rust/lancedb
1
0
Fork 0
mirror of https://github.com/lancedb/lancedb.git synced 2026-10-14 15:52:26 +00:00
Code Issues Packages Projects Releases Wiki Activity

feat: represent registered Functions by OCI image identity (#4176)

Browse Source 
Function versions identify independent Function objects and their
numeric revisions. Rust and Python expose the object ID, location,
canonical decimal version, metadata, and availability separately from
the OCI image digest. Computed-column applications carry the complete
object reference, so existing bindings retain their identity after a
name is removed and reused.

Source authoring keeps the existing
create_function/create_function_async, job wait, and column-binding
APIs. The server coordinates baking followed by registration; users do
not have to manage manifest digests to create a Function. The previous
stored Function representation is intentionally unsupported. Existing
contract tests and shared wire fixtures are migrated to the new model.

This SDK change accompanies the final integration layer
https://github.com/lancedb/sophon/pull/7887 in the Sophon stack:
https://github.com/lancedb/sophon/pull/7885 →
https://github.com/lancedb/sophon/pull/7886 →
https://github.com/lancedb/sophon/pull/7887. A metadata-only revision
can retain the same executable image; Function version numbers must not
be used as image cache keys.

---------

Co-authored-by: lancedb automation <robot@lancedb.com>
Co-authored-by: Yang Cen <bubble-cal@outlook.com>
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
This commit is contained in:
Xuanwo
2026-09-16 12:38:59 +08:00
committed by GitHub
co-authored by lancedb automation Yang Cen Claude Fable 5.1
parent f7579d2aae
commit 3b37ea2c7a
27 changed files with 314 additions and 165 deletions
Download Patch File Download Diff File Expand all files Collapse all files
python/python/tests/test_first_class_function_slice1.py
+27 -12
View File
@@ -93,16 +93,22 @@ def test_function_version_identity_is_immutable_and_exact():
value = job_result("remote_function_job.json")
version = FunctionVersion.from_json(json.dumps(value))
assert version.name == "embed"
assert version.version == "fv_01K3EXACT"
assert version.version == "1"
assert version.image.manifest_digest.startswith("sha256:")
assert version.version != version.image.manifest_digest
with pytest.raises((TypeError, ValueError)):
version.version = "fv_changed"
version.version = "1"
with pytest.raises(TypeError, match="immutable"):
version.runtime.env["TOKENIZERS_PARALLELISM"] = "true"
version.image.descriptor["format_version"] = "changed"
changed = dict(value)
changed["version"] = "fv_changed"
changed["version"] = "2"
assert FunctionVersion(**changed) != version
assert FunctionVersion(**changed).image == version.image
for invalid in [version.image.manifest_digest, "0", "01", "-1", str(2**64)]:
with pytest.raises(ValueError):
FunctionVersion(**{**value, "version": invalid})
def test_function_version_binds_named_columns_as_one_immutable_application():
@@ -137,7 +143,7 @@ def test_function_version_binding_validates_names_and_direct_columns():
def test_function_version_keeps_named_struct_outputs_in_one_application():
value = job_result("remote_function_job.json")
value["name"] = "text_features"
value["version"] = "fv_multi_output"
value["version"] = "1"
value["signature"] = {
"inputs": [
{"name": "title", "arrow_type": "utf8", "nullable": True},
@@ -182,14 +188,14 @@ def test_function_version_keeps_named_struct_outputs_in_one_application():
def test_unknown_fields_and_discriminators_are_forward_decodable():
value = job_result("remote_function_job.json")
value["future_version_metadata"] = {"retention_class": "catalog"}
value["runtime"] = {"kind": "wasm", "module_digest": "sha256:wasm"}
value["image"]["descriptor"]["future_interface"] = {"kind": "wasm"}
value["signature"]["output"]["kind"] = "future_output_shape"
version = FunctionVersion.from_json(json.dumps(value))
assert version.runtime.kind == "wasm"
assert version.runtime.python_version is None
assert version.runtime.environment is None
assert json.loads(version.to_canonical_json())["runtime"] == {"kind": "wasm"}
assert version.image.descriptor["future_interface"] == {"kind": "wasm"}
assert json.loads(version.to_canonical_json())["image"]["descriptor"][
"future_interface"
] == {"kind": "wasm"}
assert version.signature.output.kind == "future_output_shape"
@@ -222,7 +228,7 @@ def test_function_application_uses_rename_columns_only():
def test_binding_and_refresh_result_keep_stable_remote_fields():
binding = FunctionBinding.from_json(fixture("remote_function_binding.json"))
