megaMUR_LG/chatbot/Readme.md

81 lines
1.6 KiB
Markdown
Raw Permalink Normal View History

# Setup and run
Within `chat` directory
pyenv local 3.8
python -m venv venv
source venv/bin/activate
pip install -r requirements.txt
Install ESpeak, see below
**Set config:**
In `chat` folder copy file `config-sample.json` to new file with name `config.json` and set especially the OpenAI API Key,
the other values can be kept as default.
See section "Recording Audio" for getting correct Device ID on Raspberry. You can execute `record_audio.py` to see a list of available devices with IDs.
**Run:**
Execute `main.py` e.g.
python -m main
# Recording Audio
## Python package PyAudio
Installed via pip
For more examples, e.g. conversion into MP3 see
https://realpython.com/playing-and-recording-sound-python/#recording-audio
## Raspberry Pi
Setup microphone and Device Id:
https://makersportal.com/blog/2018/8/23/recording-audio-on-the-raspberry-pi-with-python-and-a-usb-microphone
# Speech Synthesis
## Raspberry/Linux: Pico TTS
https://github.com/cyberbotics/picotts
Pico TTS for Raspberry:
https://cstan.io/post/2020/03/pico-tts-unter-raspbian-uebersetzen/
## Mac
### ESpeak
See https://espeak.sourceforge.net/
Verbesserung: MBrola Voices für ESpeak (https://github.com/numediart/MBROLA)
Install
brew install espeak
Test
espeak -vde -p20 -k20 "Hallo Welt"
MBRola für Mac: https://github.com/pettarin/setup-festival-mbrola
### Maybe Pico TTS could run on Mac
https://github.com/cyberbotics/webots/wiki/Pico-Compilation
# Audio for Raspberry
https://naomiproject.github.io/docs/
# Similar projects
https://github.com/anatolybazarov/oracle.sh