Canti: Manga Processing Program

[lilman]

I finally got epubs processed by Canti to show up in iBooks... man, what a pain. I don't think it liked the way I organized the files within the epub before, even though I followed epub standards. For now epub manga sucks big time... a good portion of the screen is wasted by a virtual book, there's no zoom, and a lot of times the images don't get sized right (when you switch from portrait 2 page view to landscape single page view, iBooks keeps the images at the same small size as the 2 page view).

I will release Canti with working epubs soon as v1.61, but before I do I'll look into giving the user an option to reverse the page order so that you can flip the pages from right-to-left for manga.

[lilman]

Here's Canti v1.61

New to version 1.61:
-Epub works in iBooks
Epub is finally out of beta and the output epub files can be synced through iTunes and read in iBooks.
I do not recommend epubs for manga though... a good portion of the screen is used by the virtual book, you can't zoom in on images, and iBooks sometimes sizes the images wrong (especially when switching from 2-page to single-page view, the image may remain at the smaller 2-page size).
Even though I eventually realized epub wasn't going to work well for manga, I had already invested a lot of time in it so I figured I should finish it. And now it's done, so take that iBooks!
-Epub page/bind ordering from right to left
Hehe, this is a cool feature to try out in iBooks. By turning this option on, the ordering of the output epub will resemble that of manga (i.e. pages and bindings will be ordered right to left).
Currently iBooks automatically loads page 1 of the epub (which when read right to left is the last page), so if you use this feature you will have to either use the page slider to jump to the real first page at the end of the book or load the first chapter from the table of contents.
Kinda fun to see the virtual iBook with proper manga ordering
This is set up in the .properties file.

Enjoy and let me know if it works for you

[lilman]

I had some requests for features, so expect these in the next release:

-[coding completed] Support for zip/cbz as input files
-[coding completed] Right to left page/bind ordering for zip/cbz and pdf
-[coding completed] Delete output image folders after processing. This feature could only be used if you output zip/cbz archives as these would allow the program to temporarily recreate the output image folders in case you need to process the same manga again.

-Update-
I had to cancel right to left page/bind ordering for zip/cbz as it conflicted with the delete output image folders after processing feature. During output folder restoration from zips the page/bind ordering could be reversed, screwing everything else up. Right to left page/bind ordering for pdf is still there and already completed.

Expect v1.7 to arrive tomorrow.

[lilman]

Here's Canti v1.7

New to version 1.7:
-Source files in zip/cbz format
Program can now process manga located in -parent_folder in .zip or .cbz format. You can also set -parent_folder to be a .zip or .cbz file.
Note that this will only process the images in the root of the archived file, so if you have an archive with multiple levels of subfolders then the images in those folders won't be processed.
-Delete output image folders after processing
This feature can only be used if:

1. You output zip/cbz archives as these will allow the program to temporarily recreate the output image folders in case you need to process the same manga again (a big time saver)
2. Binding (either through bind file or auto-binding) is disabled

This is set up in the .properties file.
-Pdf page/bind ordering from right to left
Same as epub right to left feature introduced in v1.61, now for pdf.
This is set up in the .properties file.

Enjoy and let me know if it works for you

[lilman]

Here's the features I'm working on for the next release of Canti:

-[coding completed] Improved intelligence of reprocessing manga
-[coding completed] Default .properties file generator
You specify what device you are using (kdx, k2, ipad) and the program will create a default .properties file for that device. This will include automatically finding your ImageMagick install directory, which is taking some time to program since it has to be able to handle different operating systems.

-Update-
I've made a lot of progress with the default .properties file generator. All that is left to do with it is handle auto finding the ImageMagick install directory. Starting tomorrow though I'll be working on migrating my system to the new Ubuntu so I'll be unable to program for a few days. Expect v1.71 to be released in about a week.

-Update 2-
I knew what I wanted to program and didn't want it looming in my brain while I waited for my new os to be set up, so I put in a lot of time today and finished v1.71 ahead of schedule.

