refactor: unify data preparations and research input snapshots

2026-09-12 01:24:02 +08:00
parent 849f86fef7
commit 394438e753
82 changed files with 4146 additions and 2076 deletions
@@ -0,0 +1,75 @@
+"""Replace saved dataset inputs with editable preparations and independent snapshots."""
+
+import sqlalchemy as sa
+from alembic import op
+
+revision = "0015"
+down_revision = "0014"
+branch_labels = None
+depends_on = None
+
+
+def upgrade():
+    op.drop_table("template_inputs")
+    op.add_column("catalog_batches", sa.Column("job_id", sa.String(36), nullable=True))
+    op.add_column("catalog_batches", sa.Column("offset", sa.Integer(), nullable=False, server_default="0"))
+    op.execute("UPDATE catalog_batches SET job_id = id")
+    # Preserve unfinished catalog pagination; only legacy research inputs are discarded.
+    jobs = sa.table("sync_jobs", sa.column("id"), sa.column("checkpoint", sa.JSON()))
+    batches = sa.table("catalog_batches", sa.column("id"), sa.column("offset", sa.Integer()))
+    for job_id, checkpoint in op.get_bind().execute(sa.select(jobs.c.id, jobs.c.checkpoint)):
+        offset = (checkpoint or {}).get("offset", 0)
+        if type(offset) is int and offset >= 0:
+            op.get_bind().execute(batches.update().where(batches.c.id == job_id).values(offset=offset))
+    names = {"fk": "fk_%(table_name)s_%(column_0_name)s_%(referred_table_name)s"}
+    fk = next(
+        f
+        for f in sa.inspect(op.get_bind()).get_foreign_keys("catalog_batches")
+        if f["constrained_columns"] == ["id"]
+    )
+    with op.batch_alter_table("catalog_batches", naming_convention=names) as batch:
+        batch.drop_constraint(fk["name"] or "fk_catalog_batches_id_sync_jobs", type_="foreignkey")
+        batch.alter_column("job_id", existing_type=sa.String(36), nullable=False)
+        batch.create_foreign_key("fk_catalog_batches_job", "sync_jobs", ["job_id"], ["id"])
+        batch.create_index("ix_catalog_batches_job_id", ["job_id"])
+    op.create_table(
+        "data_preparations",
+        sa.Column("id", sa.String(36), primary_key=True),
+        sa.Column("name", sa.String(200), nullable=False),
+        sa.Column("note", sa.Text(), nullable=False),
+        sa.Column("scope_key", sa.String(200), nullable=False),
+        sa.Column("scope", sa.JSON(), nullable=False),
+        sa.Column("version", sa.Integer(), nullable=False),
+        sa.Column("created_at", sa.DateTime(timezone=True), nullable=False),
+        sa.Column("updated_at", sa.DateTime(timezone=True), nullable=False),
+    )
+    op.create_index("ix_data_preparations_scope_key", "data_preparations", ["scope_key"])
+    op.create_table(
+        "preparation_fields",
+        sa.Column(
+            "preparation_id",
+            sa.String(36),
+            sa.ForeignKey("data_preparations.id", ondelete="CASCADE"),
+            primary_key=True,
+        ),
+        sa.Column("field_id", sa.String(200), primary_key=True),
+        sa.Column("dataset_id", sa.String(200), nullable=False),
+        sa.Column("content", sa.JSON(), nullable=False),
+    )
+    op.create_index("ix_preparation_fields_dataset_id", "preparation_fields", ["dataset_id"])
+    op.create_table(
+        "research_input_snapshots",
+        sa.Column("id", sa.String(36), primary_key=True),
+        sa.Column("preparation_id", sa.String(36), nullable=False),
+        sa.Column("preparation_version", sa.Integer(), nullable=False),
+        sa.Column("content", sa.JSON(), nullable=False),
+        sa.Column("created_at", sa.DateTime(timezone=True), nullable=False),
+        sa.UniqueConstraint("preparation_id", "preparation_version"),
+    )
+    op.create_index(
+        "ix_research_input_snapshots_preparation_id", "research_input_snapshots", ["preparation_id"]
+    )
+
+
+def downgrade():
+    raise RuntimeError("旧输入模型已移除；回退请恢复升级前数据库备份")