search

Kartoffel0 / Mokuro2Pdf

Create pdf files with selectable text from Mokuro's manga html overlay
GNU General Public License v3.0
33 stars 3 forks source link

Potential Bug with page ordering #4

Closed EZBUTD closed 1 year ago

EZBUTD commented 1 year ago

I followed the steps in the collab notebook, and was able to get OCR json outputs for the images. However, the pages appear to be out of order. Not sure if something is wrong w/ the set up on my side, I took a look but could not find anything.

Any help is appreciated! The copy of my notebook with relevant files is: https://colab.research.google.com/drive/1_GoQtWC0JJWzWSpecaWmLcHwC9NeIWOH#scrollTo=pvm9Fggi2lUz

EZBUTD commented 1 year ago

Ah, it looks like something is indeed going strangely with sorting, where I think the default sort is putting files in the order of 1,10,100,... for some reason: image

EZBUTD commented 1 year ago

Removing the volume number and then changing the first sort on ~line 108 to: folder.append(pages.sort_by{|s| s.scan(/\d+/).first.to_i})

seems to have fixed it for me. Not sure if it's a strange edge case that I've run into or if I structured something wrong on my end.

Kartoffel0 commented 1 year ago

Dealing with filenames is really complicated, nothing wrong on your end just something I didn't account for. I've updated the repo and now it should hopefully sort it properly, I should've updated this back when Mokuro had a similar issue but I ended up forgetting about it.

Sorry for the late response, been pretty busy with work lately.

EZBUTD commented 1 year ago

Thanks so much for taking a look and building this! I'll comment again if the issue comes back up.