MangAI v1.00b released
New to version 1.00b:
-Bug fix: Invalid collection of computer data for trial licensing
As trial licenses are tied to both your email address and your computer, gathering the correct computer information is important. I accidentally left open the possibility for invalid computer information to be gathered. This could cause multiple computers to map to the same identification, and, in turn, new users could be denied a trial license.
This has been fixed.
-Check for Update feature added to the GUI
Click Help->Check for Update to see if a new version of MangAI is available.
Still working on the @prefix option that will allow you to specify a custom prefix for output filenames. As the bug fix above is important, I didn’t want to delay this release any further.
Grab the latest version from the MangAI project page.
I’ve been working on how prefix will work. Here’s the general idea so far:
@output_prefix
Prefix placed before every output publication filename.
Leave blank if you don’t want to use a prefix.
Supported shorthand:
[title] – manga series title specified by user at runtime
[#] – integer that increments in order of publication sort
[#n] – same as [#] but integers are padded with leading zeroes to length n
Examples:
If set to “[title]~”, output files will look something like “School_Rumble~School_Rumble_-_v1_c1-10.cbz”.
If set to “[#]-“, output files will look something like “1-School_Rumble_-_v1_c1-10.cbz”, “2-School_Rumble_-_v2_c11-20.cbz”.
If set to “[#4]-” output files will look something like “0001-School_Rumble_-_v1_c1-10.cbz”, “0002-School_Rumble_-_v2_c11-20.cbz”.
Worked some on @output_prefix. I decided not to implement [#] or [#n] shorthand as that can cause the output to get messy when reprocessing. The prefix numbers of previously processed manga may be different when new files are added, leading to duplicate output publications with slightly different filenames.
If you have any suggestions, please leave a comment. After this next update I hope to switch focus to iOS dev so that we can eventually have a MangAI reader app.
I think @output_prefix is nearly done. Also fixed a few bugs.
There’s another feature I want in the next release, option for MSM interpretation/sort during the other operation organization. Right now, other operation organization does full expansion + MPM interpretation, meaning input publication are unarchived and renamed based on interpreted volume-chapter info. Including MSM interpretation/sort may further increase the usefulness of this op.
Idea from comment #3 implemented. You now have a choice between the old “organize” and the new “full organize”. Full organize works just like regular organize but includes interpret/sort/rename scans of the organized publications.
Still thinking if there is anything else I want to include in the next update. Feel free to leave a comment if you have any suggestions.
Got some great bug reports. For the next few days I’ll be in debug mode. I’ve already fixed/improved some Manga Scan Model logic.
If you experience any runtime exceptions or freezes, especially during publication or scan interpretation, please contact me. I cannot make the AI perfect, but I love to make it as good as I can.
For problems during scan analysis, you can change the @scan_interpretation setting value from “msm” to “string”. This will avoid the tricky/difficult logic of Manga Scan Model analysis and instead sort scans by simple filename comparisons. When the update is released you can switch back to msm and hopefully your problem has been solved.
Some great logic updates including the prevention of infinite loops in Manga Scan Model scan analysis. At first I was trying how to create a way to detect an infinite loop in general, using something like a stack history of operations, but then I remembered a little thing called the halting problem, lol.
More logic work and testing ahead, but I think I got the reported bugs handled.
I’ve been working on overspecification of special pages. Here’s an example:
foo_p01.jpg
foo_p02.jpg
foo_p02_note.jpg
foo_p03.jpg
In this example, foo_p02_note.jpg would initially be detected as an extra page since it has the special keyword “note” in it. According to the Manga Scan Model, extra pages are placed after normal pages and before back cover and credit pages. In this case though, that particular extra page is most likely meant to be read right after reading page 2 since it probably explains something important about that page, such as the reasoning behind a translation; therefore, you don’t want to treat that page as an extra.
There are lots of little nuances like that when dealing with the Manga Scan Model. I’ll continue working on the AI to make it as smart as I can.
Rewriting the epub creation code since the current code forms invalid epubs. I really don’t like epub for digital manga since epub reader software is designed more for text layouts (i.e. standard books) than full screen image display. However, I don’t want to drop epub support in the off chance that a user has an ereader that doesn’t support folder/cbz/pdf.
i have to admit this is a great program i’ve been converting manga to my nook color for a while now and its always been a pain (since i had to reize all images to 600 by 1024) and then unually make them the same file type.
By the way would it be possible to create a nook color of the mangai.properties.
@Robert
Glad you’re enjoying the program.
Yes, I can add a nook color settings profile. If you email me your mangai.properties files I will use it as a template.
-Update-
While I’m at it, I will make profiles for the Kindle Fire and any other new reader I can think of. If anyone has one of these new devices and has determined the best settings for them, please send me an email. Otherwise I will just choose settings based on each reader’s tech specs. This usually gives pretty good settings, but there is nothing like firsthand user feedback for optimal profiling of each device.
I just sent you the properties for my nook color/tablet (both have the same properties just one has a faster processor then the other and better battery)
@Robert Savage
Received and will be included in the upcoming release. Thanks!
Updated the epub code. I only have my iPad with iBooks to test with but it seems to be working properly.
Next up is adding settings profiles for the latest ereaders, then it’s back to logic twesting (tweaking + testing).
Looked around at the currently available tablets, and they all have pretty similar specs. Any large screen tablet like the Samsung Galaxy Tab will use the same settings as iPad (don’t resize the images so that you can zoom in, color, cbz and pdf the output). The new line of 7″ tablets all have similar specs as well (600×1024, color, cbz and pdf the output).
For now I’m adding these new profiles:
nookc (Nook Color)
nookt (Nook Tablet)
kf (Kindle Fire)
Those three devices actually have the same settings, but I gave them separate profiles just in case the specs change. If anyone would like another ereader added (like a Samsung Galaxy), please let me know.
Added more profiles:
galaxy10 (Samsung Galaxy Tab 10.1)
galaxy7.7 (Samsung Galaxy Tab 7.7)
galaxy7 (Samsung Galaxy Tab 7)
prs-t1 (Sony PRS-T1)
kobov (Kobo Vox)
xoom (Motorola Xoom)
I’m done with profiles for now unless I get specific requests. Back to work on logic.
I think all that is left is testing the new logic. Hoping for a release in the next few days.
Logic testing done! Need to see if there are any little things left to do. Next release should be out either today or tomorrow.