[rel-DB] Sequential numbers cleanup

[vkrasnovyd]

Integrates changes from OpenSlides/openslides-meta#319

Current changes in the generation logic:

• Generation with the deleted instances. Before: if the only instance of the collection is deleted and a new one is created, this new instance has sequential_number=1. Now: it has a higher value (next value from the postgres-sequence). Example: test_create_sequential_numbers_deleted_motion;
• motion.create_forwarded: data for amendment instances is created before the data for the lead_motion. As a result, forwarded lead_motion has higher sequential_number that its amendments. Example: test_allowed_to_forward_amendments_indirectly_complex.

Other changes:

• test_create_set_used_as_default_in_meeting_id. Reason - 1_1 trigger: now we can't have a meeting without used_as_default_projector_for_topic_in_meeting_id, so before the request we already have one.

[vkrasnovyd]

Check from models.yml up to permissions and jsons currently fails: regenerating models would lead to removal of the Option model and I'd prefer to deal with all these changes separately in #3167

[hjanott]

To make sure that in a meeting import with sequential number models old sequences are not applied. Just use some collection and assert that an unwanted value is overwritten with the next sequence number by the database anyway.

[hjanott]

Even though it's more lines, this way is so much easier to read. Love it.

[vkrasnovyd]

To make sure that in a meeting import with sequential number models old sequences are not applied. Just use some collection and assert that an unwanted value is overwritten with the next sequence number by the database anyway.

Out of all the affected collections only topic and motion_workflow have the import action and sequential_number is not allowed in the payloads of both of them. Added this check to the meeting.import test for the calculated fields.

[vkrasnovyd]

Updated the meeting.import test to check the changes from OpenSlides/openslides-meta@a1d976c

Added issue for the similar test in meta: OpenSlides/openslides-meta#336

[hjanott]

Looks good. Thank you.

[vkrasnovyd]

Needs changes from OpenSlides/openslides-meta#343