[lilman]

Here's Canti v1.71

Special notice: The default ConvertManga.properties file will no longer be included when you download the program. Instead, you must create a default .properties file for your device/os using the new feature listed below. Make sure to make the default .properties file before processing any manga. Also make sure to read through the default .properties file and make any changes to match your individual needs.

New to version 1.71:
-Default .properties file creator
Will create a default ConvertManga.properties file for the device you specify.
This is run by calling: java -jar ConvertManga.jar -create_properties [device name]
Current list of acceptable [device name] values: kdx, k2, ipad
Note that this will attempt to determine your ImageMagick install directory based on what operating system your are using.
>For unix/linux or mac it uses default locations.
>For windows it will search "Program Files" directories of all local disks. You may have to manually correct this value if the program fails to find this install directory for you.
-Improved reprocessing intelligence
Verifies folders listed in progress.txt (i.e. the folders which were previously processed) and removes an entry if the folder no longer exists. Of course zip restoration is performed before verifying the progress folder in case you have del_output_image_folders turned on.
Also, zip restoration will only be performed on archives relative to the current processing (so if there are other archives in the ouput folder, Canti won't waste time restoring them).

Enjoy and let me know if it works for you

[lilman]

Here's the features I'm working on for the next release of Canti:

-[coding completed] Zip input image folders
For a long time I've had my manga scans organized as image folders, but since Canti can handle input .zip/.cbz files I've added this new feature. If it is enabled then all input image folders will be archived to zips (using whatever extension you set as zip_format, either .zip or .cbz) and the folders will be deleted. This will keep the file system cleaner since it reduces thousands of image files to a collection of zip archives.
-[coding in progress] .rar/.cbr support
This is tricky since rar is a proprietary format, meaning that I can't easily program a way to rar/unrar like I could with zips. Right now I'm planning on using whatever rar program you have on your system to perform the rar operations, similar to how ImageMagick is called external to Canti to perform image manipulation. I'm not sure if I can get this working right, but I'll try.

Don't expect the next release to come anytime soon... .rar/.cbr support is going to be a pain. If you have any feature requests you can reply to this thread or pm me.

-Update-
Finally got .rar/.cbr extraction working in Linux. With that done the rest of .rar/.cbr support shouldn't be too bad.

-Update 2-
Got .rar/.cbr creation working in Linux. Also wrote most of the logic for handling .rar/.cbr archives in the main program. Currently the program allows you to set zip and rar options separately (so you could output both .cbz and .cbr for each manga chapter if you wanted to) but I'm thinking about combining the two so that you just set one archive format (.zip, .cbz, .rar, or .cbr). I'll have to think about this more...

Next up I'll add Windows .rar/.cbr support. The coding is coming along nicely so expect a release in the next 5 days.

-Update 3-
.rar/.cbr support for Linux is complete, and it works like a champ! To make things easier I combined @zip and @rar into a single option @archive, and @zip_format and @rar_format into a single option @archive_format. I've programmed Windows .rar/.cbr support but haven't tested it yet. Still need to research Mac rar support since I don't have a Mac to test with, but it should be similar to Linux.

Coding is pretty much done but there are a lot of things left to test. Expect a release in the next 3 days.

[lilman]

This isn't Canti news, but a new update for ComicZeal has been released in the app store and it adds a couple of awesome features:

+Manga mode (flip pages right to left)
+In app screen brightness control
+Easily load the next comic in a series
+System wide comic file support (so can transfer comics from Dropbox, mail, etc)

You can follow ComicZeal news at the dev's twitter: http://twitter.com/bitolithic

[lilman]

Here's Canti v1.72

New to version 1.72:
-.rar/.cbr support
Canti now supports .rar/.cbr input and output files.
Because rar is a proprietary format, Canti cannot simply read or write a rar file itself. It relies on your system having a rar handler already installed.

