This project creates slack user bot for radio streaming via huddles on ubuntu VM. It's a kind of pet-project for idea testing, with straight forward code. Tested on Ubuntu 20.04.
On current setup I choose Liquidsoap as audio streaming, than route this stream via PulseAudio onto default input device. Than browser version of slack finds this input mic by default. Some bot logging in slack account as human and broadcasting in huddle. This browser automation uses node js with puppeteer bot.
Scp radio_to_mic.sh to your server. Then run
sudo apt-get -y update
sudo apt-get -y install pulseaudio pulseaudio-utils
pulseaudio --start
chmod +x radio_to_mic.sh
./radio_to_mic.sh
sudo apt-get -y install liquidsoap
mkdir liquidsoap
cd liquidsoap
touch daemon_pid.txt
Than copy to ~/liquidsoap some default.mp3 and radio.liq.
Start liquidsoap: liquidsoap radio.liq
Check created radio.log for successful start. See something like Streaming loop starts, synchronized by active sources
Test default input by recording it to file:
sudo apt-get -y install alsa-utils
arecord -f cd test.wav
# for google-chrome
wget -q -O - https://dl-ssl.google.com/linux/linux_signing_key.pub | sudo apt-key add -
sudo sh -c 'echo "deb [arch=amd64] http://dl.google.com/linux/chrome/deb/ stable main" >> /etc/apt/sources.list.d/google-chrome.list'
sudo apt-get update
sudo apt-get -y install nodejs npm google-chrome-stable xvfb
# setup display
Xvfb -ac :99 -screen 0 1280x1024x16 & export DISPLAY=:99
Copy repository, npm install.
Create .env file in project dir, create required envs from slack.js.
Start bot by npm start. Wait for connect and listen to your default.mp3 at slack huddle.
If bot can't connect, check login page and redirect onto channel page. You can do it on your local PC with GUI.
If bot connects, but silent, you can take some screenshots from audio preferences. Default mic must be available. If not found any - something wrong with pulseaudio setup. Check audio device available at pacmd list-source-outputs.