feat!: switch to Phenopackets v2 #1394

[v-rocheleau]

Updating JSON-Schema representations and datamodel from Phenopackets V1 to V2.

Phenopackets related changes:

• Every phenopacket JSON-Schema was updated to reflect the changes described here
• JSON-Schema for VRS (GA4GH's Variation Representation Specification) entities are now in full synch with the official spec as a git submodule
• VRS includes cyclic schema references, which cannot be rendered as a fully resolved schema object. To fix this, the schema validation now relies on the referencing API to resolve VRS schemas dynamically
• Django models/serializers/filters/search/views/renderers were updated to hold the new information carried in phenopackets
• Existing django tests were updated to pass after the new changes
• New management command validate can be used to validate a phenopackets in a .json document
• New management command decamelize can be used to convert the keys of a .json document (was useful to create and convert test data, which is not decamelized if saved to the db by bypassing the ingestion workflow)
• New example .json files were added as sample and test resources

Other changes:

• New docker-compose.dev.yaml file can be used to quickly spin a Postgres DB and a simple Adminer app on port 8080. (useful for local dev on standalone katsu)
• Documentation for the docker-compose.dev.yaml tools
• New migrations
• Make all JSON-Schema IDs URI compliant
• Field level data validation for Phenopackets enums (Django choices)

[davidlougheed]

just realized this, sorry - any model for which entries aren't automatically removed by foreign key relationships needs cleanup functions + tests to make sure objects are properly removed.

[davidlougheed]

lgtm pending any search fixes needed!