cocoindex 0.1.82__cp313-cp313-macosx_10_12_x86_64.whl → 0.2.0__cp313-cp313-macosx_10_12_x86_64.whl

cocoindex/convert.py

@@ -14,7 +14,6 @@ from typing import Any, Callable, Mapping, Type, get_origin
 import numpy as np
 
 from .typing import (
-    KEY_FIELD_NAME,
     TABLE_TYPES,
     AnalyzedAnyType,
     AnalyzedBasicType,
@@ -28,7 +27,6 @@ from .typing import (
     encode_enriched_type,
     is_namedtuple_type,
     is_numpy_number_type,
-    is_struct_type,
 )
 
 
@@ -88,38 +86,35 @@ def make_engine_value_encoder(type_info: AnalyzedTypeInfo) -> Callable[[Any], An
 
             return encode_struct_list
 
-    if isinstance(variant, AnalyzedDictType):
-        if not variant.value_type:
-            return lambda value: value
+        # Otherwise it's a vector, falling into basic type in the engine.
 
+    if isinstance(variant, AnalyzedDictType):
         value_type_info = analyze_type_info(variant.value_type)
-        if isinstance(value_type_info.variant, AnalyzedStructType):
-
-            def encode_struct_dict(value: Any) -> Any:
-                if not isinstance(value, dict):
-                    return value
-                if not value:
-                    return []
-
-                sample_key, sample_val = next(iter(value.items()))
-                key_type, val_type = type(sample_key), type(sample_val)
-
-                # Handle KTable case
-                if value and is_struct_type(val_type):
-                    key_encoder = (
-                        make_engine_value_encoder(analyze_type_info(key_type))
-                        if is_struct_type(key_type)
-                        else make_engine_value_encoder(ANY_TYPE_INFO)
-                    )
-                    value_encoder = make_engine_value_encoder(
-                        analyze_type_info(val_type)
-                    )
-                    return [
-                        [key_encoder(k)] + value_encoder(v) for k, v in value.items()
-                    ]
-                return {key_encoder(k): value_encoder(v) for k, v in value.items()}
+        if not isinstance(value_type_info.variant, AnalyzedStructType):
+            raise ValueError(
+                f"Value type for dict is required to be a struct (e.g. dataclass or NamedTuple), got {variant.value_type}. "
+                f"If you want a free-formed dict, use `cocoindex.Json` instead."
+            )
+        value_encoder = make_engine_value_encoder(value_type_info)
+
+        key_type_info = analyze_type_info(variant.key_type)
+        key_encoder = make_engine_value_encoder(key_type_info)
+        if isinstance(key_type_info.variant, AnalyzedBasicType):
 
-            return encode_struct_dict
+            def encode_row(k: Any, v: Any) -> Any:
+                return [key_encoder(k)] + value_encoder(v)
+
+        else:
+
+            def encode_row(k: Any, v: Any) -> Any:
+                return key_encoder(k) + value_encoder(v)
+
+        def encode_struct_dict(value: Any) -> Any:
+            if not value:
+                return []
+            return [encode_row(k, v) for k, v in value.items()]
+
+        return encode_struct_dict
 
     if isinstance(variant, AnalyzedStructType):
         struct_type = variant.struct_type
@@ -132,8 +127,8 @@ def make_engine_value_encoder(type_info: AnalyzedTypeInfo) -> Callable[[Any], An
             field_names = [f.name for f in fields]
 
             def encode_dataclass(value: Any) -> Any:
-                if not dataclasses.is_dataclass(value):
-                    return value
+                if value is None:
+                    return None
                 return [
                     encoder(getattr(value, name))
                     for encoder, name in zip(field_encoders, field_names)
@@ -154,8 +149,8 @@ def make_engine_value_encoder(type_info: AnalyzedTypeInfo) -> Callable[[Any], An
             ]
 
             def encode_namedtuple(value: Any) -> Any:
-                if not is_namedtuple_type(type(value)):
-                    return value
+                if value is None:
+                    return None
                 return [
                     encoder(getattr(value, name))
                     for encoder, name in zip(field_encoders, field_names)
@@ -248,25 +243,47 @@ def make_engine_value_decoder(
                         f"declared `{dst_type_info.core_type}`, a dict type expected"
                     )
 
