View Single Post
Old 07-14-2010, 04:38 PM   #2
lilman
Addict
lilman has learned how to read e-bookslilman has learned how to read e-bookslilman has learned how to read e-bookslilman has learned how to read e-bookslilman has learned how to read e-bookslilman has learned how to read e-bookslilman has learned how to read e-bookslilman has learned how to read e-books
 
lilman's Avatar
 
Posts: 326
Karma: 960
Join Date: Jul 2009
Location: Florida, US
Device: Kindle DX, iPad
Hehe, how many times is this thread going to be moved around? This started in the iRex forum, but I guess they weren't interested in it... oh well.

Anyway, here's Canti v1.82

Note from the author: I tested this version on my entire digital manga collection, so I can confirm this is a very stable release.

New to version 1.82:
-Improved image handling stability
ImageMagick is a whole lot better at handling images than native Java libraries, so I've added in a preprocessing stage to let ImageMagick fix any weird data within images before letting Java take a whack at them.
Currently, Java image handling is only used for landscape 2-page split detection. In case Java still has trouble interpreting an image after ImageMagick has processed it, Canti will simply ignore trying to split that scan (in the past, Canti would have crashed).
-Bug fix: @del_output_image_folders didn't always work correctly in Windows
Geez, I thought I fixed this before, but it still occasionally fails to delete an output folder on Windows machines. I added in one more layer of defense. After that, if there is still a folder left in the output directory, you can delete it yourself :P
-Improved auto sense suffix
I changed the naming scheme of auto sense suffix to make it better able to handle reprocessing a manga series.
Before, auto sense suffix would expand all suffix names to match the length of the longest suffix. For example:
Random_Manga_-_Ch.001 <-- added two extra 0's to match longest length of three
...
Random_Manga_-_Ch.015.5 <-- added an extra 0 to match longest length of three (decimal and characters after the decimal don't count towards length)
...
Random_Manga_-_Ch.100 <-- longest length (three characters)
Although this makes the filenames line up nicely, it can be a problem if you add manga to your collection and the longest suffix length changes.
The new auto sense suffix method is to make each suffix as short as possible. For example:
Random_Manga_-_Ch.1
...
Random_Manga_-_Ch.15.5
...
Random_Manga_-_Ch.100
In this way, if the length of the longest suffix changes, the filenames will not be affected (and thus you won't have to worry about reprocessing the same manga using a different name).

If you want to fix the suffixes of manga processed with previous version of Canti to match the new naming scheme, you can use the attached FixSuffixes tool. Just run:
java -jar FixSuffixes.jar [path to processed manga folder]
For example:
java -jar FixSuffixes.jar C:\\Manga\\SchoolRumble\\
Will only take an instant to rename all the files in the specified folder to the new naming scheme.

Enjoy and let me know if it works for you
Attached Files
File Type: zip Canti_v1.82.zip (78.6 KB, 449 views)
File Type: zip FixSuffixes.zip (18.7 KB, 422 views)

Last edited by lilman; 07-14-2010 at 06:10 PM.
lilman is offline   Reply With Quote