revert(assets): drop job_ids filter from GET /api/assets (#14408)

The job_ids query filter added in #13998 has no live consumer: the
frontend Generated tab kept sourcing from GET /jobs, and the cloud side
removed its equivalent filter from the shared asset spec. Carrying it on
the local server only re-introduces Core<->Cloud drift on the shared
contract, so remove it to match.

Removed: the job_ids field + validator on ListAssetsQuery, the IN(...)
clauses in list_references_page, the service/route passthrough, and the
filter-only tests.

Kept: the canonical-UUID prompt_id enforcement at job creation (also
landed in #13998). It stands on its own -- job ids are matched verbatim
by history keys, websocket correlation, and /interrupt -- and cloud
inherits it by running core for execution, so no divergence is created.

tests-unit/assets_test/queries/test_asset_info.py

@@ -158,56 +158,6 @@ class TestListReferencesPage:
         refs, _, _ = list_references_page(session, sort="name", order="asc")
         assert refs[0].name == "large"
 
-    def test_job_ids_filter(self, session: Session):
-        asset = _make_asset(session, "hash1")
-        job_a = str(uuid.uuid4())
-        job_b = str(uuid.uuid4())
-        ref_a = _make_reference(session, asset, name="from_job_a")
-        ref_a.job_id = job_a
-        ref_b = _make_reference(session, asset, name="from_job_b")
-        ref_b.job_id = job_b
-        _make_reference(session, asset, name="no_job")
-        session.commit()
-
-        # Single job filter
-        refs, _, total = list_references_page(session, job_ids=[job_a])
-        assert total == 1
-        assert refs[0].name == "from_job_a"
-
-        # Multi-job filter (IN)
-        refs, _, total = list_references_page(session, job_ids=[job_a, job_b])
-        names = sorted(r.name for r in refs)
-        assert total == 2
-        assert names == ["from_job_a", "from_job_b"]
-
-        # Unknown job id matches nothing
-        refs, _, total = list_references_page(session, job_ids=[str(uuid.uuid4())])
-        assert total == 0
-        assert refs == []
-
-        # Empty/None means no filter -> all three references
-        refs, _, total = list_references_page(session, job_ids=[])
-        assert total == 3
-        refs, _, total = list_references_page(session, job_ids=None)
-        assert total == 3
-
-    def test_job_ids_combined_with_other_filters(self, session: Session):
-        asset = _make_asset(session, "hash1")
-        job_a = str(uuid.uuid4())
-        ref_match = _make_reference(session, asset, name="match.bin")
-        ref_match.job_id = job_a
-        ref_wrong_name = _make_reference(session, asset, name="other.bin")
-        ref_wrong_name.job_id = job_a
-        ref_wrong_job = _make_reference(session, asset, name="match.bin")
-        ref_wrong_job.job_id = str(uuid.uuid4())
-        session.commit()
-
-        refs, _, total = list_references_page(
-            session, job_ids=[job_a], name_contains="match"
-        )
-        assert total == 1
-        assert refs[0].id == ref_match.id
-
 
 class TestFetchReferenceAssetAndTags:
     def test_returns_none_for_nonexistent(self, session: Session):

tests-unit/assets_test/queries/test_list_assets_query.py

@@ -1,60 +0,0 @@
-"""Schema-level unit tests for ListAssetsQuery (no DB required)."""
-import uuid
-
-import pytest
-from pydantic import ValidationError
-
-from app.assets.api.schemas_in import ListAssetsQuery
-
-
-class TestJobIdsValidator:
-    def test_csv_string_parses_and_canonicalizes(self):
-        a = "AAAAAAAA-BBBB-CCCC-DDDD-EEEEEEEEEEEE"
-        b = "11111111-2222-3333-4444-555555555555"
-        q = ListAssetsQuery.model_validate({"job_ids": f"{a},{b}"})
-        # Canonicalized to lowercase
-        assert q.job_ids == [a.lower(), b]
-
-    def test_repeated_query_params_as_list(self):
-        a = "11111111-1111-1111-1111-111111111111"
-        b = "22222222-2222-2222-2222-222222222222"
-        q = ListAssetsQuery.model_validate({"job_ids": [a, b]})
-        assert q.job_ids == [a, b]
-
-    def test_dedup_preserves_first_seen_order(self):
-        a = "11111111-1111-1111-1111-111111111111"
-        b = "22222222-2222-2222-2222-222222222222"
-        q = ListAssetsQuery.model_validate({"job_ids": [a, b, a]})
-        assert q.job_ids == [a, b]
-
-    def test_default_empty(self):
-        q = ListAssetsQuery.model_validate({})
-        assert q.job_ids == []
-
-    def test_invalid_uuid_rejected(self):
-        with pytest.raises(ValidationError) as exc:
-            ListAssetsQuery.model_validate({"job_ids": "not-a-uuid"})
-        assert "must be UUIDs" in str(exc.value)
-
-    def test_non_string_list_item_rejected(self):
-        with pytest.raises(ValidationError) as exc:
-            ListAssetsQuery.model_validate(
-                {"job_ids": ["11111111-1111-1111-1111-111111111111", 42]}
-            )
-        assert "must be strings" in str(exc.value)
-
-    def test_non_string_non_list_value_rejected(self):
-        with pytest.raises(ValidationError) as exc:
-            ListAssetsQuery.model_validate({"job_ids": {"bad": "shape"}})
-        assert "must be a string or list of strings" in str(exc.value)
-
-    def test_max_length_enforced(self):
-        too_many = [str(uuid.uuid4()) for _ in range(501)]
-        with pytest.raises(ValidationError) as exc:
-            ListAssetsQuery.model_validate({"job_ids": too_many})
-        assert exc.value.errors()[0]["type"] == "too_long"
-
-    def test_max_length_boundary_accepted(self):
-        at_cap = [str(uuid.uuid4()) for _ in range(500)]
-        q = ListAssetsQuery.model_validate({"job_ids": at_cap})
-        assert len(q.job_ids) == 500

tests-unit/assets_test/test_prompt_id_enforcement.py

@@ -1,9 +1,9 @@
 """POST /prompt enforces canonical-UUID job ids at creation time.
 
-Lives in assets_test because it uses this suite's booted-server fixture and
-because the invariant exists for the assets pipeline: the GET /api/assets
-``job_ids`` filter matches stored job ids exactly, so a job minted with a
-non-canonical id would produce assets the filter can never find.
+Lives in assets_test because it uses this suite's booted-server fixture. The
+invariant itself is pipeline-wide: a job id is stored and compared verbatim
+downstream — history keys, websocket correlation, and /interrupt matching —
+so a job minted with a non-canonical id would miss every exact-match lookup.
 
 The prompt bodies here are intentionally invalid workflows — prompt_id
 validation happens before workflow validation, so a rejected id returns