antp.be forum

Code: Select all

(***************************************************

Ant Movie Catalog importation script
www.antp.be/software/moviecatalog/

[Infos]
Authors=Legrad
Title=HoyCinema(ES)
Description=
Site=www.hoyCinema.com
Language=ES
Version=1.1
Requires=3.5.0
Comments=
License=
GetInfo=1

[Options]

***************************************************)

 program HoyCinema;
var 
  MovieName: string; 
  MovieURL: string; 
//------------------------------------------------------------------------------------ 
function BorraComillas(var S: string): string; 
var 
   n,len, tag: Integer; 
   c: char; 
   t: String; 
begin 
   tag := 0; 
   t := ''; 
   len := length(s); 
   for n :=1 to len do 
   begin 
     c := Copy(s,n,1); 
     if c = '''' then 
        c := ' '; 
        t := t + c; 
   end 
   s := t; 
   result := t; 
end; 

//------------------------------------------------------------------ 
                              function FindLine(Pattern: string; List: TStringList; StartAt: Integer): Integer; 
var 
  i: Integer; 
begin 
  Result := -1; 
  if StartAt < 0 then 
    StartAt := 0; 
  for i := StartAt to List.Count-1 do 
    if Pos(Pattern, List.GetString(i)) <> 0 then 
    begin 
      Result := i; 
      Break; 
    end; 
end; 
//------------------------------------------------------------------------------------ 
                      function TextBetween(var S: string; StartTag: string; EndTag: string): string; 
var 
  InitialPos: Integer; 
begin 
  InitialPos := Pos(StartTag, S); 
  if InitialPos = 0 then 
    result := '' 
  else 
  begin 
    Delete(S, 1, InitialPos + Length(StartTag) - 1); 
    InitialPos := Pos(EndTag, S); 
    if InitialPos = 0 then 
      result := S 
    else 
    begin 
      result := copy(S, 1, InitialPos - 1); 
      Delete(S, 1, InitialPos + 1); 
    end; 
  end; 
end; 

//------------------------------------------------------------------------------------ 
                               function DeleteTags(var S: string): string; 
var 
   n,len, tag: Integer; 
   c: char; 
   t: String; 
begin 

   tag := 0; 
   t := ''; 
   len := length(s); 

   for n :=1 to len do 
   begin 
      c := Copy(s,n,1); 


      if c = #9 then 
         c := ' '; 

      if(tag=1) then 
      begin 
         if(c='>') then tag := 0; 
         continue; 
      end 
      else 
      begin 
         if(c='<') then 
         begin 
            tag := 1; 
            continue; 
         end; 
         t := t + c; 
      end; 
   end 
   s := t; 
   result := t; 
end; 
//------------------------------------------------------------------------------------
function UpFirstLetterWord(texto:string):string;
var espaco:integer;
sst:string;
begin
texto:=AnsiUpFirstLetter(AnsiLowerCase(texto));
repeat
    espaco:=Pos(' ',texto);
    sst:=AnsiUpperCase(Copy(texto,espaco+1,1));

texto:=Copy(texto,1,espaco-1)+'/|\'+sst+Copy(texto,espaco+2,length(texto));
until Pos(' ',texto)=0;
texto := StringReplace(texto, '/|\', ' ');
if Copy(texto,1,1)=' ' then
  texto:=Copy(texto,2,length(texto));
result:=texto;
end;

//------------------------------------------------------------------------------------ 
procedure AnalyzePage(Address: string); 
var 
  strPage, MovieAddr, MovieTitle, MovieDate, MovieID, Movie: string; 
  BeginPos, EndPos: Integer; 
  BeginPoss, EndPoss: Integer; 
begin 
  strPage := GetPage(Address); 
  BeginPos := Pos('en el dominio <b>www.hoycinema.com<', strPage);
  if(BeginPos > -1)then 
    begin 
      PickTreeClear; 
      Delete(strPage, 1, BeginPos); 
      BeginPos := Pos('href="http://www.hoycinema.com/sinopsis/', strPage);
      EndPos := 1; 
      while ((BeginPos > 0) and (EndPos > 0)) do 

