM) Title: 猥E熟・E8 岡・EE/a>
</p></td>

<td><a href="/search/=/searchstr=djk/limit=30/n1=FgRCTw9VBA4GCF5WR14KTg__/n2=Aw1fVhQKX19XC0VQX085WgALX1c_/sort=rank_asc/view=text/num=1/">マニア
(12/27/2011 5:07:59 PM) URL: http://www.dmm.co.jp/digital/videoa/-/detail/=/cid=djks01/
(12/27/2011 5:07:59 PM) Title: 渋谷女子校生 少女の道・E8時間
(12/27/2011 5:07:59 PM) URL: http://www.dmm.co.jp/digital/videoa/-/detail/=/cid=h_275tdjk00016/
(12/27/2011 5:07:59 PM) Title: 宅配露出 2
(12/27/2011 5:07:59 PM) URL: http://www.dmm.co.jp/digital/videoa/-/detail/=/cid=h_275tdjk00004/
(12/27/2011 5:07:59 PM) Title: 宅配露出
(12/27/2011 5:07:59 PM) URL: http://www.dmm.co.jp/digital/videoa/-/detail/=/cid=h_275tdjk00015/
(12/27/2011 5:07:59 PM) Title: ・・E灰鷯・憤E中・E弧穏泪Eぅ芛/a></p>
</td>

<td><a href="/search/=/searchstr=djk/limit=30/n1=FgRCTw9VBA4GAVhfWkIHWw__/n2=Aw1fVhQKX1ZRAlhMUlo5QQgBU1lR/sort=rank_asc/view=text/num=1/">ビデオ
(12/27/2011 5:07:59 PM) URL: http://www.dmm.co.jp/digital/videoa/-/detail/=/cid=h_157djk010/
(12/27/2011 5:07:59 PM) Title: 女子校生見せつけオナニー
(12/27/2011 5:07:59 PM) URL: http://www.dmm.co.jp/monthly/mania/-/detail/=/cid=h_157djk010/
(12/27/2011 5:07:59 PM) Title: 女子校生見せつけオナニー
(12/27/2011 5:07:59 PM) URL: http://www.dmm.co.jp/digital/videoa/-/detail/=/cid=41djk012/
(12/27/2011 5:07:59 PM) Title: 猥らなほどに悩ましい 古都ひか・E/a></p>
</td>

<td><a href="/search/=/searchstr=djk/limit=30/n1=FgRCTw9VBA4GAVhfWkIHWw__/n2=Aw1fVhQKX1ZRAlhMUlo5QQgBU1lR/sort=rank_asc/view=text/num=1/">ビデオ
(12/27/2011 5:07:59 PM) URL: http://www.dmm.co.jp/monthly/hmp/-/detail/=/cid=41djk012/
(12/27/2011 5:07:59 PM) Title: 猥らなほどに悩ましい 古都ひか・E/a>
</p></td>

<td><a href="/search/=/searchstr=djk/limit=30/n1=FgRCTw9VBA4GCF5WR14KTg__/n2=Aw1fVhQKX19XC0VQX085XwwV/sort=rank_asc/view=text/num=1/">h.m.p
(12/27/2011 5:07:59 PM) URL: http://www.dmm.co.jp/digital/videoa/-/detail/=/cid=djks03/
(12/27/2011 5:07:59 PM) Title: 渋谷女子校生 少女の道・E8時間 3
(12/27/2011 5:07:59 PM) URL: http://www.dmm.co.jp/digital/videoa/-/detail/=/cid=41hodv00211/
(12/27/2011 5:07:59 PM) Title: ・Eぅ廚気E拭・古都ひか・E/a></p>
</td>

<td><a href="/search/=/searchstr=djk/limit=30/n1=FgRCTw9VBA4GAVhfWkIHWw__/n2=Aw1fVhQKX1ZRAlhMUlo5QQgBU1lR/sort=rank_asc/view=text/num=1/">ビデオ
(12/27/2011 5:07:59 PM) URL: http://www.dmm.co.jp/monthly/hmp/-/detail/=/cid=41hodv00211/
(12/27/2011 5:07:59 PM) Title: ・Eぅ廚気E拭・古都ひか・E/a>
</p></td>

