feat: Support arbitrary properties for x-sql-datatype in x-sql-datatype-properties

Signed-off-by: Edgar Ramírez-Mondragón <[email protected]>

Author: edgarrmondragon

docs/guides/sql-target.md

@@ -64,13 +64,19 @@ You can register new type handlers for the `x-sql-datatype` extension:
 
 ```python
 from sqlalchemy.types import SMALLINT
+from my_sqlalchemy_dialect import VectorType
+
+
+def vector_to_sql(jsonschema: dict) -> VectorType:
+    return VectorType(**jsonschema.get("x-sql-datatype-properties", {}))
 
 
 class MyConnector(SQLConnector):
     @functools.cached_property
     def jsonschema_to_sql(self):
         to_sql = JSONSchemaToSQL()
         to_sql.register_sql_datatype_handler("smallint", SMALLINT)
+        to_sql.register_sql_datatype_handler("vector", custom_vector_to_sql)
         return to_sql
 ```
 
@@ -86,6 +92,10 @@ plugins:
       addresses:
         number:
           x-sql-datatype: smallint
+      values:
+        x-sql-datatype: vector
+        x-sql-datatype-properties:
+          dim: 10
 ```
 ````
 

singer_sdk/singerlib/schema.py

@@ -46,6 +46,7 @@
     "oneOf",
     # JSON Schema extensions
     "x-sql-datatype",
+    "x-sql-datatype-properties",
 ]
 
 
@@ -89,6 +90,7 @@ class Schema:
     contentEncoding: str | None = None  # noqa: N815
     # JSON Schema extensions
     x_sql_datatype: str | None = None
+    x_sql_datatype_properties: dict[str, t.Any] | None = None
 
     deprecated: bool | None = None
     oneOf: t.Any | None = None  # noqa: N815

tests/singerlib/test_schema.py

@@ -28,6 +28,14 @@
     "additionalProperties": True,
     "required": ["a_string"],
 }
+SCHEMA_WITH_CUSTOM_SQL_DATATYPE = Schema(
+    x_sql_datatype="vector",
+    x_sql_datatype_properties={"dimensions": 10},
+)
+SCHEMA_WITH_CUSTOM_SQL_DATATYPE_DICT = {
+    "x-sql-datatype": "vector",
+    "x-sql-datatype-properties": {"dimensions": 10},
+}
 
 
 @pytest.mark.parametrize(
@@ -53,6 +61,11 @@
             OBJECT_DICT,
             id="object_to_dict",
         ),
+        pytest.param(
+            SCHEMA_WITH_CUSTOM_SQL_DATATYPE,
+            SCHEMA_WITH_CUSTOM_SQL_DATATYPE_DICT,
+            id="schema_with_custom_sql_datatype_to_dict",
+        ),
     ],
 )
 def test_schema_to_dict(schema, expected):
@@ -82,6 +95,11 @@ def test_schema_to_dict(schema, expected):
             OBJECT_SCHEMA,
             id="schema_from_object_dict",
         ),
+        pytest.param(
+            SCHEMA_WITH_CUSTOM_SQL_DATATYPE_DICT,
+            SCHEMA_WITH_CUSTOM_SQL_DATATYPE,
+            id="schema_from_custom_sql_datatype_dict",
+        ),
     ],
 )
 def test_schema_from_dict(pydict, expected):