I had lots of troubles getting all the service account and keys setup yesterday manually. You explained it pretty well in the readme, but i still went nuts trying to second guess in via the Google console gui. I still have not gotten all the config correct. Google Vision works but not Google Translator

So perhaps this might be work using ?

https://github.com/caos/orbos

josselynennie@cloudshell:~$ cloudshell_open --repo_url "https://github.com/cameronkinsella/manga-translator" --page "editor" --open_in_editor "scripts/cloudshell/create-service-account-key.md" --force_new_clone 2022/10/14 21:36:54 Cloning https://github.com/cameronkinsella/manga-translator into /home/josselynennie/cloudshell_open/manga-translator Cloning into '/home/josselynennie/cloudshell_open/manga-translator'... remote: Enumerating objects: 372, done. remote: Counting objects: 100% (17/17), done. remote: Compressing objects: 100% (12/12), done. remote: Total 372 (delta 8), reused 5 (delta 5), pack-reused 355Receiving objects: 100% (372/372), 4.45 MiB | 26.18 MiB/s, done. Resolving deltas: 100% (191/191), done. Traceback (most recent call last): File "/usr/local/bin/cloudshell", line 496, in main() File "/usr/local/bin/cloudshell", line 493, in main args.handler(args) File "/usr/local/bin/cloudshell", line 110, in edit_files open(filename, 'a').close() FileNotFoundError: [Errno 2] No such file or directory: '/home/josselynennie/cloudshell_open/manga-translator/scripts/cloudshell/create-service-account-key.md'

On several manga that I am working to translate, the boxes which cover the translated text appear to not be capturing the entire speech buble and as such is producing nonsense translations. Would there be a way to manually create these boxes around the text so it knows which exact characters to try and translate or is this a limitation of the OCR software?

Very cool software !

Its possible for the system to recognise automatically where the Text is in each picture, so that a user does not have to click on it.

This example does is : https://github.com/sachaos/jisui

Basically a PDF is sent to google, and it can extract where the Text is in the document.

If you look at the 2 PDFs in a PDF viewer and use the Find tool you can see that it put the real text in during the processing such that the Find functionality now works. So you end up with a fully work PDF with FIND !

https://github.com/sachaos/jisui/blob/master/example/scanned.pdf

• Cannot Find the text.

https://github.com/sachaos/jisui/blob/master/example/result.pdf

• Can Find the text

Im trying to translate manhua by putting this (https://rawkuma.com/) site link but its say cannot download images.plss fix this! Screenshot_2022-11-01-03-04-04-344_com.wiryaimd.mangatranslator.jpg

Add new option to translate an image from your clipboard.

The built-in Windows snipping tool (there are also Linux and Mac equivalents) allows you to freely select a portion of your screen and copies it as an image to your clipboard. Using the new clipboard option in conjunction with the Windows snipping tool makes it much easier to do quick adjustments and translations without needing to save a new image every time.

Resolves #3

Tried it out but when i open the app it fails silently. I guess i need to figure out how to capture logs.

TO run in sim:

gogio -o manga-trans.app -target ios .
xcrun simctl install booted manga-trans.app