Enable timezone type for timestamp pyarrow type

src/fondant/core/schema.py

@@ -197,7 +197,13 @@ def from_dict(cls, json_schema: dict):
             fields = [(name, cls.from_dict(prop)) for name, prop in properties.items()]
             return cls.struct(fields)
 
+        if type_name == "timestamp":
+            return cls(type_name)
+
         if isinstance(type_name, str):
+            type_format = json_schema.get("format", None)
+            if type_format == "date-time":
+                return cls(pa.timestamp("us", tz="UTC"))
             return cls(type_name)
 
         msg = f"Invalid 'type' value: {type_name}"
@@ -218,6 +224,9 @@ def to_dict(self) -> dict:
             fields = [(field.name, Type(field.type).to_dict()) for field in self.value]
             return {"type": "object", "properties": dict(fields)}
 
+        elif isinstance(self.value, pa.TimestampType):
+            return {"type": "string", "format": "date-time"}
+
         type_ = None
         for type_name, data_type in _TYPES.items():
             if self.value.equals(data_type):

src/fondant/core/schemas/common.json

@@ -38,6 +38,10 @@
           "type": "string",
           "$ref": "#/definitions/subset_data_type"
         },
+        "format": {
+          "type": "string",
+          "description": "additional format information for the field"
+        },
         "properties": {
           "type": "object",
           "properties": {

tests/component/examples/component_specs/component.yaml

@@ -30,6 +30,9 @@ produces:
         number:
           type: int32
 
+  date:
+    type: string
+    format: date-time
 
 args:
   flag:

tests/component/test_component.py

@@ -55,6 +55,11 @@ def mocked_load_dataframe(self):
         return dd.from_dict(
             {
                 "images_data": [1, 2, 3],
+                "date": [
+                    "2024-02-29T12:30:45",
+                    "2024-02-29T12:30:45",
+                    "2024-02-29T12:30:45",
+                ],
                 "element": [
                     ("1", 1),
                     ("2", 2),

tests/core/test_schema.py

@@ -11,6 +11,10 @@ def test_valid_type():
     assert Type.list(Type("int8")).value == pa.list_(pa.int8())
     assert Type.list(Type.list(Type("string"))).value == pa.list_(pa.list_(pa.string()))
     assert Type("int8").to_dict() == {"type": "int8"}
+    assert Type(pa.timestamp("us", tz="UTC")).to_dict() == {
+        "type": "string",
+        "format": "date-time",
+    }
     assert Type.list("float32").to_dict() == {
         "type": "array",
         "items": {"type": "float32"},
@@ -42,6 +46,9 @@ def test_valid_json_schema():
     assert Type.from_dict(
         {"type": "array", "items": {"type": "array", "items": {"type": "int8"}}},
     ).value == pa.list_(pa.list_(pa.int8()))
+    assert Type.from_dict(
+        {"type": "string", "format": "date-time"},
+    ).value == pa.timestamp("us", tz="UTC")
     assert Type.from_dict(
         {
             "type": "object",