OCRbot now reports the character count of its transcriptions in the content warning. see its reply to this post for an example.

cc @OCRbot

OCRbot will no longer mention people if their handles are included in images.

@OCRbot

preliminary testing using starwall-e

looks like it works! it's not generating any fake mentions anymore!

i've made a regex filter that will remove the fake mentions that appear at the start of ebooks bot posts. this behaviour will be optional, but it's been lacking from mstdn-ebooks for a while.

a bug with OCRbot that prevented it from working with chinese text has now been fixed!

@OCRbot chi

OCRbot should now support multiple languages. simply say @OCRbot jpn and it will parse the image as japanese. support for other languages can easily be added.

is designed to be intuitive and simple to use, but that's not always easy to achieve, and i don't want to overwhelm the user with huge explanations of every step. if you need more information during any step of the way, you can click the "help" button to learn more about the exact page you're on.

here's what can do so far:

- detect whether a given instance is mastodon/pleroma/misskey/etc
- redirect you to the account creation page if necessary
- authenticate with the instance

in other words, the basics are nearly done 0u0

here's an example of a macOS user going through the guide as far as is currently possible.

when i use tracing, the error doesn't happen, and it runs fine

but when i run the program normally, it gets stuck after the "create an account" step, and the whole thing freezes

why is it different when i trace it even though i'm running the same code with the same python version ;-;

so fedibooks is confirmed to work on linux (which is where i've been developing it), windows, and macOS :blobcathappy:

i don't know why my macOS virtual machine starts up with init logs, but i have to say, i vastly prefer it to the apple logo

i want to bundle the font used (noto sans) so windows/macOS users can have the layout look right

i checked the way that chocolatey (a "package manager" for windows) does it and it seems to just hardcode the download URL...

if that's the best solution they've got then i guess i'll do that too

it works

(the username/password entry format is only used for diaspora*, don't worry! mastodon/pleroma/misskey all use oauth)

in the process of testing fedibooks on windows for the first time...

i'll definitely want to include an installer for this

also, for the curious, this is what it looks like when you try to fit every single window fedibooks has onto one screen!

will allow for both complex and simple functionality! you can set custom rulesets with multiple conditions and actions, or you can just use simple ebooks-style posting!

Lynnestodon

@lynnesbian@fedi.lynnesbian.space's anti-chud pro-skub instance for funtimes