Enforce Dataset Filters via Configuration File

[Damien Ayers]

At the moment, DEA is filtering datasets that get sent to the Queue which Alchemist is processing. For example, datasets with low geometric accuracy, or low maturity aren't supposed to be processed and aren't added to the queue.

While this works okay most of the time, it has several flaws:

• The configuration is split between the config file, and the infrastructure which is stored somewhere completely different.
• It's hard to reprocess, since the filtering must be done by hand.
• It's possible to process datasets which shouldn't be processed!

We need to allow configuring a dataset filter inside the configuration file.

A possible way of doing this is using the same --predicate python form as is used on the add-missing-to-queue CLI mode.