Resolve warning in Dataset Alias migration

This PR resolves an SQLAlchemy warning in the migration by correctly setting the `alias_id` column as part of the primary key in the `dag_schedule_dataset_alias_reference` table. Previously, only dag_id was marked as the primary key, causing a mismatch with the local definition, which triggered an SAWarning.

Example: https://github.com/apache/airflow/actions/runs/11526187767/job/32090094094?pr=43243#step:6:745

```
/opt/airflow/airflow/migrations/versions/0026_2_10_0_dag_schedule_dataset_alias_reference.py:46 SAWarning: Table 'dag_schedule_dataset_alias_reference' specifies columns 'dag_id' as primary_key=True, not matching locally specified columns 'alias_id', 'dag_id'; setting the current primary key columns to 'alias_id', 'dag_id'. This warning may become an exception in a future release
```

I verified that both columns are marked as primary keys already in 2.10.

airflow/migrations/versions/0026_2_10_0_dag_schedule_dataset_alias_reference.py

@@ -45,7 +45,7 @@ def upgrade():
     """Add dag_schedule_dataset_alias_reference table."""
     op.create_table(
         "dag_schedule_dataset_alias_reference",
-        sa.Column("alias_id", sa.Integer(), nullable=False),
+        sa.Column("alias_id", sa.Integer(), primary_key=True, nullable=False),
         sa.Column("dag_id", StringID(), primary_key=True, nullable=False),
         sa.Column("created_at", airflow.utils.sqlalchemy.UtcDateTime(timezone=True), nullable=False),
         sa.Column("updated_at", airflow.utils.sqlalchemy.UtcDateTime(timezone=True), nullable=False),

docs/apache-airflow/img/airflow_erd.sha256

@@ -1 +1 @@
-5067b6eac290220d23ce1bef6ce6d02478757f626f5a81a53ca2bba8e31c094c
+5958e108ee31f4ca94d6a878767d8d1ca00ef4e664ecda9308b367c173a09226