just a reminder that i have a patreon now: https://www.patreon.com/lynnesbian
the main tiers to be interested in are:
- ebooks bot ($2.50) - your own ebooks bot, hosted on my server
- custom bot ($6) - i'll make you your own simple bot. it can make images black and white, find rhymes, play tic-tac-toe, whatever you'd like
if you're interested in the custom bot i implore you to read the whole tier description
my main projects are:
- #FediBooks, my main project, and the biggest thing i've ever worked on! it's a GUI for creating fediverse bots that supports all sorts of features, from simple "ebooks" type posting to custom reply rules - https://github.com/Lynnesbian/FediBooks/
- @OCRbot, a bot that uses tesseract OCR to recognise the text in an image and reply with what it believes the image says -- check the replies to this post to see it in action - https://github.com/Lynnesbian/mstdn-ebooks/
- mstdn-ebooks, a command line program for creating ebooks bots. it's very simple and kind of old, but i do still support it. - https://github.com/Lynnesbian/mstdn-ebooks
Okay, I've fixed mstdn-ebooks' compatibility with Pleroma.
If you're getting me to host your ebooks bot(s), there's nothing you need to do.
If you're hosting your own (or someone else is hosting yours for you), you only need to do anything if any of the users your bot learns from use Pleroma. In that case, you need to delete `toots.db` and restart the reply service.
Thanks for understanding.
i've pushed an experimental fix for the OCRbot federation bug (this one: https://github.com/Lynnesbian/OCRbot/issues/8). hopefully this fixes OCRbot's longest standing bug!
i'm currently doing some maintenance on https://searx.lynnesbian.space. there might be some server errors, slow results, etc.
i made a post about why i should never have made OCRbot, and where and how i went wrong with it.
(the text in this screenshot is taken from https://lynnesbian.space/csl)
OCR Output (chars: 1247)
. eel 86 8 4 O 18:25
@ https://lynnesbian.space/csI) 6 @& :
The Cooperative Software License (CSL) is a copy-far-
left (as opposed to copyleft) license that aims to rectify
some of the issues that arise when using a copyleft
license such as the GPL, namely, that your work can still
be used in some extent by commercial entities and
other for-profit ventures. It is based on the Peer
Production License, and is adapted to be more
applicable to software.
Licenses like the CSL and any Creative Commons Non-
Commercial License are not libre (free as in freedom)
software licenses, as they restrict the freedoms of
certain user groups and use cases. This is why the CSL
is classed as a copy-far-left (as opposed to copyleft)
license. It is therefore incompatible with (and thus
cannot link to) software released under licenses such as
the GPLv3 or the MPLy2. It's important to keep this in
mind when using the CSL for your project. The CSL is,
however, compatible with less restrictive open source
licenses, such as the MIT License or Creative Commons
0. The Free Software Foundation lists reasons why the
Peer Production License (the basis of the CSL) is not a
free as in libre license.
OCRbot now reports the character count of its transcriptions in the content warning. see its reply to this post for an example.
a list of all config.json options for mstdn-ebooks can be found here: https://github.com/Lynnesbian/mstdn-ebooks/blob/master/README.md#configuration
OCRbot will no longer mention people if their handles are included in images.