Telegram bot, that helps transcribe and then sum up content of your voices, audios and meetings.
- clone repo
- install dependances using poetry
app/bot.py
to run service- there is also dockerfile, but it is mostly for prod
- limitation:
- audiofiles more than 25mb
- chunking
- silence
- archiving
- there is limit (technocal (4096 utf) and cognitive) on size of message sent back -- need to cut it wisely
- audiofiles more than 25mb
- beautifying answers
- DevOps:
- docker
- refact it to 2 sage build with only venv
- docker compose
- CI/CD -- github action on push/merge
- administration
- logging
- docker
- performanse
- ffmpeg - takes time for conversion
- use python instead of cli
- stream io -- less iops
- ffmpeg - takes time for conversion
- social functioanality
- user accounting
- statistic
- count channels to expect number of speaking persons