search

diracdeltas / icowid-generator

markov bot based on ICO whitepapers and erowid trip reports
GNU General Public License v3.0
37 stars 2 forks source link

forking to toot into the fediverse #3

Open Be-ing opened 3 years ago

Be-ing commented 3 years ago

Hi @diracdeltas, thank you for this bot which has brought me some hilarious moments. I am going to revive it and make it toot into the fediverse instead of Twitter. I have applied for an account on the https://botsin.space Mastodon server and am currently awaiting approval for that. I have forked the code including @rveeblefetzer's PR #1 and pushed it to a Codeberg repository. I plan to have the bot check Erowid for new data and publish a new toot daily.

Be-ing commented 3 years ago

Unfortunately it looks like the ICO whitepaper submodule is offline now. Does anyone know of a current source collecting this data?

Be-ing commented 3 years ago

This is tangentially related, but you may also be interested in this study that used natural language processing techniques on Erowid trip reports to identify which drug the report was writing about.

diracdeltas commented 3 years ago

hi @Be-ing! cool idea. i don't have the PDFs from that repo, but i do have my text file extractions from the PDFs. i've uploaded that to https://github.com/diracdeltas/ico-whitepapers-text

FWIW i curated the output for the twitter account

let me know if you have any questions

Be-ing commented 3 years ago

i've uploaded that to https://github.com/diracdeltas/ico-whitepapers-text

That looks like it's the same as what is already in the corpi/ico_text folder in this repository.

FWIW i curated the output for the twitter account

Ah, I was wondering about that. Running the scripts locally, the output was not as consistently hilarious as the Twitter feed. Maybe I'll have it post more often, perhaps every hour, and curate it by boosting toots I like. I want to set it up so it will keep going indefinitely without anyone needing to take care of it.

Be-ing commented 3 years ago

Do you have any suggestions for free speech-to-text software? I want to add transcripts of videos and podcasts I have found searching the web for "cryptopsychedelic". I might expand it further adding text from other psychedelic web publications, conference recordings, and podcasts.

diracdeltas commented 3 years ago

That looks like it's the same as what is already in the corpi/ico_text folder in this repository.

ah yea, in that case you don't really need the original PDFs. i think i just deleted them locally after extracting the text.

Do you have any suggestions for free speech-to-text software? I want to add transcripts of videos and podcasts I have found searching the web for "cryptopsychedelic". I might expand it further adding text from other psychedelic web publications, conference recordings, and podcasts.

https://talonvoice.com/ maybe?