feat(prevent): add table for prevent AI configs

[JerrySentry]

Add a configuration model to control prevent AI product features (code review, test analytics, bug prediction).

[github-actions]

This PR has a migration; here is the generated SQL for src/sentry/prevent/migrations/0001_create_prevent_ai_configuration.py

for 0001_create_prevent_ai_configuration in prevent

--
-- Create model PreventAIConfiguration
--
CREATE TABLE "prevent_ai_configuration" ("id" bigint NOT NULL PRIMARY KEY GENERATED BY DEFAULT AS IDENTITY, "date_updated" timestamp with time zone NOT NULL, "date_added" timestamp with time zone NOT NULL, "integration_id" bigint NULL, "data" jsonb NOT NULL, "organization_id" bigint NOT NULL);
CREATE UNIQUE INDEX CONCURRENTLY "prevent_ai_configuration_organization_id_integrat_2daf0f56_uniq" ON "prevent_ai_configuration" ("organization_id", "integration_id");
ALTER TABLE "prevent_ai_configuration" ADD CONSTRAINT "prevent_ai_configuration_organization_id_integrat_2daf0f56_uniq" UNIQUE USING INDEX "prevent_ai_configuration_organization_id_integrat_2daf0f56_uniq";
ALTER TABLE "prevent_ai_configuration" ADD CONSTRAINT "prevent_ai_configura_organization_id_69ec9fa2_fk_sentry_or" FOREIGN KEY ("organization_id") REFERENCES "sentry_organization" ("id") DEFERRABLE INITIALLY DEFERRED NOT VALID;
ALTER TABLE "prevent_ai_configuration" VALIDATE CONSTRAINT "prevent_ai_configura_organization_id_69ec9fa2_fk_sentry_or";
CREATE INDEX CONCURRENTLY "prevent_ai_configuration_integration_id_a6f19c01" ON "prevent_ai_configuration" ("integration_id");
CREATE INDEX CONCURRENTLY "prevent_ai_configuration_organization_id_69ec9fa2" ON "prevent_ai_configuration" ("organization_id");

[suejung-sentry]

I looked over other migrations in the repo and it seems like most follow the convention of sentry_{thingwithnospaces}. Should we update this to match, or is there something that automagically formats it for us?

[JerrySentry]

its called prevent_ai_configuration now as per this suggestion, not sure about the "thingwithnospaces", some tables has multiple underscores but most tables has 1..
#101831 (comment)

[suejung-sentry]

Looks like if integration is nullable and part of the unique_together, we'll have to make sure on application side not to write multiple rows with organization_id=abc and integration_id=NULL (probably can just disallow writing if integration is null)

[wedamija]

The unique index actually won't enforce (organization_id=x, integration_id=null) as unique, because in postgres every null value is considered unique. But I'm confused as to what the use of these rows is when there's no integration?