Telegram bot chatGPT + audio speech

Improvement of functionality of the TG bot via ChatGRT API

The main code is located on personal hosting, but it needs to be improved and fixed, namely: 

1. Refine the request for buttons

- rephrase correctly and briefly and translate into business English 

- rephrase briefly in a bulleted list 

- create meeting notes briefly in a bulleted list 

Add three new buttons: 

- NEW! add a new  button for free requests, so that questions different from these commands can be sent in the chat  

- NEW! Audio transcription 

- NEW! Document translation [EN, RU..]

2. Refine audio transcription 

- The bot is connected to the Speech GPT transcriber Gladia, however, when requested it a) transcribes only 5% of the audio length regardless of the file size   b) does not separate the text by speakers.. meaning that one transcription can have 3-5-7 speakers ranging from 20 minutes to 2 hours. It is necessary to either change the AI itself or understand why it does not separate speakers separately (the functionality is uploaded, but there is an error that crashes the code)

2.1 For audio text recognition

- it is necessary to check the language of the text in which language the audio is (Russian/English/Polish/German   (may be default)

- the function should be able to recognize multilingual audio, is this possible? For example, the audio can be in Russian, as well as in English, there may be replicas in Polish and German 

3. Develop text and file translation functionality

- it is necessary to use multi-account support from deepl and switch between them when the limit is reached (limit 500,000 characters, need to check)

- Text translation functionality from English to Russian and vice versa, like in DeepL

- Supported file formats: .pptx, .doc, Google Docs (such as .xlsx, .docx).

- the output file should be in the same format, so if it was ppxt, we get ppxt. 

- The file name should be the same as the original + stack [translation language]

- After processing the file, enter a message that processing is complete (currently this is done through manual verification, via the "check" button)

the bot itself, there may be a need to create another bot through botphaser and connect if necessary. I do not have access to botphaser.. the previous programmer disappeared, possibly due to events in the country.

I will provide access to the code on hosting and provide a link to the current bot.