For unrar support, make sure you have installed one of the following:

• For Linux/Unix: rar, unrar, unrar-nonfree
• For Windows: WinRAR (installed to [drive_letter]:\Program Files\[folder with "rar" in its name, such as WinRAR]\)
• For Mac: rar

For rar support, make sure you have installed one of the following:

• For Linux/Unix: rar
• For Windows: WinRAR (installed to [drive_letter]:\Program Files\[folder with "rar" in its name, such as WinRAR]\)
• For Mac: rar

Note that the .properties file has changed. Instead of @zip and @zip_format, there is now @archive and @archive_format. @archive_format can be set to .zip, .cbz, .rar, or .cbr
This is set up in the .properties file.
-Archive input image folders
For a long time I've had my manga scans organized as image folders, but since Canti can handle input archive files I've added this new feature.
If it is enabled then all input image folders will be archived (using whatever format you set as archive_format, either .zip/.cbz/.rar/.cbr) and the folders will be deleted. This will keep the file system cleaner since it reduces thousands of image files to a collection of zip archives.
This is set up in the .properties file.
-Other tweaks
I don't remember all the little stuff I changed, just know that the latest release of Canti is always the best version to use.
To give you an idea of Canti's growth:

# lines of code in Canti v1.1a: ~1600
# lines of code in Canti v1.72: ~3900

Enjoy and let me know if it works for you

[lilman]

Here's the features I'm working on for the next release of Canti:

-[coding completed] Various updates
Random things I want implemented, like restore archived input image folders, and fixing a problem where delete_output_image_folders leaves an output image folder or two (no idea why this is happening... it may be a Windows only problem).
-[coding completed] Auto split 2-page portrait scans
I've wanted this feature for a long time but it's been at the bottom of my todo list because it's tricky and I will have to write it from scratch. I think I will be able o implement this but it may take some time. If it works then it should be able to detect if the scan is really 2-pages (i.e. you lazy scanned the open book) but not split the page if it is one of those 2-page wide spreads.

That's about it. This should be the last Canti release as I have a new project coming up, so if you have any feature or bug fix requests let me know now.

-Update-
Mostly done with the various updates.
I have a good start on auto splitting 2-page portrait scans. I'm writing code to analyze a group of pixels and independently compare the rows and columns within a color threshold. I think this should allow me to find the two main tells of a 2-page portrait scan:
1) Big white column in the middle of the scan
2) Thin dark column border surrounded by white running down the middle of the scan (i.e. the shadow created by the page fold)

-Update 2-
Finished most of the coding for auto splitting 2-page portrait scans. I can't believe how well it works!... it is very good at telling which 2-page spreads should be split and which ones should be left alone (i.e. the 2-pages form a single image that is better to view as a whole). It's also incredibly fast and shouldn't impact processing time too much (less than a second per page analysis, I guess I could time it for a more accurate reading). All that is left is tweaking the parameters to get the split detection just right.

-Update 3-
Coding is done. I'll do a bit more testing and then release the new version.
Auto split pages is awesome!

[lilman]

Here's Canti v1.8

Biggest addition here is auto split 2-page portrait scans! I am so proud of this feature as it uses my own two-page detection algorithm, and it works correctly almost all the time on the manga I've tested. However, since I could only work on it based on tests of my own manga there is no guarantee it will work right for you. So if you give it a try and its accuracy sucks then let me know and I'll add in the ability for the user to tweak it. For now I've kept all two-page detection settings hidden so that the user doesn't have to worry about them.

