feat(assessment): Consolidate assessment_run table

backend/app/alembic/versions/086_fold_assessment_run_into_assessment.py

@@ -0,0 +1,155 @@
+"""Fold assessment_run into assessment
+
+Revision ID: 086
+Revises: 085
+Create Date: 2026-10-06 00:00:00.000000
+
+The child was 1:1 with its parent everywhere, so its config + runtime columns move
+onto `assessment` and the table is dropped. `error_message` lands in
+`execution.error`; keys the code no longer reads are stripped. `input` goes: nothing
+writes it since the console pipeline was retired. Downgrade recreates `assessment_run`
+empty: take a `pg_dump -t assessment_run` before upgrading, the rows come back from it.
+An assessment with no run fails the `SET NOT NULL` on purpose rather than migrating half.
+"""
+
+import sqlalchemy as sa
+from alembic import op
+from sqlalchemy.dialects import postgresql
+
+revision = "086"
+down_revision = "085"
+branch_labels = None
+depends_on = None
+
+CONFIG_INDEX = "idx_assessment_config"
+CONFIG_FK = "fk_assessment_config_id_config"
+RUN_TABLE = "assessment_run"
+# Derivable from config_version or already in the stage dumps.
+DROPPED_EXECUTION_KEYS = (
+    "stage_output_urls",
+    "provider",
+    "model",
+    "input_schema",
+    "verdicts",
+    "gate_passed",
+)
+
+
+def upgrade() -> None:
+    op.add_column(
+        "assessment",
+        sa.Column("config_id", postgresql.UUID(as_uuid=True), nullable=True),
+    )
+    op.add_column(
+        "assessment", sa.Column("config_version", sa.Integer(), nullable=True)
+    )
+    op.add_column(
+        "assessment",
+        sa.Column(
+            "total_items",
+            sa.Integer(),
+            nullable=True,
+            comment="Submitted row count; sizes the result without reading the rows",
+        ),
+    )
+    op.add_column(
+        "assessment",
+        sa.Column(
+            "execution",
+            postgresql.JSONB(astext_type=sa.Text()),
+            nullable=True,
+            comment="Staged-batch runtime (AssessmentExecution); NULL for RESPONSE",
+        ),
+    )
+
+    strip = " ".join(f"- '{key}'" for key in DROPPED_EXECUTION_KEYS)
+    op.execute(
+        f"""
+        UPDATE assessment AS a
+        SET config_id = r.config_id,
+            config_version = r.config_version,
+            total_items = r.total_items,
+            execution = CASE
+                WHEN r.error_message IS NULL THEN r.execution {strip}
+                ELSE (COALESCE(r.execution, '{{}}'::jsonb) {strip})
+                     || jsonb_build_object('error', r.error_message)
+            END
+        FROM {RUN_TABLE} AS r
+        WHERE r.assessment_id = a.id
+          AND a.config_id IS NULL
+        """
+    )
+
+    op.alter_column("assessment", "config_id", nullable=False)
+    op.alter_column("assessment", "config_version", nullable=False)
+    op.alter_column("assessment", "total_items", nullable=False)
+    op.create_foreign_key(CONFIG_FK, "assessment", "config", ["config_id"], ["id"])
+    op.create_index(CONFIG_INDEX, "assessment", ["config_id", "config_version"])
+
+    op.drop_column("assessment", "input")
+    op.drop_table(RUN_TABLE)
+
+
+def downgrade() -> None:
+    op.create_table(
+        RUN_TABLE,
+        sa.Column("id", sa.Integer(), primary_key=True),
+        sa.Column(
+            "assessment_id",
+            postgresql.UUID(as_uuid=True),
+            sa.ForeignKey(
+                "assessment.id",
+                name="fk_assessment_run_assessment_id",
+                ondelete="CASCADE",
+            ),
+            nullable=False,
+        ),
+        sa.Column(
+            "config_id",
+            postgresql.UUID(as_uuid=True),
+            sa.ForeignKey("config.id", name="fk_assessment_run_config_id"),
+            nullable=False,
+        ),
+        sa.Column("config_version", sa.Integer(), nullable=False),
+        sa.Column(
+            "status",
+            postgresql.ENUM(name="assessment_status", create_type=False),
+            server_default="PENDING",
+            nullable=False,
+        ),
+        sa.Column(
+            "batch_job_id",
+            sa.Integer(),
+            sa.ForeignKey(
+                "batch_job.id",
+                name="fk_assessment_run_batch_job_id",
+                ondelete="SET NULL",
+            ),
+            nullable=True,
+        ),
+        sa.Column("total_items", sa.Integer(), server_default="0", nullable=False),
+        sa.Column("execution", postgresql.JSONB(astext_type=sa.Text()), nullable=True),
+        sa.Column(
+            "post_processing_config",
+            postgresql.JSONB(astext_type=sa.Text()),
+            nullable=True,
+        ),
+        sa.Column("error_message", sa.Text(), nullable=True),
+        sa.Column("inserted_at", sa.DateTime(), nullable=False),
+        sa.Column("updated_at", sa.DateTime(), nullable=False),
+    )
+    op.create_index("idx_assessment_run_assessment_id", RUN_TABLE, ["assessment_id"])
+    op.create_index(
+        "idx_assessment_run_config", RUN_TABLE, ["config_id", "config_version"]
+    )
+    op.create_index("idx_assessment_run_status", RUN_TABLE, ["status"])
+    op.add_column(
+        "assessment",
+        sa.Column("input", postgresql.JSONB(astext_type=sa.Text()), nullable=True),
+    )
+    op.drop_index(CONFIG_INDEX, table_name="assessment")
+    op.drop_constraint(CONFIG_FK, "assessment", type_="foreignkey")
+    op.drop_column("assessment", "execution")
+    op.drop_column("assessment", "total_items")
+    op.drop_column("assessment", "config_version")
+    op.drop_column("assessment", "config_id")

