vgi-python 0.8.3__tar.gz → 0.8.5__tar.gz

{vgi_python-0.8.3 → vgi_python-0.8.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: vgi-python
-Version: 0.8.3
+Version: 0.8.5
 Summary: Vector Gateway Interface - Connect DuckDB to external programs via Apache Arrow
 Project-URL: Homepage, https://query.farm
 Project-URL: Repository, https://github.com/Query-farm/vgi-python

{vgi_python-0.8.3 → vgi_python-0.8.5}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "vgi-python"
-version = "0.8.3"
+version = "0.8.5"
 description = "Vector Gateway Interface - Connect DuckDB to external programs via Apache Arrow"
 readme = "README.md"
 keywords = [

{vgi_python-0.8.3 → vgi_python-0.8.5}/tests/catalog/test_declarative.py

@@ -1207,6 +1207,20 @@ class TestCatalogDescriptor:
         catalog = Catalog(name="myapp", schemas=[main, analytics])
         assert len(catalog.schemas) == 2
 
+    def test_catalog_source_url_defaults_to_none(self) -> None:
+        """Catalog source_url is None unless set."""
+        catalog = Catalog(name="myapp", schemas=[Schema(name="main")])
+        assert catalog.source_url is None
+
+    def test_catalog_source_url_settable(self) -> None:
+        """Catalog source_url can advertise the repo/docs homepage."""
+        catalog = Catalog(
+            name="myapp",
+            schemas=[Schema(name="main")],
+            source_url="https://github.com/example/myapp",
+        )
+        assert catalog.source_url == "https://github.com/example/myapp"
+
 
 class TestCatalogValidation:
     """Tests for Catalog validation."""
@@ -1311,6 +1325,25 @@ class TestReadOnlyCatalogWithCatalog:
         infos = catalog_interface.catalogs()
         assert [i.name for i in infos] == ["testapp"]
 
+    def test_catalogs_omits_source_url_when_unset(self, catalog_interface: ReadOnlyCatalogInterface) -> None:
+        """catalogs() yields source_url=None when the Catalog doesn't set one."""
+        infos = catalog_interface.catalogs()
+        assert infos[0].source_url is None
+
+    def test_catalogs_serializes_source_url(self, users_table: Table) -> None:
+        """catalogs() surfaces the Catalog.source_url into CatalogInfo."""
+
+        class SourcedCatalog(ReadOnlyCatalogInterface):
+            catalog = Catalog(
+                name="sourced",
+                default_schema="main",
+                schemas=[Schema(name="main", tables=[users_table])],
+                source_url="https://github.com/example/sourced",
+            )
+
+        infos = SourcedCatalog().catalogs()
+        assert infos[0].source_url == "https://github.com/example/sourced"
+
     def test_catalog_attach(self, catalog_interface: ReadOnlyCatalogInterface) -> None:
         """catalog_attach returns result with correct defaults."""
         result = catalog_interface.catalog_attach(

{vgi_python-0.8.3 → vgi_python-0.8.5}/tests/catalog/test_example_worker_catalog.py

@@ -78,6 +78,13 @@ class TestExampleWorkerCatalog:
         catalogs = client.catalogs()
         assert "example" in [c.name for c in catalogs]
 
+    def test_catalogs_advertises_source_url(self) -> None:
+        """The declarative catalog's source_url round-trips through discovery."""
+        client = Client(EXAMPLE_WORKER)
+        catalogs = client.catalogs()
+        example = next(c for c in catalogs if c.name == "example")
+        assert example.source_url == "https://github.com/query-farm/vgi-python"
+
     def test_catalog_attach_works(self) -> None:
         """Can attach to the 'example' catalog."""
         client = Client(EXAMPLE_WORKER)