        begin 
          Delete(strPage, 1, BeginPos); 
          EndPos := Pos('" class', strPage);
          MovieId := Copy(strPage,+40, EndPos-40);
          MovieAddr := 'http://www.hoycinema.com/sinopsis/' + MovieId;
          BeginPoss := Pos(')"',strPage);
          EndPoss := Pos(' -', strPage);
          MovieTitle := Copy(strPage,BeginPoss, EndPoss);
          MovieTitle  := TextBetween(MovieTitle , '>', '-');
          MovieTitle  := StringReplace(MovieTitle , 'á', 'á');
          MovieTitle  := StringReplace(MovieTitle , 'é', 'é');
          MovieTitle  := StringReplace(MovieTitle , 'í', 'í');
          MovieTitle  := StringReplace(MovieTitle , 'ó', 'ó');
          MovieTitle  := StringReplace(MovieTitle , 'ú', 'ú');
          MovieTitle  := StringReplace(MovieTitle , 'Â', '');
          MovieTitle  := StringReplace(MovieTitle , 'ñ', 'ñ');
          DeleteTags(MovieTitle);
          PickTreeAdd(MovieTitle, MovieAddr);
          BeginPos := Pos('href="http://www.hoycinema.com/sinopsis/', strPage);
          if(Pos('</body>', strPage) < BeginPos) then 
           BeginPos := -1; 
           PickTreeSort;
        end; 

    end; 
    PickTreeExec(Address);
    AnalyzeMoviePage(Address); 
    AnalyzeVideoPage(Address);
end;
//------------------------------------------------------------------------------------
procedure AnalyzeMoviePage(Address: string);
var
  Page: TStringList;
  LineNr: Integer;
  Line: string;
  Item: string;
  Comments: string;
  Actors: string;
  Directors: string;
  Description: string;
  Busca: integer;
  begin
   //URL
  SetField(fieldURL, Address);
  Page := TStringList.Create;
  Page.Text := GetPage(Address);
  // sinopsis
 LineNr := FindLine('<span id="esp_sup__ctl0_sinopsis_lbl">', Page, 0);
  if LineNr > 0 then
  begin
    Item := copy(Page.Text, pos('<span id="esp_sup__ctl0_sinopsis_lbl">',Page.Text), length(Page.Text));
    Item := TextBetween (Item, '<span id="esp_sup__ctl0_sinopsis_lbl">', '</td>');
    Item := StringReplace(Item , #13#10, '');
    DeleteTags (Item);
    Item  := Trim(Item );
    HTMLDecode(Item);
    SetField(fieldDescription, Trim (Item));
  end;
  end;
//------------------------------------------------------------------------------------ 
procedure AnalyzeVideoPage(Address: string);
var 
  Page: TStringList; 
  LineNr: Integer; 
  Line: string; 
  Item: string; 
  Comments: string; 
  Actors: string; 
  Directors: string; 
  Description: string;
  Busca: integer;


begin 
  Description := ''; 

  // URL 
  SetField(fieldURL, Address); 
  Page := TStringList.Create;
  Address := StringReplace (Address,'sinopsis/','');
  Page.Text := GetPage(Address); 

  // Titulo traducido 
  LineNr := FindLine('<TITLE>', Page, 0);
  Line := Page.GetString(LineNr); 
  Item := TextBetween (Line, '<TITLE>', '-');
  Item  := UpFirstLetterWord(Item);
  HTMLDecode(Item); 
  SetField(fieldTranslatedTitle, Trim (Item)); 