New to version 1.8:
-Auto split 2-page portrait scans
If enabled then Canti will attempt to automatically split portrait scans that contain two pages into separate images. Canti will try to only split two-page portrait scans when the two pages don't form one giant image.
Note that this is not perfect... sometimes it will split scans which shouldn't be split or doesn't split scans that should have been split. It uses an algorithm which I came up with myself to try to detect the space between two pages in a scan. The algorithm isn't perfect, but it's fast and it works pretty damn well.
There are additional options that you can specify when using this feature:
@auto_split_r_first: This value determines if the right side of the image should come before the left side (like with manga books).
@allow_auto_split_color_scans: Allows or prevents the auto splitter from splitting color images. I've defaulted this to off since two-page color scans in manga usually form one giant image.
This is set up in the .properties file.
-Improved page ordering
Previous page ordering moved cover pages to the front and credits to the back (based on filenames).
Now also attempts to place inserts after the covers, checks if a cover is a back cover, and doesn't alter the ordering if the order is already forced through a number naming scheme.
-Unarchive input image folders
If enabled then any input image archive will be unarchived to a folder and then the archive will be deleted.
This value supersedes @archive_input_image_folders and is useful if you need to test individual images within an archive collection.
This is set up in the .properties file.
-Other tweaks
Too hard to remember everything, but as always the latest release of Canti is always the best version to use.

Enjoy and let me know if it works for you

[lilman]

Oops, very minor mistake in v1.8, so download v1.8a instead. Sorry about that.

New to version 1.8a:
-Fixed a minor mistake with auto split 2-page portrait scans.
I was splitting pages at the left edge of the detected page spacing instead of the center. This should make the x coordinate of page splits a few pixels better.
(I know, not a big deal... any split within the detected page spacing should be fine, but I'm a perfectionist :P)

Enjoy and let me know if it works for you

[lilman]

Here's the features I'm working on for the next release of Canti:

-[coding completed] Bug fix: Single folder mode doesn't work on a single file (only a single folder)
Wow, single folder mode was broken big-time if you specify a single file and not a single folder. Looking through the code I'm surprised it ever worked. I've already fixed the problem and will release the update soon... sorry about that.
-[coding completed] .pdf support for input
Canti already had .pdf support for output, but now it will also have .pdf support for input. This probably won't be used much for manga/comics, but it's useful if you have a pdf doc that you want trimmed or want some pages of it as images. Note that this feature works by converting each page of the pdf doc to individual images; thus, an output pdf will not have the ability to rescale font like you can with typical pdf files.

The new version of Canti will be released soon. I'm already starting my next project so I want to leave Canti in as final a state as possible. If you want any other features or have any bugs to report, let me know now.

-Update-
Coding and testing done. Fixed some bugs, added pdf input support (which works great... I already used it on a 150 page technical pdf to trim the margins. If you have an iPad you shouldn't need to use this since Goodreader has a manual crop feature, but for other readers it may be useful). I will wait a day or two before releasing the new update to see if there is anything else I want to add; after that, I'm moving on to a new project.

-Update 2-
Program is pretty much finalized. I'm working on a proper readme doc... better do that now since the program is still fresh in my head and I'll probably forget most of it when I start my new project. I've added default settings for a bunch of other ereader devices so I'll have to find a good place to upload the final program (no way I'm going into each ereader board to post this).

Expect a release in the next few days. It will be v1.81F, the F meaning that it should be the final release. Any further release after that is only if a bug is found or I add in functionality that I think is worth releasing to the public. But for the most part, I won't be working on this program anymore.

-Update 3-
I started making a proper doc for Canti and realized it was going to take a long time to do it right. I had a choice to make: work more on the code, or write a proper doc. I chose to work more on the code since that is more fun than technical documentation :P

Lots of updates in the final release of Canti. At this point I think I'm done with the coding, I just need to use it a while longer to make sure everything is working properly. Expect a release in the next few days.

-Update 4-
Final release is taking longer than I expected. The problem is I can't work on it every day, and the days I can work on it I can only get in an hour or two of programming/testing. Also, I keep coming up with features/improvements and I don't want to leave anything out of the final release. But no worries, the final release is coming... eventually

[lilman]

Still working on the final release of Canti (so much coding!), but I wanted to release this now. There is a bug in the previous releases of Canti that if you use @archive_input_image_folders using @archive_format .zip or .cbz on a Windows machine than the filenames of the images within the archives will have a \ placed at the beginning (for example, \01.jpg instead of 01.jpg). This bug has already been fixed in the final release of Canti, but I'm releasing this tool so that you can rearchive affected folders.

