Removed home counts and new source count

[ajstewart]

• The expensive counts have been removed from the homepage.
• Removed the sky coverage map from the homepage.
• The new sources count has been removed from the run detail page.
• n_new_sources column added to run model.
• migration file will calculate n_new_sources to populate new column.
• Updated scheme diagrams.
• Fixed link to jupyterhub.

Note: Currently the time field of the pipeline run is set to auto_now which means that the time that is reflected on the pipeline run is updated whenever the model object itself is updated. This means that running the migration in this PR will update the 'time run' of each job to the time of the migration.

Fixes #605.

[ajstewart]

@ddobie for whenever is convenient this PR removes the expensive count queries on the home page and also the count of the new sources on the run detail page.

The migration has been setup to fill in a new n_new_sources column in the run model with one off counts so that migration will take a little bit to run.

Also note that currently all tests are failing on GitHub as our test instance doesn't respect the poetry lock file and instead installs the latest versions with respect to the conditions set out in the pyproject file. This means that an update of something is not currently compatible with the pipeline as it is. All my tests pass locally however with an environment installed from the lock file.

[ddobie]

I will trust that it works :p