{vgi_python-0.8.3 → vgi_python-0.8.5}/tests/test_argument_spec.py

@@ -12,6 +12,7 @@ import pytest
 from vgi.argument_spec import (
     VGI_ARG_KEY,
     VGI_ARG_NAMED,
+    VGI_DOC_KEY,
     VGI_TYPE_ANY,
     VGI_TYPE_KEY,
     VGI_TYPE_TABLE,
@@ -674,3 +675,50 @@ class TestArgumentSpecRepr:
         )
         result = repr(spec)
         assert "any_type" in result
+
+
+class TestArgumentDoc:
+    """Per-argument description via the vgi_doc field-metadata key."""
+
+    def test_doc_serialized_as_utf8_field_metadata(self) -> None:
+        """A non-empty doc is emitted as the vgi_doc field metadata (UTF-8)."""
+        specs = [ArgumentSpec(name="unit", position=0, arrow_type=pa.utf8(), doc="Source unit, e.g. 'mi'")]
+        schema = argument_specs_to_schema(specs)
+        meta = schema.field("unit").metadata or {}
+        assert meta.get(VGI_DOC_KEY) == b"Source unit, e.g. 'mi'"
+
+    def test_doc_absent_when_empty(self) -> None:
+        """Presence-only: an empty doc emits no vgi_doc key (absent = undocumented)."""
+        specs = [ArgumentSpec(name="x", position=0, arrow_type=pa.int64())]
+        schema = argument_specs_to_schema(specs)
+        meta = schema.field("x").metadata or {}
+        assert VGI_DOC_KEY not in meta
+
+    def test_doc_round_trips(self) -> None:
+        """Doc survives schema -> specs round-trip; missing doc decodes to ''."""
+        specs = [
+            ArgumentSpec(name="a", position=0, arrow_type=pa.int64(), doc="first arg"),
+            ArgumentSpec(name="b", position=1, arrow_type=pa.utf8()),
+        ]
+        out = schema_to_argument_specs(argument_specs_to_schema(specs))
+        by_name = {s.name: s for s in out}
+        assert by_name["a"].doc == "first arg"
+        assert by_name["b"].doc == ""
+
+    def test_doc_unicode_round_trips(self) -> None:
+        """Non-ASCII docs (µ, ≥, em-dash) survive the UTF-8 round-trip."""
+        doc = "value in µm — must be ≥ 1"
+        specs = [ArgumentSpec(name="v", position=0, arrow_type=pa.float64(), doc=doc)]
+        out = schema_to_argument_specs(argument_specs_to_schema(specs))
+        assert out[0].doc == doc
+
+    def test_extract_doc_from_arg(self) -> None:
+        """extract_argument_specs captures the Arg(doc=...) string."""
+
+        class FunctionWithDocs(TableInOutFunction):  # type: ignore[type-arg]
+            unit: str = Arg[str](0, doc="Unit string, e.g. 'mi'")  # type: ignore[assignment]
+            scale: float = Arg[float](1)  # type: ignore[assignment]
+
+        specs = {s.name: s for s in extract_argument_specs(FunctionWithDocs)}
+        assert specs["unit"].doc == "Unit string, e.g. 'mi'"
+        assert specs["scale"].doc == ""

{vgi_python-0.8.3 → vgi_python-0.8.5}/tests/test_metadata.py

@@ -613,3 +613,118 @@ class TestFunctionTypeInference:
 