-                key_field_schema = engine_fields_schema[0]
-                field_path.append(f".{key_field_schema.get('name', KEY_FIELD_NAME)}")
-                key_decoder = make_engine_value_decoder(
-                    field_path,
-                    key_field_schema["type"],
-                    analyze_type_info(key_type),
-                    for_key=True,
-                )
-                field_path.pop()
+                num_key_parts = src_type.get("num_key_parts", 1)
+                key_type_info = analyze_type_info(key_type)
+                key_decoder: Callable[..., Any] | None = None
+                if (
+                    isinstance(
+                        key_type_info.variant, (AnalyzedBasicType, AnalyzedAnyType)
+                    )
+                    and num_key_parts == 1
+                ):
+                    single_key_decoder = make_engine_value_decoder(
+                        field_path,
+                        engine_fields_schema[0]["type"],
+                        key_type_info,
+                        for_key=True,
+                    )
+
+                    def key_decoder(value: list[Any]) -> Any:
+                        return single_key_decoder(value[0])
+
+                else:
+                    key_decoder = make_engine_struct_decoder(
+                        field_path,
+                        engine_fields_schema[0:num_key_parts],
+                        key_type_info,
+                        for_key=True,
+                    )
                 value_decoder = make_engine_struct_decoder(
                     field_path,
-                    engine_fields_schema[1:],
+                    engine_fields_schema[num_key_parts:],
                     analyze_type_info(value_type),
                 )
 
                 def decode(value: Any) -> Any | None:
                     if value is None:
                         return None
-                    return {key_decoder(v[0]): value_decoder(v[1:]) for v in value}
+                    return {
+                        key_decoder(v[0:num_key_parts]): value_decoder(
+                            v[num_key_parts:]
+                        )
+                        for v in value
+                    }
 
         return decode
 

cocoindex/flow.py

@@ -105,18 +105,26 @@ def _spec_kind(spec: Any) -> str:
 
 def _transform_helper(
     flow_builder_state: _FlowBuilderState,
-    fn_spec: FunctionSpec,
+    fn_spec: FunctionSpec | Callable[..., Any],
     transform_args: list[tuple[Any, str | None]],
     name: str | None = None,
 ) -> DataSlice[Any]:
-    if not isinstance(fn_spec, FunctionSpec):
+    if isinstance(fn_spec, FunctionSpec):
+        kind = _spec_kind(fn_spec)
+        spec = fn_spec
+    elif callable(fn_spec) and (
+        op_kind := getattr(fn_spec, "__cocoindex_op_kind__", None)
+    ):
+        kind = op_kind
+        spec = op.EmptyFunctionSpec()
+    else:
         raise ValueError("transform() can only be called on a CocoIndex function")
 
     return _create_data_slice(
         flow_builder_state,
         lambda target_scope, name: flow_builder_state.engine_flow_builder.transform(
-            _spec_kind(fn_spec),
-            dump_engine_object(fn_spec),
+            kind,
+            dump_engine_object(spec),
             transform_args,
             target_scope,
             flow_builder_state.field_name_builder.build_name(
@@ -245,7 +253,7 @@ class DataSlice(Generic[T]):
             f(scope)
 
     def transform(
-        self, fn_spec: op.FunctionSpec, *args: Any, **kwargs: Any
+        self, fn_spec: op.FunctionSpec | Callable[..., Any], *args: Any, **kwargs: Any
     ) -> DataSlice[Any]:
         """
         Apply a function to the data slice.
@@ -513,7 +521,7 @@ class FlowBuilder:
         )
 
     def transform(
-        self, fn_spec: FunctionSpec, *args: Any, **kwargs: Any
+        self, fn_spec: FunctionSpec | Callable[..., Any], *args: Any, **kwargs: Any
     ) -> DataSlice[Any]:
         """
         Apply a function to inputs, returning a DataSlice.

cocoindex/op.py

@@ -388,53 +388,39 @@ def executor_class(**args: Any) -> Callable[[type], type]:
     return _inner
 
 
-class _EmptyFunctionSpec(FunctionSpec):
+class EmptyFunctionSpec(FunctionSpec):
     pass
 
 
 class _SimpleFunctionExecutor:
-    spec: Any
+    spec: Callable[..., Any]
 
     def prepare(self) -> None:
-        self.__call__ = self.spec.__call__
+        self.__call__ = staticmethod(self.spec)
 
 
-def function(**args: Any) -> Callable[[Callable[..., Any]], FunctionSpec]:
+def function(**args: Any) -> Callable[[Callable[..., Any]], Callable[..., Any]]:
     """
     Decorate a function to provide a function for an op.
     """
     op_args = OpArgs(**args)
 