  // año 
  LineNr := FindLine('<strong>· Año:', Page, 0);
  if LineNr > 0 then
  begin
    Item := copy(Page.Text, pos('<strong>· Año:',Page.Text), length(Page.Text));
    Item := TextBetween (Item, '"contenido_txt">', '</a>');
    Item := StringReplace(Item , #13#10, '');
    DeleteTags (Item);
    Item   := AnsiUpFirstLetter(Item );
    Item  := Trim(Item );
    HTMLDecode(Item);
    SetField(fieldYear, Trim (Item)); 
  end; 
  // Categoria 
  LineNr := FindLine('<strong>· Género: ', Page, 0);
  if LineNr > 0 then 
  begin 
    Item := copy(Page.Text, pos('<strong>· Género: ',Page.Text), length(Page.Text));
    Item := TextBetween (Item, '"contenido_txt">', '</a>');
    Item := StringReplace(Item , #13#10, ''); 
    Item  := StringReplace(Item , '   ', ''); 
    DeleteTags (Item); 
    Item   := AnsiUpFirstLetter(Item ); 
    Item  := Trim(Item ); 
    HTMLDecode(Item); 
    SetField(fieldCategory, Trim (Item)); 
  end; 
   // pais 
  LineNr := FindLine('<strong>· País:', Page, 0);
  if LineNr > 0 then 
  begin 
    Item := copy(Page.Text, pos('<strong>· País:',Page.Text), length(Page.Text));
    Item := TextBetween (Item, 'ctl0_pais_lbl">', '</span>');
    Item := StringReplace(Item , #13#10, ''); 
    Item  := StringReplace(Item , '   ', ''); 
    DeleteTags (Item); 
    Item   := AnsiUpFirstLetter(Item ); 
    Item  := Trim(Item ); 
    HTMLDecode(Item); 
    SetField(fieldCountry, Trim (Item)); 
  end; 
   // director 
  LineNr := FindLine('<strong>· Director:', Page, 0);
  if LineNr > 0 then 
  begin 
    Item := copy(Page.Text, pos('<strong>· Director:',Page.Text), length(Page.Text));
    Item := TextBetween (Item, '"contenido_txt">', '</a>');
    Item := StringReplace(Item , #13#10, ''); 
    DeleteTags (Item); 
    Item   := AnsiUpFirstLetter(Item ); 
    Item  := Trim(Item ); 
    HTMLDecode(Item); 
    SetField(fieldDirector, Trim (Item)); 
  end;
  // reparto
  LineNr := FindLine('Ficha artística</td>', Page, 0);
  if LineNr > 0 then
  begin
    Item := copy(Page.Text, pos('Ficha artística</td>',Page.Text), length(Page.Text));
    Item := TextBetween (Item, 'htm">·', '</table>');
    Item := StringReplace(Item , #13#10, '');
    Item := StringReplace(Item ,'      ', '');
    Item := StringReplace(Item ,'·', ', ');
    DeleteTags (Item);
    Item   := AnsiUpFirstLetter(Item );
    Item  := Trim(Item );
    HTMLDecode(Item);
    Item  := StringReplace(Item , '      ,', ',');
    SetField(fieldActors, Trim (Item));
  end;
 // guion
 LineNr := FindLine('<span id="tdsupizq__ctl0_guion_lbl"><a', Page, 0);
  if LineNr > 0 then 
  begin 
    Item := copy(Page.Text, pos('<span id="tdsupizq__ctl0_guion_lbl"><a',Page.Text), length(Page.Text));
    Item := TextBetween (Item, 'class="contenido_txt" >', '</span></td>');
    Item := StringReplace(Item ,'</a>', ', ');
    DeleteTags (Item); 
    Item  := Trim(Item ); 
    HTMLDecode(Item); 
    SetField(fieldComments, Trim ('Guión: '+Item));
  end; 
  // Caratula
  LineNr := FindLine('src="http://media.hoycinema.com/pic.aspx?img=', Page, 0);
  if LineNr <> -1 then
  begin
     Line := Page.GetString(LineNr);
     Item := TextBetween (Line, 'src="', '.jpg');
     GetPicture (Item+'.jpg');
     end;
end;
 //-------------------------------------------------------------------------
begin 
       if (CheckVersion(3,5,0)=FALSe) then 
   begin 
      ShowMessage('Se requiere Ant Movie Catalog versión 3.5 o superior'); 
      exit; 
   end; 

   MovieName := GetField(fieldTranslatedTitle); 
   if MovieName = '' then 
            MovieName := GetField(fieldOriginalTitle); 
Input('HoyCinema', 'Buscar:', MovieName); 

     if(GetOption('Sin resultado') = 0) then  Input('HoyCinema', 'Buscar:', MovieName); 

   AnalyzePage('http://www.google.es/search?num=100&hl=es&as_qdr=all&q=' + UrlEncode(MovieName)+'+site%3Awww.hoycinema.com&btnG=B%C3%BAsqueda&meta=');
end.