         meta = resolve_metadata(TestFunc)
         assert meta.function_type == CatalogFunctionType.TABLE
+
+
+class TestFunctionTags:
+    """A worker can attach arbitrary free-form tags to a function's metadata.
+
+    These flow ``Meta.tags`` -> ``ResolvedMetadata.tags`` -> ``FunctionInfo.tags``
+    -> DuckDB's per-function ``tags`` map, which is what the ``vgi-lint-check``
+    metadata linter reads (e.g. VGI307 requires ``vgi.columns_md`` on a
+    dynamic-schema table function).
+    """
+
+    def test_meta_tags_resolved(self) -> None:
+        """Arbitrary Meta.tags land on ResolvedMetadata.tags verbatim."""
+
+        class TaggedFunction(TableInOutFunction):  # type: ignore[type-arg]
+            class Meta:
+                name = "tagged_func"
+                tags = {
+                    "vgi.columns_md": "| col | type |\n| --- | --- |\n| id | BIGINT |",
+                    "custom.team": "data",
+                }
+
+            data: TableInput = Arg[TableInput](0, doc="Input table")  # type: ignore[assignment]
+
+        meta = resolve_metadata(TaggedFunction)
+        assert meta.tags == {
+            "vgi.columns_md": "| col | type |\n| --- | --- |\n| id | BIGINT |",
+            "custom.team": "data",
+        }
+
+    def test_tags_survive_arrow_roundtrip(self) -> None:
+        """Function tags survive the metadata Arrow wire roundtrip (map column)."""
+
+        class TaggedFunction(TableInOutFunction):  # type: ignore[type-arg]
+            class Meta:
+                name = "tagged_func"
+                tags = {"vgi.columns_md": "| col |", "k": "v"}
+
+            data: TableInput = Arg[TableInput](0, doc="Input table")  # type: ignore[assignment]
+
+        restored = arrow_to_functions(functions_to_arrow([TaggedFunction]))[0]
+        assert restored.tags == {"vgi.columns_md": "| col |", "k": "v"}
+
+    def test_tags_and_examples_are_independent(self) -> None:
+        """Worker tags do not clobber the structured examples, and vice versa.
+
+        ``examples`` is its own structured channel (serialized to the
+        ``examples`` struct column / ``FunctionInfo.examples``); it is never
+        folded into the ``tags`` map. A worker that *also* sets a
+        ``vgi.example_queries`` tag (the tag-encoded examples convention the
+        linter understands) keeps both — the tag is preserved untouched and the
+        structured examples remain separate.
+        """
+
+        class BothFunction(TableInOutFunction):  # type: ignore[type-arg]
+            class Meta:
+                name = "both_func"
+                examples = ["SELECT * FROM both_func((SELECT 1))"]
+                tags = {
+                    "vgi.example_queries": '[{"description":"x","sql":"SELECT tag"}]',
+                    "vgi.columns_md": "| col |",
+                }
+
+            data: TableInput = Arg[TableInput](0, doc="Input table")  # type: ignore[assignment]
+
+        meta = resolve_metadata(BothFunction)
+        # Worker-provided tags untouched (example_queries tag not derived, not dropped).
+        assert meta.tags == {
+            "vgi.example_queries": '[{"description":"x","sql":"SELECT tag"}]',
+            "vgi.columns_md": "| col |",
+        }
+        # Structured examples live independently and are not merged into tags.
+        assert [ex.sql for ex in meta.examples] == ["SELECT * FROM both_func((SELECT 1))"]
+
+    def test_tags_flow_into_function_info(self) -> None:
+        """Tags reach ``FunctionInfo.tags`` (catalog -> DuckDB -> linter path).
+
+        Also asserts that the structured ``examples`` and the worker-set
+        ``vgi.example_queries`` tag coexist on ``FunctionInfo`` without either
+        clobbering the other.
+        """
+        from vgi.catalog.catalog_interface import (
+            FunctionInfo,
+            ReadOnlyCatalogInterface,
+        )
+        from vgi.table_function import TableFunctionGenerator
+
+        class DynamicTable(TableFunctionGenerator):  # type: ignore[type-arg]
+            class Meta:
+                name = "dynamic_table"
+                examples = ["SELECT * FROM dynamic_table()"]
+                tags = {
+                    "vgi.columns_md": "| col | type |\n| --- | --- |\n| id | BIGINT |",
+                    "vgi.example_queries": '[{"description":"x","sql":"SELECT 1"}]',
+                    "custom.team": "data",
+                }
+
+            FIXED_SCHEMA = pa.schema([pa.field("id", pa.int64())])
+
+            def process(self, params):  # type: ignore[no-untyped-def]
+                yield pa.record_batch({"id": [1]})
+
+        class Cat(ReadOnlyCatalogInterface):
+            catalog_name = "functions"
+            functions = [DynamicTable]
+
+        info = Cat()._function_to_info(DynamicTable, "main")
+        assert isinstance(info, FunctionInfo)
+        assert info.tags == {
+            "vgi.columns_md": "| col | type |\n| --- | --- |\n| id | BIGINT |",
+            "vgi.example_queries": '[{"description":"x","sql":"SELECT 1"}]',
+            "custom.team": "data",
+        }
+        # Structured examples are carried separately from the tag-encoded ones.
+        assert [ex.sql for ex in info.examples] == ["SELECT * FROM dynamic_table()"]

