3.1 - Community Patch [PSD Support]

• Fixes multiple typos and misspellings (Courtesy of @kianmeng)
• Adds PSD support as both input files and output (Courtesy of @BishrGhalil)

Huge thanks to both of you!

[Note about PSD Support]
-PSD loading and saving is both more ram intensive (Higher memory usage than usual) and much slow (Will take significantly longer than usual).
-Output .psd file are still essentially images but in .psd format, therefore it will not contain the same layers as the original

3.0 - A new beginning

SmartStitch V3.0

• Fundamental rewrite of the entire software.
• Retains all previous features and more.
• Cuter and Better Looking UI, now using PyQT.
• No more tinkter bugs.
• Faster performance, around 10-15%.
• A much lower ram usage especially for large chapter/input images.
• Easier setup process for mac and Linux.
• Everything is now modular, so we can expand on it much faster and easier than before.
• Quality field for lossy output image types.
• Automated Batch Mode, automatically navigate input directory for viable work directory.
• A better logging system.
• Less crashes i hope.

2.3.1 - QoL Changes and Crash Condition Fix

• Fixes the reoccurring crash that happens sometimes there are no more rows to iterate over. This happens sometimes when the program reaches the end of the image strip. (For ppl experiencing crashes please update to this version!)
• Changes the minimum allowed image height from 50% of the rough height to 40%, this is to give more lee way for those that upload to MangaDex, where small sizes are more welcomed than larger ones.
• Program now saves the latest input folder path/directory and can set it as the initial directory that the Browse Dialog opens at. Should be helpful to people that use the same input folder, no more renavigation to back to that folder.
• Now the progress bar will take into a count if you have a subprocess to run after stitch (it will reserve 20% of the process of each input to display it's status/completion). So now it should be a bit more representative of the entire process progress.

2.3.1 Fixes & Changes

• Fixed another crash condition related to resizing image files, a pretty rare case when you get a raw image with a single pixel height and it can't resize it down.

2.2 - The Shiny and Polished Update

WOW, so SHINY, MY EYES ARE BLESSED, MY SWAN SONG, MY MANGNUM OPUS... My story of how my program was reincarnated as a slime in a tkinter harm, but there is not enough brains cells to keep working with this awful python package... seems like a pretty good story to read.

Anyways, my mental breakdown as side, the main focus of this update was to polish and clean up any minor issue with the program, as well as add to a shiny new GUI modern theme. (It is likely my last major update for a long while)

• Adds new modern and clean theme, that fits with the new Logo. (No one would actually believe that this is tkinter)
• Added scrollbars for the subprocess console, and now it will not grow crazy in size like last update.
• Added alot of blood, sweet and tears to make tkinter behave. (Word of advice, don't make GUIs with python lol)
• Fixed an issue that caused some of the subprocess message to only appear after it was complete, now all subprocess message will appear in real-time in the subprocess console.
• Fixed some minor padding issues here and there.

2.1.2 - Adds new subprocess after completion feature (The Waifu2x Update)

• Adds the After Completion Subprocess Feature
• (GUI Only) With this option, one can set a specific console process to be fire on the output files of the application. For example, you can set it to fire waifu2x on the output files, so you can have the best raw processing experience. So how do we set that up,

1. Enable the Show Advanced Settings
2. Enable the Show Subprocess Settings
3. Enable the run subprocess after completion flag.
4. Set the process path/location, you can essentially browse to the process' exe file
5. Set the arguments you want to pass to the process (Use the argument [output_folder] [stitched] to pass the output directory to your process).
6. Optional: Use the argument [processed] to pass a custom output directory to your process for those that can't create their own output.

• TLDR; Basically now you can run waifu2x/pinga or whatever process you want on the output images, It's the ultimate person's dream come true.

2.1.1 Fix

• Changes the argument [output_folder] to [stitched] for consistency
• Added a [processed] argument to create a custom output folder for process that can't create their own output folder (Looking at you waifu2x-ncnn-vulkan)
• Error message of the console should now be displayed in the GUI under the subprocess console.

Quick Guide for those using waifu2x-ncnn-vulkan

this is how I set up mine, with the input being [stitched] and the output being [processed], the rest of the arguments are up to you,

2.1.2 Fix

• Fixes subprocess not working, if it's path contained a space, now it will add quotation around the path if it detects it contains a space.
• If the path does not contain a space it will not add the quotation marks, since it does not like them when there is no spaces.
• Cleaned up how the subprocess display the start and end phrases.
• Accepts merge request for using .png for the internal application Icon.
• Added an noicon spec file to compile for other platforms.

2.0 - The Overhual Update

SmartStitch V2.0

• New Shiny Logo, with more stitches Teehee xD.
• Alot of rewrites for many things.
• All core stitch related functions were moved To a Core File.
• Having a common file allow both the Console and GUI version to be updated with each other.
• Rewriting the Console, essentially finally updating it to the same functions as the ones GUI uses. (Now runs from the Core File)
• Added the SmartStitchGUI_SingleFile.spec, now you can build a single file version of the GUI using Pyinstaller.
• Rewriting the GUI, to be cleaner (code wise) and work better with the centralized Core File.
• Separated the settings in the GUI to basic and advanced settings.
• Added 2 advanced settings, which are the ignorable pixels edges, and the scan line step. (These features can be called from the Console Version as well)
• Version 2.0, should be a better platform to build on for newer version, so look forward to possible new features :3.
• Full details on everything in the application will now be found in the application's read me file.
• Documentation

1.9 - Fixes to random freeze up and crash conditions

• Should fix the random freeze up that happen with very specific document size.
• It treats what the last version could not.
• Removes the numpy warning about subtraction unsigned int... since images are made of arrays of unsigned integers (SO.... changed a bit how the pixel value is calculated). bye bye warning :D.
• Sets up some functions for a possible future feature.

1.8.5 - Important Fix for a Crash Condition

This update does not add any new features, but rather fixes a critical crash condition on some edge cases.

• The crash occurs on some edge cases (Happens when a single row of pixel is left to be sliced alone at the end), the program will run outside the image size by iterating and adding 1 to the last row index,
• Basically, it does not stop at the last row of pixels in some edge situations.
• This should fix this issue so it should not any edge case of that sort anymore.
• Please use this version instead of the 1.8, since it has all the same features plus this fix.

1.8 Adds Width Enforcement Modes

So essentially it's very straightforward. It adds a setting to select one of three modes.

1. No Enforcement, where you load the files as is, and work on them, if they vary in size, you will get some black lines in the side (Highest quality as there is no changes to the pixel values)
2. Automatic uniform width, where you force all files to have the same width as the smallest file in the input folder.
3. User Customized width, where the user specifies the width they want.

• For the last two options, it uses LANCZOS resampling to resize the images, it's decent for small changes, but really should not be used as means for upscale images. (Please just use waifu2x for upscaling raws)
• The new settings are saved, and will load on relaunch for lazy ppl like me. xD
• Additional note: This does add a bit of processing time depending on the image size, I have observed an additional 1-3 seconds.

1.7 - Batch Mode and other QoL Changes

For the ultimate time saving experience, we now have batch mode.

• By setting batch mode to on, the program will stitch all the folders (one by one) with input path.
• Ex. you have ch1, ch2, ch3 folders in a folder, then use batch mode and it will automatically stitch it for you.
• Now execution duration will now be displayed to the max accurate digit possible (satisfying numbers, hmmm delicious).
• Now the GUI will not freeze when the stitching is in progress, because the work is being done on a separate thread, so the GUI will no longer freeze.
• Changed how the progress bar works, and now it's more dynamic and accurate depending on the number of inputs (this is mainly for batch mode.)