Another improvement to the IMDB (large pic) script

[jwm]

Hey, thx. Great improvements!

Now there is one more thing to resolve: sometimes the script finds a picture, but it is the picture of a different movie.

Take "Song jia huang chao" as an example. The imdb link takes you to an Amazon page where the movie is found, but no picture available. However, Amazon proposes 'closely related' movies and the script takes the first picture of the related movies! What it should do is return empty because this movie doesn't have a picture.

I would appreciate your efforts to look into this.

[Ork]

You forgot to turn to uppercase the article "a".

[trekkie]

Ork Silly me .you're wright.

jwm

Ok . Here is the latest code. only AnalyzeMoviePage section

Code: Select all

[Hawkie]

I keep getting small pictures when I use your script. for example, for the movie "Auto Focus", when I use your script I get a small pictures with a women silluette, but when I use the normal, large picture script I get a LARGE picture with a men in an uniform..
What's wrong ?

[Ork]

Seems like the "Made for video" and "TV series" sections should be swapped.
I think all IMDb scripts have this error.

Code: Select all

    LineNr := FindLine('<H2><A NAME="vid">Made for video</A></H2>', Page, LineNr);
    if LineNr > -1 then
    begin
      PickTreeAdd('Made for video', '');
      AddMoviesTitles(Page, LineNr);
    end;
    LineNr := FindLine('<H2><A NAME="tvs">TV series</A></H2>', Page, LineNr);
    if LineNr > -1 then
    begin
      PickTreeAdd('TV Series', '');
      AddMoviesTitles(Page, LineNr);
    end;

[antp]

Which error
The two are not in the same order on the IMDB page than in the script ?
If it is the case, you are right, they must be swapped to make that work (since LineNr is passed in parameters, it does not restart the search from the top of the page)

[jwm]

Seems we are slowly getting there. A lot of the previous problems have been solved. However, there are still some movies that don't give a proper picture. Examples:
"Dances with wolves" has large picture available, but the script returns empty.
"Arlington Road" has a small picture available, but the script returns empty.
"The prince of Egypt" has small pictures available, and also a large for DTS. The script returns empty.
"Evolver" has a small picture available, but the script returns empty.
"Last to surrender" has a small picture available, but the script returns empty.
"Kansas City", idem.

A lot of my movies return a small picture, while there is a large picture available in VHS. I am planning to experiment a bit with importing VHS picture first, and if not found try DVD. (the script now does DVD first, and then VHS).

In the meantime I have made the following changes to the script:

- truncated the year, so that years like "1998/I" also get imported;
- changed category import to show all categories, instead of only the first one;
- added a picklist that allows you to choose the Translated Title from all available 'also known as' titles, or leave it empty; by setting ImportAllAKATitles := True it will bypass the picklist and import all available AKA titles;
- changed the layout of the Actors box to one actor per line and some blanks between actor and role (ideally this will later become two columns);
- changed the IMDB source page from"main details" to "combined details";
- added import of producer(s);
- changed import of actors/roles to include the full cast;
- added name and year of the movie you are looking up to the header of the title selection picktree;

Here is the code:

Code: Select all

[kinetix]

Great improvements! I thank you for that.

But there was problem (not an error or anything, but...)
After I've imported all data, I look in the Actors' field and notice.. after the last actor is mentioned, there is a line break and an empty line appears after the last name.......
Corrections, any1?

another thought (perheps not 'needed' for everyone, but..)
if I want the article before the title (IE - THE before, and not after Terminator).
any1 have a code for this?

[jwm]

There is not really an extra line in the actors field, it is merely the CRLF at the end. To fix this make the following change in the code.

Replace these lines (653):

Code: Select all

          Value := Value +  #13+#10;
          FullValue := FullValue + Value;

with this line:

Code: Select all

          if FullValue = '' then FullValue := FullValue + Value else FullValue := FullValue +  #13+#10 + Value;

[jwm]

I have made a further improvement to the layout of the actors field.

Replace line 652:

Code: Select all

          if Value2 <> '' then Value := Value + '      (as      ' + Value2 + ')';

with this line:

Code: Select all

          if Value2 <> '' then Value := Value + #9#9#9 + 'as:  ' + Value2;

It looks less messy and the roles are nicely aligned in a column.

