local_batch_job.py ensure all integrations are in-memory or not available (=no errors)

[JeroenVerstraelen]

EPIC: #806

• ensure all integrations are in-memory or not available (=no errors)
  • zookeeper, ejr, etl, elastic, hadoop, yarn, kubernetes, sentinelhub, aws, custom processes

[EmileSonneveld]

The InMemoryJobRegistry is used. No zookeeper.
InMemoryJobRegistry removed the need of elastic.
The batchjob get launched directly, without yarn/hadoop or kubernetes.
To launch an openEO backend that can submit batchjobs, some form of kubernetes would be needed. Maybe minicube, or otherwise some kind of stub.
the layercatalog is cleared in the current implementation of local batch_jobs. So no references to sentinelhub or aws are used.
Maybe the the batch job results need to be written trough an S3 mocker like boto3
For 'etl' and 'custom processes', I don't know the status

[EmileSonneveld]

Note: run_graph_locally works without any internet connection.