assert binding.function.version == "fv_01K3TEXT"
assert binding.function.version == "1"
assert [output.output_ordinal for output in binding.outputs] == [0, 1]
assert binding.input_schema is not None
assert binding.output_schema is not None
@@ -339,7 +345,16 @@ def test_rename_requires_named_struct_and_keeps_partial_mapping_immutable():
scalar = FunctionApplication.from_json(
json.dumps(
{
"function": {"name": "embed", "version": "fv_exact"},
"function": {
"name": "embed",
"version": "1",
"object_id": "fixture",
"location": "memory:///fixture",
"manifest_digest": (
"sha256:"
"7e22f815b6648e14f093a3979a8e5a2082fa773ebe1ec84b135cae7e84d6f8e6"
),
},
"inputs": [],
"output": {
"kind": "scalar",
python/python/tests/test_first_class_function_slice2.py
+28 -19
View File
@@ -45,6 +45,11 @@ FIXTURES = (
)
FUNCTION_VERSION = json.loads(
(FIXTURES / "remote_function_version.canonical.json").read_text()
)["version"]
@udf(
pip=["numpy>=2"],
env={"MODE": "test"},
@@ -1199,11 +1204,11 @@ def test_local_function_catalog_operations_are_not_supported(tmp_path):
with pytest.raises(NotImplementedError, match=message):
db.create_function_async(normalize_score)
with pytest.raises(NotImplementedError, match=message):
db.get_function("normalize_score", version="fv_exact")
db.get_function("normalize_score", version=FUNCTION_VERSION)
with pytest.raises(NotImplementedError, match=message):
db.list_functions()
with pytest.raises(NotImplementedError, match=message):
db.drop_function("normalize_score", version="fv_exact")
db.drop_function("normalize_score", version=FUNCTION_VERSION)
@contextlib.contextmanager
@@ -1230,15 +1235,17 @@ def _mock_remote_function_catalog():
if self.path == "/v1/function/normalize_score/create":
state["version"] = {
"name": "normalize_score",
"version": "fv_exact",
"artifact": {
key: body["artifact"][key]
for key in ("kind", "digest", "entrypoint")
},
"version": FUNCTION_VERSION,
"object_id": "fixture",
"location": "memory:///fixture",
"metadata": {},
"disabled": False,
"image": json.loads(
(
FIXTURES / "remote_function_version.canonical.json"
).read_text()
)["image"],
"signature": body["signature"],
"runtime": body["runtime"],
"runtime_digest": "sha256:runtime",
"environment_digest": "sha256:environment",
"created_at": "2026-08-21T00:00:00Z",
}
response = {"job_id": "job-register"}
@@ -1252,10 +1259,10 @@ def _mock_remote_function_catalog():
"result": state["version"],
}
elif self.path == "/v1/function/normalize_score/describe":
assert body == {"version": "fv_exact"}
assert body == {"version": FUNCTION_VERSION}
response = state["version"]
elif self.path == "/v1/function/normalize_score/drop":
assert body == {"version": "fv_exact"}
assert body == {"version": FUNCTION_VERSION}
response = {"dropped": True}
else:
status = 404
@@ -1278,7 +1285,7 @@ def _mock_remote_function_catalog():
"functions": [
{
"name": "normalize_score",
"version": "fv_exact",
"version": FUNCTION_VERSION,
"definition": state["version"],
}
],
@@ -1314,7 +1321,7 @@ def test_remote_registration_job_and_exact_version_reopen_round_trip():
assert created == reopened
assert reopened.name == "normalize_score"
assert reopened.version == "fv_exact"
assert reopened.version == FUNCTION_VERSION
create_request = state["requests"][0][1]
expected_request = json.loads(
normalize_score.registration_request.to_canonical_json()
@@ -1334,7 +1341,7 @@ def test_blocking_remote_registration_returns_function_version():
created = db.create_function(normalize_score)
assert created.name == "normalize_score"
assert created.version == "fv_exact"
assert created.version == FUNCTION_VERSION
assert [path for path, _ in state["requests"]] == [
"/v1/function/normalize_score/create",
"/v1/jobs/describe",
@@ -1392,12 +1399,12 @@ def test_remote_drop_function_sends_exact_version():
host_override=host,
client_config={"retry_config": {"retries": 0}},
)
assert db.drop_function("normalize_score", version="fv_exact") is True
assert db.drop_function("normalize_score", version=FUNCTION_VERSION) is True
assert state["requests"] == [
(
"/v1/function/normalize_score/drop",
{"version": "fv_exact"},
{"version": FUNCTION_VERSION},
)
]
@@ -1411,11 +1418,13 @@ async def test_async_remote_drop_function_sends_exact_version():
host_override=host,
client_config={"retry_config": {"retries": 0}},
)
assert await db.drop_function("normalize_score", version="fv_exact") is True
assert (
await db.drop_function("normalize_score", version=FUNCTION_VERSION) is True
)
assert state["requests"] == [
(
"/v1/function/normalize_score/drop",
{"version": "fv_exact"},
{"version": FUNCTION_VERSION},
)
]