-    def _inner(fn: Callable[..., Any]) -> FunctionSpec:
+    def _inner(fn: Callable[..., Any]) -> Callable[..., Any]:
         # Convert snake case to camel case.
-        op_name = "".join(word.capitalize() for word in fn.__name__.split("_"))
+        op_kind = "".join(word.capitalize() for word in fn.__name__.split("_"))
         sig = inspect.signature(fn)
-        full_name = f"{fn.__module__}.{fn.__qualname__}"
-
-        # An object that is both callable and can act as a FunctionSpec.
-        class _CallableSpec(_EmptyFunctionSpec):
-            __call__ = staticmethod(fn)
-
-            def __reduce__(self) -> str | tuple[Any, ...]:
-                return full_name
-
-        _CallableSpec.__name__ = op_name
-        _CallableSpec.__doc__ = fn.__doc__
-        _CallableSpec.__qualname__ = fn.__qualname__
-        _CallableSpec.__module__ = fn.__module__
-        callable_spec = _CallableSpec()
-
+        fn.__cocoindex_op_kind__ = op_kind  # type: ignore
         _register_op_factory(
             category=OpCategory.FUNCTION,
             expected_args=list(sig.parameters.items()),
             expected_return=sig.return_annotation,
             executor_factory=_SimpleFunctionExecutor,
-            spec_loader=lambda: callable_spec,
-            op_kind=op_name,
+            spec_loader=lambda: fn,
+            op_kind=op_kind,
             op_args=op_args,
         )
 
-        return callable_spec
+        return fn
 
     return _inner
 

cocoindex/sources.py

@@ -2,6 +2,7 @@
 
 from . import op
 from .auth_registry import TransientAuthEntryReference
+from .setting import DatabaseConnectionSpec
 import datetime
 
 
@@ -67,3 +68,22 @@ class AzureBlob(op.SourceSpec):
 
     sas_token: TransientAuthEntryReference[str] | None = None
     account_access_key: TransientAuthEntryReference[str] | None = None
+
+
+class Postgres(op.SourceSpec):
+    """Import data from a PostgreSQL table."""
+
+    _op_category = op.OpCategory.SOURCE
+
+    # Table name to read from (required)
+    table_name: str
+
+    # Database connection reference (optional - uses default if not provided)
+    database: TransientAuthEntryReference[DatabaseConnectionSpec] | None = None
+
+    # Optional: specific columns to include (if None, includes all columns)
+    included_columns: list[str] | None = None
+
+    # Optional: column name to use for ordinal tracking (for incremental updates)
+    # Should be a timestamp, serial, or other incrementing column
+    ordinal_column: str | None = None

cocoindex/typing.py

@@ -330,35 +330,50 @@ def analyze_type_info(t: Any) -> AnalyzedTypeInfo:
 
 def _encode_struct_schema(
     struct_type: type, key_type: type | None = None
-) -> dict[str, Any]:
+) -> tuple[dict[str, Any], int | None]:
     fields = []
 
-    def add_field(name: str, t: Any) -> None:
+    def add_field(name: str, analyzed_type: AnalyzedTypeInfo) -> None:
         try:
-            type_info = encode_enriched_type_info(analyze_type_info(t))
+            type_info = encode_enriched_type_info(analyzed_type)
         except ValueError as e:
             e.add_note(
                 f"Failed to encode annotation for field - "
-                f"{struct_type.__name__}.{name}: {t}"
+                f"{struct_type.__name__}.{name}: {analyzed_type.core_type}"
             )
             raise
         type_info["name"] = name
         fields.append(type_info)
 
+    def add_fields_from_struct(struct_type: type) -> None:
+        if dataclasses.is_dataclass(struct_type):
+            for field in dataclasses.fields(struct_type):
+                add_field(field.name, analyze_type_info(field.type))
+        elif is_namedtuple_type(struct_type):
+            for name, field_type in struct_type.__annotations__.items():
+                add_field(name, analyze_type_info(field_type))
+        else:
+            raise ValueError(f"Unsupported struct type: {struct_type}")
+
+    result: dict[str, Any] = {}
+    num_key_parts = None
     if key_type is not None:
-        add_field(KEY_FIELD_NAME, key_type)
+        key_type_info = analyze_type_info(key_type)
+        if isinstance(key_type_info.variant, AnalyzedBasicType):
+            add_field(KEY_FIELD_NAME, key_type_info)
+            num_key_parts = 1
+        elif isinstance(key_type_info.variant, AnalyzedStructType):
+            add_fields_from_struct(key_type_info.variant.struct_type)
+            num_key_parts = len(fields)
+        else:
+            raise ValueError(f"Unsupported key type: {key_type}")
 
-    if dataclasses.is_dataclass(struct_type):
-        for field in dataclasses.fields(struct_type):
-            add_field(field.name, field.type)
-    elif is_namedtuple_type(struct_type):
-        for name, field_type in struct_type.__annotations__.items():
-            add_field(name, field_type)
+    add_fields_from_struct(struct_type)
 
