feat: Cohere Telegraph

[alterxyz]

Kindly try my own bot https://t.me/tealmoon_ai_v1_bot

• Integrate Cohere's proactive and native web search capability
• Output the entire LLM message with its source link on Telegraph
• Add README file with tutorials on obtaining Cohere and Telegraph API keys

The new feature need:

pip install beautifulsoup4 # for html to Telegraph foramte
pip install Markdown # for markdown to html
pip install cohere # Cohere

The requirements.txt and env variables are not set up yet.

[yihong0618]

nice

[yihong0618]

Nice work

[yihong0618]

@alterxyz can we use something like

https://github.com/python273/telegraph or https://github.com/aiogram/aiograph

that user can easily use it? since telegraph do not need to login

[alterxyz]

What do you mean by easily use it?

Can you make some simple example about that?

[yihong0618]

will dig it seems these package do not need cookie

[alterxyz]

Tbh creating a telegraph page is just simple request.

The issue there is convert text(markdown) to the telegra.ph format](https://telegra.ph/api#NodeElement), that why there is import for markdown and beautiful.

pip install Markdown
 pip install beautifulsoup4

A quick look is both those two does not did that.

Our latest code does not need account or cookie or token.

It will auto generate a account and store it (to a json file... not sure is it good / best practice or not) if there is no env telegraph token.

(we can even skip the store account if needed, like everytime restart it we generate a new account, if we do not need edit it in future, and btw the telegraph page can not be delete but can be edit if we have the token.)

[yihong0618]

copy that and make sense

[yihong0618]

Thanks