Include PyTimber in the NXCALS nightly view

As discussed today in the SWAN meeting, it would be useful to have a version of PyTimber available to users out-of-the-box. It is also the case that Pytimber and NXCALS development is very active and is seeing very regular releases, therefore it is not currently practical to crystalise the versions in a fixed LCG stack.

There is already an NXCALS view which is built nightly on top of a standard LCG stack, and it is proposed that the installation of PyTimber be included in this process. It looks like this might happen in https://gitlab.cern.ch/sft/lcgjenkins/-/blob/master/nxcals_view_transaction.sh (also mentioned in https://lcgdocs.web.cern.ch/lcgdocs/lcg/specialviews/#nxcals).

Unlike NXCALS, it is recommended to install pytimber via the standard Python package manger. The following steps will install pytimber (and its dependencies) into the site-packages directory and will result in around 300 JARs being installed in the lib/python*/site-packages/cmmnbuild_dep_manager/lib directory:

# Install pytimber and its Python dependencies

python -m pip install --index-url https://acc-py-repo.cern.ch/repository/vr-py-releases/simple --trusted-host acc-py-repo.cern.ch --upgrade pytimber==3.*

# Install pytimber's Java dependencies

python -m cmmnbuild_dep_manager resolve

For the record, this will result in a different set of JARs to the one that is provided by the NXCALS bundle. This is expected and desired, we do not forsee there being a collision as PySpark is runing Py4J (a separate process) whereas pytimber is using JPype (JVM in the same process as the Python interpreter).

cc @piotr.sowinski as the responsible for PyTimber