Update python to 3.10

Python 3.7 went EOL earlier this year. Update to py3.10 to get back onto a supported version. py3.10 was chosen as it is the newest version of python supported by pyspark 3.1. We will need to upgrade to spark 3.4+ to use a newer version of python, but we should prefer to stay on spark 3.1.2 to match production hosts.

While attempting to update python the conda dependency resolver gets stuck. Updated CI to use the conda-analytics package, the same one used in production, which has a newer dependency resolver that doesn't get stuck.

Includes an update of maven to 3.9.5. This allows us to enable the BF resolver collector which downloads pom's in parallel. This reduced time in CI by > 8 minutes, previously it was timing out at 20 minutes.

Additionally moves us over to the wmcs gitlab runners. The job was getting SIGKILL depending on which runner it happened to land on due to overrunning memory limits. Switching to the wmcs runners gives higher memory availability, along with larger timeouts.

Bug: T346373