{vgi_python-0.8.3 → vgi_python-0.8.5}/uv.lock

@@ -2250,7 +2250,7 @@ requires-dist = [
 
 [[package]]
 name = "vgi-python"
-version = "0.8.2"
+version = "0.8.4"
 source = { editable = "." }
 dependencies = [
     { name = "click" },

{vgi_python-0.8.3 → vgi_python-0.8.5}/vgi/_test_fixtures/table/_common.py

@@ -61,7 +61,7 @@ class CountBatchArgs:
     """
 
     count: Annotated[int, Arg(0, doc="Number of rows to generate", ge=0)]
-    batch_size: Annotated[int, Arg("batch_size", default=1000, doc="Batch size for output", ge=1)]
+    batch_size: Annotated[int, Arg("batch_size", default=2048, doc="Batch size for output", ge=1)]
 
 
 @dataclass(slots=True, frozen=True)
@@ -91,7 +91,7 @@ class _BaseSequenceFunction(TableFunctionGenerator[Any, CountdownState]):
     NUMPY_DTYPE: ClassVar[type[np.generic]] = np.int64
     STATS_ARROW_TYPE: ClassVar[pa.DataType] = pa.int64()
     STATS_COLUMN_NAME: ClassVar[str] = "n"
-    BATCH_SIZE_FALLBACK: ClassVar[int] = 1000
+    BATCH_SIZE_FALLBACK: ClassVar[int] = 2048
 
     @classmethod
     def initial_state(cls, params: ProcessParams[Any]) -> CountdownState:

{vgi_python-0.8.3 → vgi_python-0.8.5}/vgi/_test_fixtures/worker.py

@@ -320,6 +320,7 @@ _EXAMPLE_CATALOG = Catalog(
     default_schema="main",
     comment="Example VGI catalog for testing",
     tags={"source": "vgi-fixture-worker", "version": "1"},
+    source_url="https://github.com/query-farm/vgi-python",
     schemas=[
         Schema(
             name="main",

{vgi_python-0.8.3 → vgi_python-0.8.5}/vgi/argument_spec.py

@@ -38,6 +38,7 @@ __all__ = [
     "VGI_VARARGS_TRUE",
     "VGI_CONST_KEY",
     "VGI_CONST_TRUE",
+    "VGI_DOC_KEY",
 ]
 
 # =============================================================================
@@ -61,6 +62,11 @@ VGI_VARARGS_TRUE = b"true"
 VGI_CONST_KEY = b"vgi_const"
 VGI_CONST_TRUE = b"true"
 
+# Key carrying the per-argument description (UTF-8 text). Presence-only: the key
+# is omitted entirely when there is no doc (absent = undocumented). The
+# ``vgi_doc_*`` prefix is reserved for future per-argument doc variants.
+VGI_DOC_KEY = b"vgi_doc"
+
 
 def _argument_spec_sort_key(spec: "ArgumentSpec") -> tuple[int, int | str]:
     """Sort key: positional first (by index), then named (alphabetically)."""
@@ -98,6 +104,9 @@ class ArgumentSpec:
         is_const: True if this argument is constant-folded ([`ConstParam`][]).
             Constant arguments are scalar values known at planning time,
             rather than columnar data processed at runtime.
+        doc: Optional human/agent-facing description of the argument. Surfaced
+            through the catalog as the ``vgi_doc`` Arrow field metadata key
+            (UTF-8); empty string means undocumented.
 
