feat(ingest, backend): do not approve curated revisions (by users other than insdc_ingest_user), send notification if ingest wants to revise a curated sequence

[anna-parker]

resolves #3087, #3084, #2613

partially: #3085,

preview URL: https://insdc-curation-ingest-qui.loculus.org/

Summary

1. Ingest no longer approves all revisions in the insdc_ingest_group but only revisions it submitted (e.g. user= insdc_ingest_user) - this allows curators to revise a sequence and not have it be approved directly by ingest. In order to do this the backend now allows filtering by submitter not just submissionGroup
2. Send slack notification if ingest wants to revise a curated sequence - see if revision is in a sequence where any version was submitted by a user other than insdc_ingest_user - in this case send a notification and put the sequence into state blocked. Again the backend must be modified to also return the submitter of the original-metadata.

Screenshot

PR Checklist

• Requires feat(website): add revise form #3111 to be submitted
• Confirm the state (metadata) is the same as currently using comparison script: Checkout pathoplexus/pathoplexus@ce68b92

micromamba activate pp-integrity
cd tests/regression-testing 
snakemake results/ebola-zaire.meta.main.insdc-curation-ingest-qui.diff 
snakemake results/cchf.meta.main.insdc-curation-ingest-qui.diff 
snakemake results/west-nile.meta.main.insdc-curation-ingest-qui.diff 

There is no diff in the ebola-zaire, cchf and west nile metadata.

• Curate three sequences via the revise page as super-user: curate metadata, curate sequence, curate metadata and sequence -> check these are not overwritten by ingest
• Make sure ingest doesn't approve curated sequences
• Connect to preview db - change original data hash of a curated sequence (three cases as above), make sure that we get a notification that ingest doesn't know what to do and will not revise these sequences.
  I see this in logs:

15:24:26  WARNING (   compare_hashes.py:  63) - Sequence MZ424862 has been curated before - do not know how to proceed

Somehow the slack hook was removed from the config, I had to add it back. Now we are notified:

• Double check slack hook is still in pathoplexus ingest config

[anna-parker]

Testing:

1. LOC_000373F.1: curated metadata -> stable
2. LOC_000372H.1: curated sequence -> stable
3. LOC_00036WV.1: both -> stable
4. LOC_000RSPC.2: both also multi-segmented -> stable
5. Curated LOC_000371K.1 but did not approve - should not show up
6. Then edited hash of LOC_000373F and LOC_000RSPC for v1 and v2 in db to test notification system for single and multi-segmented
7. Modify hash of non-curated LOC_000RSNE and LOC_00036ZP to trigger v2

[corneliusroemer]

Very cool stuff, this was on my backlog and I'm very grateful you've done already so much more! Still reviewing, but wanted to ask what you think about adding/modifying one or two backend tests to make the new behaviour tested?

[corneliusroemer]

This looks great, backend changes are very minimal, would just be good to keep things tested in backend.

Python is good, I didn't verify things super closely, but I trust your testing, the changes seem to be fairly small, mostly a refactor - the new logic is very localized I think, simply bailing when a sequence was ever touched by non-insdc_ingest_user.