Extend CI/CD and produce deployable packages

Summary

Right now our CI/CD pipeline covers code with unit tests, but few things are still missing:

• docker container is rather old and is not kept up to date
• installation procedure isn't straightforward
• we do not provide a deployable package
• integration tests are missing

What is the expected correct behavior?

1. External dependencies that are temporarily included in our codebase have to be factored out:

• molecularprofiles: see https://gitlab.cta-observatory.org/cta-array-elements/ccf/mdps/-/issues/9
• rdams client for the NCAR GDAS access: see https://github.com/NCAR/rda-apps-clients/issues/9

2. Docker container must be updated

• Consider nightly builds with the "develop" version of the ctapipe
• Consider stable builds with the latest stable release of the ctapipe
• Consider different linux distros (CC7, Rocky Linux, Alma Linux)
• Consider interactive container creation

3. The calibpipe packaging must be implemented

• Dependencies must be cleaned
• The package must be available at PyPI
• Installation instructions must be easy, clear, and as automated as possible

4. Continuous integration

• Implement automated integration tests with CWL (for the moment)
  • May require certain service accounts
• Consider cvmfs deployment of the code for further integration with other DPPS modules

Relevant logs and/or screenshots