-    result: dict[str, Any] = {"fields": fields}
+    result["fields"] = fields
     if doc := inspect.getdoc(struct_type):
         result["description"] = doc
-    return result
+    return result, num_key_parts
 
 
 def _encode_type(type_info: AnalyzedTypeInfo) -> dict[str, Any]:
@@ -374,7 +389,7 @@ def _encode_type(type_info: AnalyzedTypeInfo) -> dict[str, Any]:
         return {"kind": variant.kind}
 
     if isinstance(variant, AnalyzedStructType):
-        encoded_type = _encode_struct_schema(variant.struct_type)
+        encoded_type, _ = _encode_struct_schema(variant.struct_type)
         encoded_type["kind"] = "Struct"
         return encoded_type
 
@@ -384,10 +399,8 @@ def _encode_type(type_info: AnalyzedTypeInfo) -> dict[str, Any]:
         if isinstance(elem_type_info.variant, AnalyzedStructType):
             if variant.vector_info is not None:
                 raise ValueError("LTable type must not have a vector info")
-            return {
-                "kind": "LTable",
-                "row": _encode_struct_schema(elem_type_info.variant.struct_type),
-            }
+            row_type, _ = _encode_struct_schema(elem_type_info.variant.struct_type)
+            return {"kind": "LTable", "row": row_type}
         else:
             vector_info = variant.vector_info
             return {
@@ -402,12 +415,14 @@ def _encode_type(type_info: AnalyzedTypeInfo) -> dict[str, Any]:
             raise ValueError(
                 f"KTable value must have a Struct type, got {value_type_info.core_type}"
             )
+        row_type, num_key_parts = _encode_struct_schema(
+            value_type_info.variant.struct_type,
+            variant.key_type,
+        )
         return {
             "kind": "KTable",
-            "row": _encode_struct_schema(
-                value_type_info.variant.struct_type,
-                variant.key_type,
-            ),
+            "row": row_type,
+            "num_key_parts": num_key_parts,
         }
 
     if isinstance(variant, AnalyzedUnionType):

{cocoindex-0.1.82.dist-info → cocoindex-0.2.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: cocoindex
-Version: 0.1.82
+Version: 0.2.0
 Requires-Dist: click>=8.1.8
 Requires-Dist: rich>=14.0.0
 Requires-Dist: python-dotenv>=1.1.0

