feat: add airflow operator and hook for ClickHouse (#3699)

Author: Spar9a

docs/integrations/engines/clickhouse.md

@@ -394,6 +394,26 @@ If a model has many records in each partition, you may see additional performanc
 ## Local/Built-in Scheduler
 **Engine Adapter Type**: `clickhouse`
 
+## Airflow Scheduler
+**Engine Name:** `clickhouse`
+
+In order to share a common implementation across local and Airflow, SQLMesh ClickHouse implements its own hook and operator.
+
+By default, the connection ID is set to `sqlmesh_clickhouse_default`, but can be overridden using the `engine_operator_args` parameter to the `SQLMeshAirflow` instance as in the example below:
+```python linenums="1"
+from sqlmesh.schedulers.airflow import NO_DEFAULT_CATALOG
+
+sqlmesh_airflow = SQLMeshAirflow(
+    "clickhouse",
+    default_catalog=NO_DEFAULT_CATALOG,
+    engine_operator_args={
+        "sqlmesh_clickhouse_conn_id": "<Connection ID>"
+    },
+)
+```
+
+Note: `NO_DEFAULT_CATALOG` is required for ClickHouse since ClickHouse doesn't support catalogs. 
+
 ### Connection options
 
 | Option               | Description                                                                                                                                                               |  Type  | Required |

sqlmesh/schedulers/airflow/hooks/clickhouse.py

@@ -0,0 +1,34 @@
+from __future__ import annotations
+
+import typing as t
+
+from airflow.providers.common.sql.hooks.sql import DbApiHook
+
+if t.TYPE_CHECKING:
+    from clickhouse_connect.dbapi.connection import Connection
+
+
+class SQLMeshClickHouseHook(DbApiHook):
+    """
+    Uses the ClickHouse Python DB API connector.
+    """
+
+    conn_name_attr = "sqlmesh_clickhouse_conn_id"
+    default_conn_name = "sqlmesh_clickhouse_default"
+    conn_type = "sqlmesh_clickhouse"
+    hook_name = "SQLMesh ClickHouse"
+
+    def get_conn(self) -> Connection:
+        """Returns a ClickHouse connection object"""
+        from clickhouse_connect.dbapi import connect
+
+        db = self.get_connection(getattr(self, t.cast(str, self.conn_name_attr)))
+
+        return connect(
+            host=db.host,
+            port=db.port,
+            username=db.login,
+            password=db.password,
+            database=db.schema,
+            **db.extra_dejson,
+        )

sqlmesh/schedulers/airflow/operators/clickhouse.py

@@ -0,0 +1,32 @@
+from __future__ import annotations
+
+import typing as t
+
+
+from sqlmesh.schedulers.airflow.hooks.clickhouse import SQLMeshClickHouseHook
+from sqlmesh.schedulers.airflow.operators.base import BaseDbApiOperator
+from sqlmesh.schedulers.airflow.operators.targets import BaseTarget
+
+
+class SQLMeshClickHouseOperator(BaseDbApiOperator):
+    """The operator that evaluates a SQLMesh model snapshot on a ClickHouse target
+
+    Args:
+        target: The target that will be executed by this operator instance.
+        postgres_conn_id: The Airflow connection id for the postgres target.
+    """
+
+    def __init__(
+        self,
+        *,
+        target: BaseTarget,
+        clickhouse_conn_id: str = SQLMeshClickHouseHook.default_conn_name,
+        **kwargs: t.Any,
+    ) -> None:
+        super().__init__(
+            target=target,
+            conn_id=clickhouse_conn_id,
+            dialect="clickhouse",
+            hook_type=SQLMeshClickHouseHook,
+            **kwargs,
+        )

sqlmesh/schedulers/airflow/util.py

@@ -122,6 +122,10 @@ def discover_engine_operator(name: str, sql_only: bool = False) -> t.Type[BaseOp
     name = name.lower()
 
     try:
+        if name == "clickhouse":
+            from sqlmesh.schedulers.airflow.operators.clickhouse import SQLMeshClickHouseOperator
+
+            return SQLMeshClickHouseOperator
         if name == "spark":
             from sqlmesh.schedulers.airflow.operators.spark_submit import (
                 SQLMeshSparkSubmitOperator,