Saraswati

================

- Add guidelines for running Saraswati on Docker
- Fix program flow when no ``--upload`` option is given

================

- Use 5 minutes recording chunk size as default
- Protect against running out of disk space
- Add uploader based on ``rsync``. Thanks Roh and Diren!

================

- Adjust recording pipeline: Add ``audioconvert ! queue`` GStreamer elements after audio source
- Improve documentation
- Fix creating spool directory
- Improve installation convenience by not depending on ``PyGObject-stubs`` at runtime

================

- Add project infrastructure
- Adjust to API of GStreamer 1.18.4
- Run software tests on CI/GHA
- Refactor recorder program into "saraswati" module
- Use ISO8601-like timestamps again
- Significantly improve package and program structure
- Make all relevant parameters configurable
- Add command line entrypoint ``saraswati``. Invoke ``saraswati --help`` in
order to find out about its usage.
- Improve documentation
- Add ``CHANGES.rst``
- Ensure spool directory exists by automatically creating it
- Fix selecting a custom recording location (--spool)

================

- Add basic example for demonstrating the GStreamer Python bindings
- Honor GStreamer 1.10.4
- Store audio chunks to /var/spool/saraswati
- Use chunk length of 10 seconds
- Run on BeagleBone
- Use ``.mka`` as filename extension for audio-only Matroska files


bump2version==1.0.0
twine==3.4.1
keyring==23.0.1