Clone this repo
git clone git@github.com:francisco-perez-sorrosal/satnbot.git
Install project dependencies
cd satnbot
poetry install
docker build -t "satnbot" .
docker run -it --rm -e DISCORD_TOKEN="MTA-xxx" -e OPENAI_API_KEY="sk-xxx" satnbot
I'm assuming you have your fancy Discord account! 😃
Assign the right permissions to let your bot do things! 🥶
3.1. In OAuth2's URL generator, select the scopes (e.g. bot
and application.command
)
3.2 Then select the bot permissions (e.g. administrator
)
See TODOs below
Edit the tdb.sh
and put your tokens/keys there:
DISCORD_TOKEN="<YOUR_DISCORD_APP_TOKEN_HERE>"
OPENAI_API_KEY="<YOUR_OPENAI_KEY_HERE>"
./tdb.sh
Bear in mind when using the bot capabilities below that:
Both limitations can be workaround but this is not done at this point. Will be addressed later
@fpsbot what should I do today apart from my daily learning duties reading about quantum physics, biomechatronics, synthetic biology and the wsj?
Yeah... I guess my life sucks...
There's also a slash command (/history
) to check the history or requests...
...so you can ask again (with /h <idx_no>
) to get another thrilling response..
You can use this thig also with the slash command /syf
(syf: Stats Yahoo Finance) slash command to get a stats report of the tickers of the day from Yahoo Finance
By default you get a human readable output...
...but if you send the command with the parameter format
as tsv /syf tsv
it will return a tsv-formatted file for you to digest wisefully with a High Speed ML trading algorithm to become rich some day (Mental note: TODO ASAP).
In summary, basically this does to do what this engineer/scientist (representing many of us) had to do very painfully here
Important Note When passing the text extracted from the pdf document to the bot with the /ax
command input has to be limited when sending it to ChatGPT. This can be done with the parameters chunks
and chars_per_chunk
available in the slash command.
You can use the bot also with the slash command /axs
(arxiv sanity) to access the latest articles in the topic you specify:
/axs filter_tags cs.AI
Or you can retrieve stuff on two+ different topics:
/axs filter_tags cs.AI&cs.CV
Then you can copy the Arxiv ID and then use the /ax
(arxiv) slash command to retrieve a summary of the article:
/ax arxiv_id: 2202.02098
You can specify also the language
(english is default) in which you want the summary and the style
(paragraph(default)|bulletpoints|sonnet) and number of style_items
/ax arxiv_id: 2202.02098 language: english style: bulletpoints style_items: 10
/ax arxiv_id:2202.02098 language:spanish style:paragraph style_items:2
For things here to do before I grow old or before I die see this.