[trekkie]

found a few Bugs in my previous script.

1.
In AnalyzeMoviePage the line after looking for 'exec/obidos/ASIN'
should be

Code: Select all

           If not PickTreeSelected Then 
                TitleLine:= FindLine('/exec/obidos/ASIN/',AmazonPage, 0); 
           if (LineNr > TitleLine) Then LineNr:=-1 [b] ; [/b]

I forgot the ";" (that was a nasty bug to find)

Hawkie This causes the problem with "Auto Focus".

2. Also before displaying picture selection window the variable should be set to false so

Code: Select all

      PicktreeSelected:=False; 
      // Display the Picture Selection Window 
      .... 

3. jwm

A lot of problems getting a picture occurs when there is no large one
and as i said before i can't verify GetPicture success
( I've tested "Dances with wolves","The Prince of Egypt" ,"Arlington
Road" ..)
I am working on an improved version which will follow the link in
every paragraph entry and check for large picture existance.

In case no large picture exists . Should it get the Thumbnail from
AMAZON or revert to IMDB ??
( "Arlington road" has no IMDB picture but has a small AMAZON
thumbnail)

[jwm]

I know it is not so easy as it may seem to some people.

I appreciate your efforts in finding improvements. Thanks.

My preference would be:
-find large image in DVD;
-if not found then find large image in VHS;
-if not found then find small image in IMDB;
-if not found then find small image in DVD;
-if not found then find small image in VHS;

[Guest]

I just installed Ant and have put some movies into it. I am using the IMDB script that was included in the installation to download the rest of the information for my movies including the images. I have noticed that some of the movies I have previously retrieved information for, have gotten their image get's changed to another movie's image.

What I'm thinking is that the file names of the images that are stored on my hard drive are exactly as their are named on amazon or wherever the IMDB script gets it. So it probably is overwriting an image that I had already gotten for another movie.

Is this what is happening? Or am I doing something wrong with the way i'm saving.

What I'm doing now, is after I use the script to retrieve the information, I immeditately go to the picture on the bottom left and click on the folder to select a picture, but what i do is rename the file, usually it's named 30m.jpg, 20.jpg to the moviename.jpg. So far it's working but I'm sure there's a better way to go about this.

Booker

[antp]

That may be a problem, right... I shall add an option to generate an unique file name, or to use movie title as file name.

[booker]

After doing some other testing, the easiest thing to do is to just save the catalog after the script retrieves the information. From what I can tell the catalog saves the image in the amc file also so once you've saved it doesn't matter if the image files downloaded by the script overwrite each other. Correct me if I am wrong here on my assumption that the image file is saved in the .amc file.

For the time being have to save after every update via this script so that it correct saves the image isn't so bad. But for future features I hope something could be done about it.

booker

[antp]

That's quite strange.
The picture is directly saved in the catalog, or as an external file, but it does not change from one to another.
It depends of the True/False parameter of the GetPicture function call in the script.
Only .XML will force the picture to be saved as external file when saving the catalog. If you use a .AMC file it will keep the picture as it is, internal or external.

[antp]

I cleaned a little the large pic script.
I'll update the program soon, so it would be nice if you could say if this version of the script is ok for you, and use this one as basis for future versions/modifications :
http://www.antp.be/temp/IMDB%20(large%20pic).ifs

For the future, it would be nice to put a link to a file on a http/ftp server instead of putting the whole script in the forum, since this script becomes quite large (25 KB)

[booker]

What i've been doing is cataloging my cd folders (200cd capacity) by importing them from a spreadsheet file (openoffice, save as cvs format). It would be nice in future versions for a folder field (i.e. dvd folder 1) but I noticed the next version should have some custom field options.

So I'm going to be starting another catalog file and import the this cvs file of another caselogic folder. It would be nice if there was a feature to run this import script for each entry to great the information from imdb etc. But it's fine for now since it keeps messing up the the imported picture. (Haven't tried to new script tho).

I'll try the new script out.

[Guest]

Is the above script

http://www.antp.be/temp/IMDB%20(large%20pic).ifs

newer than than the one included in the 3.4.2 full install?

Booker

[antp]

yes.
I still have to update this script in the program.