backend/app/api/routes/assessment/__init__.py

@@ -1,7 +1,7 @@
 from fastapi import APIRouter, Depends
 
 from app.api.permissions import require_feature
-from app.api.routes.assessment import assessments, datasets, runs
+from app.api.routes.assessment import datasets
 from app.core.feature_flags import FeatureFlag
 
 router = APIRouter(
@@ -11,7 +11,5 @@
 )
 
 router.include_router(datasets.router)
-router.include_router(assessments.router)
-router.include_router(runs.router)
 
 __all__ = ["router"]

backend/app/api/routes/assessment/api.py

@@ -85,7 +85,7 @@ def list_assessments(
     offset: Annotated[int, Query(ge=0)] = 0,
 ) -> APIResponse[list[AssessmentSummary]]:
     """BATCH assessments newest-first, optionally narrowed to one config or version."""
-    rows = api_crud.list_assessments_with_execution(
+    rows = api_crud.list_assessments(
         session=session,
         organization_id=auth_context.organization_.id,
         project_id=auth_context.project_.id,
@@ -96,8 +96,8 @@ def list_assessments(
     )
     return APIResponse.success_response(
         data=[
-            results.build_summary(assessment, execution, submission_name)
-            for assessment, execution, submission_name in rows
+            results.build_summary(assessment, submission_name)
+            for assessment, submission_name in rows
         ]
     )
 

backend/app/celery/tasks/job_execution.py

@@ -15,6 +15,7 @@
 import logging
 from collections.abc import Callable
 from typing import TYPE_CHECKING, TypeVar
+from uuid import UUID
 
 from asgi_correlation_id import correlation_id
 from celery import Task, current_task
@@ -370,7 +371,7 @@ def run_assessment_pipeline(
 @gevent_timeout(settings.CELERY_TASK_SOFT_TIME_LIMIT, "run_assessment_api_batch")
 def run_assessment_api_batch(
     self,
-    execution_id: int,
+    assessment_id: str,
     organization_id: int,
     project_id: int,
     trace_id: str,
@@ -389,15 +390,15 @@ def run_assessment_api_batch(
     result = _run_with_otel_parent(
         self,
         lambda: run_batch_stage(
-            execution_id=execution_id,
+            assessment_id=UUID(assessment_id),
             organization_id=organization_id,
             project_id=project_id,
         ),
     )
     if result and result.get("requeue"):
         self.apply_async(
             kwargs={
-                "execution_id": execution_id,
+                "assessment_id": assessment_id,
                 "organization_id": organization_id,
                 "project_id": project_id,
                 "trace_id": trace_id,