{cocoindex-0.1.82.dist-info → cocoindex-0.2.0.dist-info}/RECORD

@@ -1,23 +1,23 @@
-cocoindex-0.1.82.dist-info/METADATA,sha256=totDI7Pxf_1q8U_5YRvAq0npG0A0loaeUpM7PeKvfkU,12188
-cocoindex-0.1.82.dist-info/WHEEL,sha256=NvZaK6sFPuu8Uh9tMXGKWxMrGLxBro5TQniO6rRq4wQ,106
-cocoindex-0.1.82.dist-info/entry_points.txt,sha256=_NretjYVzBdNTn7dK-zgwr7YfG2afz1u1uSE-5bZXF8,46
-cocoindex-0.1.82.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+cocoindex-0.2.0.dist-info/METADATA,sha256=UqU2UILgJFn7J42_8ohdY0MOewTw4X5l_rmmclMP6kM,12187
+cocoindex-0.2.0.dist-info/WHEEL,sha256=iXfRWk7-127zCPB-_BNFDQE-qLd9Rsj-fJMRKNRg-kg,106
+cocoindex-0.2.0.dist-info/entry_points.txt,sha256=_NretjYVzBdNTn7dK-zgwr7YfG2afz1u1uSE-5bZXF8,46
+cocoindex-0.2.0.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
 cocoindex/__init__.py,sha256=sLpSVO5Cotgn_82lawxvXnaqfa-qj33rytWBAe2MTtU,2201
-cocoindex/_engine.cpython-313-darwin.so,sha256=focgTojoap8Q40ZHM_pXRQqK7ng44FJmWWQbZeTmoZ8,68725524
+cocoindex/_engine.cpython-313-darwin.so,sha256=kg6EHfT1E6POzhNdToBDS42lHPIfrCrRK1KHwqm4Qjs,68876620
 cocoindex/auth_registry.py,sha256=PE1-kVkcyC1G2C_V7b1kvYzeq73OFQehWKQP7ln7fJ8,1478
 cocoindex/cli.py,sha256=8Q2D0gI_yHorxfFC7tzV96oJAR7PtTn_T4qWLorPPoM,20675
-cocoindex/convert.py,sha256=bf61K1gEYhZx05LQtV_-E7SDOUYpiYaHZtH8iZ1W_VE,21473
-cocoindex/flow.py,sha256=W0PjZ0vLQFEW6tvbblje85jnkz4Mn9B1_Z4_slUffYM,35981
+cocoindex/convert.py,sha256=x3eYx-NQD-z9h-jmpVgUMiSTqll6Baa_SfZJF3mZSUc,22071
+cocoindex/flow.py,sha256=7133Z6QU4LF_GMV81AmkO_aGhfAQCStwOnMW7BABymU,36258
 cocoindex/functions.py,sha256=09erNt3WbzY9l1KER-akBF2O5-6xEahV2ORBECaL6yk,12260
 cocoindex/index.py,sha256=j93B9jEvvLXHtpzKWL88SY6wCGEoPgpsQhEGHlyYGFg,540
 cocoindex/lib.py,sha256=f--9dAYd84CZosbDZqNW0oGbBLsY3dXiUTR1VrfQ_QY,817
 cocoindex/llm.py,sha256=Pv_cdnRngTLtuLU9AUmS8izIHhcKVnuBNolC33f9BDI,851
-cocoindex/op.py,sha256=kdcWAOvvMxTVTinOZK6cVi8tY59G6VOdVuTV8fUJy0A,22714
+cocoindex/op.py,sha256=w9CnrCEE6P5Voo2LDvMDlYPrBQuzNC5FfOHP6WF0Buw,22241
 cocoindex/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 cocoindex/runtime.py,sha256=povilB3HH3y1JF-yxKwU-pD8n2WnAqyQxIgvXXHNc60,1080
 cocoindex/setting.py,sha256=n5ipJ_uMFZBh5-Ib_gyjOqS_XPXsQwTFVYEuSAnoscg,5283
 cocoindex/setup.py,sha256=7uIHKN4FOCuoidPXcKyGTrkqpkl9luL49-6UcnMxYzw,3068
-cocoindex/sources.py,sha256=69COA4qbZDipzGYfXv-WJSmicFkA509xIShRGDh6A0A,2083
+cocoindex/sources.py,sha256=o57FZARXaLKuQr4nKYLj9kxXBRX3cWEfDhFG3lldxoE,2779
 cocoindex/subprocess_exec.py,sha256=davnBvbw119NBaFX6s9IyHW9MTZyrYYuKYex0HManHY,7902
 cocoindex/targets.py,sha256=Nfh_tpFd1goTnS_cxBjIs4j9zl3Z4Z1JomAQ1dl3Sic,2796
 cocoindex/tests/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -26,8 +26,8 @@ cocoindex/tests/test_optional_database.py,sha256=snAmkNa6wtOSaxoZE1HgjvL5v_ylitt
 cocoindex/tests/test_transform_flow.py,sha256=MdIzLdIMBPlIfKRFAB7n8gIZixqChL6xb4vScvcyiNU,5148
 cocoindex/tests/test_typing.py,sha256=9OF3lO2uSpZBefkEJx7WRbnkXjwQtvlQIeeARYQID68,12391
 cocoindex/tests/test_validation.py,sha256=X6AQzVs-hVKIXcrHMEMQnhfUE8at7iXQnPq8nHNhZ2Q,4543
-cocoindex/typing.py,sha256=qQ0ANF3iuQDeSqipHgL2SDiiXL2reTMUN0aj4ve_T0w,13359
+cocoindex/typing.py,sha256=lEQYIzAGVKQp6RnhyeopY9Q7xEED7yQj3ZMxvTPblV8,14200
 cocoindex/user_app_loader.py,sha256=o2n25y7CGw9Qrb-RN6tiNlv2YJ9YLM8NSMhJZjLhhKQ,1914
 cocoindex/utils.py,sha256=hUhX-XV6XGCtJSEIpBOuDv6VvqImwPlgBxztBTw7u0U,598
 cocoindex/validation.py,sha256=PZnJoby4sLbsmPv9fOjOQXuefjfZ7gmtsiTGU8SH-tc,3090
-cocoindex-0.1.82.dist-info/RECORD,,
+cocoindex-0.2.0.dist-info/RECORD,,

{cocoindex-0.1.82.dist-info → cocoindex-0.2.0.dist-info}/WHEEL

@@ -1,4 +1,4 @@
 Wheel-Version: 1.0
-Generator: maturin (1.9.3)
+Generator: maturin (1.9.4)
 Root-Is-Purelib: false
 Tag: cp313-cp313-macosx_10_12_x86_64