     Note:
         For named arguments, the Python attribute name (``name``) and the SQL
@@ -119,6 +128,7 @@ class ArgumentSpec:
     is_any_type: bool = False
     is_varargs: bool = False
     is_const: bool = False
+    doc: str = ""
 
     def __repr__(self) -> str:
         """Return concise repr showing key attributes."""
@@ -197,6 +207,10 @@ def argument_specs_to_schema(specs: Sequence[ArgumentSpec]) -> pa.Schema:
         if spec.is_const:
             metadata[VGI_CONST_KEY] = VGI_CONST_TRUE
 
+        # Per-argument description (UTF-8; presence-only — omit when empty)
+        if spec.doc:
+            metadata[VGI_DOC_KEY] = spec.doc.encode("utf-8")
+
         # Create field with or without metadata
         field = pa.field(
             spec.name,
@@ -246,6 +260,10 @@ def schema_to_argument_specs(schema: pa.Schema) -> list[ArgumentSpec]:
         # Check const
         is_const = metadata.get(VGI_CONST_KEY) == VGI_CONST_TRUE
 
+        # Per-argument description (UTF-8; absent = undocumented)
+        doc_bytes = metadata.get(VGI_DOC_KEY)
+        doc = doc_bytes.decode("utf-8") if doc_bytes else ""
+
         specs.append(
             ArgumentSpec(
                 name=field.name,
@@ -255,6 +273,7 @@ def schema_to_argument_specs(schema: pa.Schema) -> list[ArgumentSpec]:
                 is_any_type=is_any_type,
                 is_varargs=is_varargs,
                 is_const=is_const,
+                doc=doc,
             )
         )
 
@@ -312,6 +331,7 @@ def extract_argument_specs(
                 is_any_type=param_arg.is_any,
                 is_varargs=param_arg.varargs,
                 is_const=False,
+                doc=param_arg.doc or "",
             )
         )
 
@@ -328,6 +348,7 @@ def extract_argument_specs(
                 is_any_type=const_arg.is_any,
                 is_varargs=const_arg.varargs,
                 is_const=True,
+                doc=const_arg.doc or "",
             )
         )
 
@@ -392,6 +413,7 @@ def extract_argument_specs(
                     is_any_type=is_any_type,
                     is_varargs=arg_instance.varargs,
                     is_const=getattr(arg_instance, "const", False),
+                    doc=getattr(arg_instance, "doc", "") or "",
                 )
             )
 
@@ -466,6 +488,7 @@ def extract_argument_specs(
                         is_any_type=is_any_type,
                         is_varargs=is_varargs,
                         is_const=is_const,
+                        doc=getattr(arg_legacy, "doc", "") or "",
                     )
                 )
 

{vgi_python-0.8.3 → vgi_python-0.8.5}/vgi/catalog/catalog_interface.py

@@ -2319,6 +2319,7 @@ class ReadOnlyCatalogInterface(CatalogInterface):
                 implementation_version=None,
                 data_version_spec=None,
                 attach_option_specs=[spec.serialize() for spec in self.attach_option_specs],
+                source_url=self.catalog.source_url if self.catalog is not None else None,
             )
         ]
 

{vgi_python-0.8.3 → vgi_python-0.8.5}/vgi/catalog/descriptors.py

@@ -873,6 +873,10 @@ class Catalog:
         schemas: Sequence of Schema objects defining the catalog contents.
         comment: Optional comment describing the catalog.
         tags: Optional key-value tags associated with the catalog.
+        source_url: Where this worker's code lives — repo, build, or docs
+            homepage. ``None`` (the default) when the worker doesn't advertise
+            a source location. Surfaced via the ``catalog_catalogs()`` discovery
+            record (``CatalogInfo.source_url``).
 
     """
 
@@ -881,6 +885,7 @@ class Catalog:
     schemas: Sequence[Schema] = ()
     comment: str | None = None
     tags: dict[str, str] = field(default_factory=dict)
+    source_url: str | None = None
 
     def __post_init__(self) -> None:
         """Validate catalog configuration."""