Just extract the program from the attached zip and run:
java -jar FixArchives.jar [path to folder containing affected archives]

For example:
java -jar FixArchives.jar C:\\Manga\\SchoolRumble\\

It will go through all the .zip/.cbz archives in the specified folder, unarchive them to temporary folders, fix the filenames, and then rearchive. Depending on how many archives you have in the specified folder it may take a few minutes for the program to complete.

It's kinda strange that this bug only occurred on Windows machines and only if you used @archive_input_image_folders (not sure why the output folder archives weren't affected). Anyway, sorry about that.

[lilman]

Here's Canti v1.81F, most likely the final release

New to version 1.81F:
Note: The F stands for final, so this is probably the last release of Canti.
Also note that the jar has been renamed from ConvertManga.jar -> Canti.jar and the .properties file has been renamed from ConvertManga.properties -> Canti.properties
-Bug fix: Single folder mode doesn't work on a single file (only a single folder)
Wow, single folder mode was broken big-time if you specify a single file and not a single folder. Looking through the code I'm surprised it ever worked.
The problem was complicated but has been fixed... sorry about that.
-Bug fix: .bmp support fixed
.bmp support was always in Canti but Canti didn't know that .bmp was a valid image format. Now it knows, problem solved.
-Bug fix: Internal filename problem with .zip/.cbz archives
Files within .zip/.cbz archives had a filesystem separator char at the beggining of the filename (for example, \0000.jpg instead of 0000.jpg). I believe this bug only occurred on Windows machines, but it has been fixed.
-Bug fix: @del_output_image_folders didn't always work correctly in Windows
Sometimes in Windows if you enabled @del_output_image_folders, a folder or two would be left undeleted.
This bug was not my fault. The code was correct, but for whatever reason Windows has a problem sometimes with java deleting folders. I had to add in code to retry deleting output image folders if it fails the first time.
-.pdf support for input
Canti already had .pdf support for output, but now it also has .pdf support for input.
Note that this feature works by converting each page of the pdf doc to indvidual images; thus, an output pdf will not have the ability to rescale font like you can with typical pdf files.
-Improved auto sense suffix
Auto sense suffix can now handle hyphens in suffix (for example, Manga_Title_001-023) and spaces/underscores.
-Improved auto split 2-page portrait scans intelligence
Improved the page split detection algorithm by increasing the number of samples it uses to try to detect a page split (at a small cost to processing time).
Even if you have @auto_split_color_scans disabled it will still split scans that are half grayscale half color (because those are guaranteed not to form one giant image).
Added options to the .properties file to specify auto split grayscale and color strictness. These values will allow you to fine-tune the page split detection algorithm to your liking (so the user can make it easier or harder for a page split to be detected).
Added @auto_sense_portrait_scans to the .properties file to specify whether the page split detection should only be performed on folders in which most of the scans are portrait.
-Added default .properties profiles for more devices
You can always edit the .properties file for whatever ereader device you have, but I've added default profiles for the following devices:

B&N: nook
Sony: prs-900, prs-700, prs-600, prs-505, prs-500, prs-300
Default: default (used to make a default .properties file that you can modify if you use a device Canti doesn't have an optimized profile for)

-Set default manga_prefix in the .properties file
The value of -manga_prefix can still be set in the command line at runtime, but now you can also set a default @manga_prefix in the .properties file.
If you set @manga_prefix to "auto" then Canti will attempt to use the same -manga_prefix that you used previously for the current series. This only works if you are using the same output_folder and -title as before.
If you specify -manga_prefix at runtime then this value will supersede the one in the .properties file.

*phew* That was a lot of updates. Since this is the last release I wanted to get everything in there. Actually, there are a lot of features that I didn't get to program. If I get more time later, or if I ever release my own comic reader app, I will use Canti as a base and make it even more awesome. But for now I need to move on to a new project...

As always, enjoy and let me know if it works for you