<td><a href="/search/=/searchstr=djk/limit=30/n1=FgRCTw9VBA4GCF5WR14KTg__/n2=Aw1fVhQKX19XC0VQX085XwwV/sort=rank_asc/view=text/num=1/">h.m.p
(12/27/2011 5:07:59 PM) URL: http://www.dmm.co.jp/digital/videoa/-/detail/=/cid=djks02/
(12/27/2011 5:07:59 PM) Title: 渋谷女子校生 少女の道・E8時間 2
(12/27/2011 5:07:59 PM) URL: http://www.dmm.co.jp/mono/dvd/-/detail/=/cid=djks02/
(12/27/2011 5:07:59 PM) Title: 渋谷女子校生 少女の道・E8時間 2
(12/27/2011 5:07:59 PM) URL: http://www.dmm.co.jp/mono/dvd/-/detail/=/cid=djks01/
(12/27/2011 5:07:59 PM) Title: 渋谷女子校生 少女の道・E8時間
(12/27/2011 5:07:59 PM) URL: http://www.dmm.co.jp/mono/dvd/-/detail/=/cid=h_275tdjk16/
(12/27/2011 5:07:59 PM) Title: 宅配露出 2
(12/27/2011 5:07:59 PM) URL: http://www.dmm.co.jp/mono/dvd/-/detail/=/cid=10dnsh001/
(12/27/2011 5:07:59 PM) Title: パンチラ☆ちら見電車 〜チラチラ片道切符〜
(12/27/2011 5:07:59 PM) URL: http://www.dmm.co.jp/monthly/mania/-/detail/=/cid=29djkc04/
(12/27/2011 5:07:59 PM) Title: 痴女×M・E2
(12/27/2011 5:07:59 PM) URL: http://www.dmm.co.jp/monthly/mania/-/detail/=/cid=29djkc01/
(12/27/2011 5:07:59 PM) Title: 痴女×M・E/a>
</p></td>

<td><a href="/search/=/searchstr=djk/limit=30/n1=FgRCTw9VBA4GCF5WR14KTg__/n2=Aw1fVhQKX19XC0VQX085WgALX1c_/sort=rank_asc/view=text/num=1/">マニア
(12/27/2011 5:07:59 PM) URL: http://www.dmm.co.jp/monthly/mania/-/detail/=/cid=29djkb01/
(12/27/2011 5:07:59 PM) Title: 爆乳痴態娘 1 水沢ダイア
(12/27/2011 5:07:59 PM) URL: http://www.dmm.co.jp/monthly/mania/-/detail/=/cid=29djkb02/
(12/27/2011 5:07:59 PM) Title: 爆乳痴態娘 2 真咲菜々
(12/27/2011 5:07:59 PM) URL: http://www.dmm.co.jp/monthly/mania/-/detail/=/cid=29djkj01/
(12/27/2011 5:07:59 PM) Title: 痴的女教師

This is the parsed results page:

Code: [Select]

http://www.dmm.co.jp/search/=/searchstr=djk/analyze=V1EBCFcEUAc_/limit=30/sort=rank_asc/view=text/num=1/
And finally this is the code of :

Code: [Select]

procedure ParseSearchResults(HTML : String);
var
 curPos, EndPos, P : Integer;
 Title, URL : String;
begin
 EndPos := 1;
 
 LogMessage('CodePage:' + IntToStr(GetCodePage)); 
 curPos := PosFrom('<table summary="', HTML, EndPos);
 EndPos := curPos + Length('<table summary="');
 while curPos > 0 do begin
  //EndPos := curPos + Length('<table summary="');

  curPos := PosFrom('<td><p class="ttl">', HTML, EndPos);
  if curPos > 0 then begin
   curPos := PosFrom('<a href="', HTML, EndPos);
   if curPos > 0 then begin
   curPos := curPos + Length('<a href="');
   EndPos := PosFrom('">', HTML, curPos);
   URL := BASE_URL + Copy(HTML, curPos, EndPos - curPos);
   LogMessage('URL: ' + URL); //debug
  curPos := EndPos + 2;
  EndPos := PosFrom('</a>', HTML, curPos);
  Title := Copy(HTML, curPos, EndPos - curPos);
  EndPos := PosFrom('<td>', HTML, EndPos) + 2;
  EndPos := PosFrom('<td>', HTML, EndPos) + 2;
  end;
   end;

   LogMessage('Title: ' + Title); //debug
   AddSearchResult(Title, '', '', URL, '');
   curPos := PosFrom('<p class="ttl"', HTML, EndPos);

 end;
end;
Thank you for further investigating of the problem.


[attachment deleted by admin]

Title: Re: Broken character encoding
Post by: nostra on December 27, 2011, 05:44:57 pm

Could you, please, check if the problem persists with the current beta version? (1.0.2.2)

Title: Re: Broken character encoding
Post by: gibbon on December 27, 2011, 07:04:36 pm

Quote from: nostra on December 27, 2011, 05:44:57 pm

Could you, please, check if the problem persists with the current beta version? (1.0.2.2)

Yes, it does. :( In 1.0.2.2 beta exactly the same problem.