corvic-engine 0.3.0rc54__tar.gz → 0.3.0rc56__tar.gz

{corvic_engine-0.3.0rc54 → corvic_engine-0.3.0rc56}/Cargo.lock

@@ -22,7 +22,7 @@ checksum = "baf1de4339761588bc0619e3cbc0120ee582ebb74b53b4efbf79117bd2da40fd"
 
 [[package]]
 name = "corvic-engine"
-version = "0.3.0-rc.54"
+version = "0.3.0-rc.56"
 dependencies = [
  "numpy",
  "pyo3",
@@ -63,7 +63,7 @@ checksum = "48c757948c5ede0e46177b7add2e67155f70e33c07fea8284df6576da70b3719"
 
 [[package]]
 name = "engine-walk"
-version = "0.3.0-rc.54"
+version = "0.3.0-rc.56"
 dependencies = [
  "version_check",
 ]

{corvic_engine-0.3.0rc54 → corvic_engine-0.3.0rc56}/Cargo.toml

@@ -3,7 +3,7 @@ resolver = "2"
 members = ["python"]
 
 [workspace.package]
-version = "0.3.0-rc.54"
+version = "0.3.0-rc.56"
 edition = "2021"
 publish = false
 

{corvic_engine-0.3.0rc54 → corvic_engine-0.3.0rc56}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: corvic-engine
-Version: 0.3.0rc54
+Version: 0.3.0rc56
 Classifier: Environment :: Console
 Classifier: License :: Other/Proprietary License
 Classifier: Programming Language :: Python :: Implementation :: CPython

{corvic_engine-0.3.0rc54 → corvic_engine-0.3.0rc56}/pyproject.toml

@@ -20,7 +20,7 @@ authors = [
 ]
 readme = "README.md"
 description = "Seamless embedding generation and retrieval."
-version = "0.3.0-rc.54" # sync this with tool.poetry.version below
+version = "0.3.0-rc.56" # sync this with tool.poetry.version below
 
 # TODO(ddn): Pull dependencies from tool.poetry.dependencies. We use maturin as
 # the build system but poetry for the development environment.
@@ -74,7 +74,7 @@ packages = [
 ]
 description = ""
 authors = []
-version = "0.3.0-rc.54"
+version = "0.3.0-rc.56"
 
 [tool.poetry.scripts]
 check = "corvic_check.cli:main"
@@ -95,7 +95,7 @@ cachetools = "^5.3.0"
 duckdb = "^1.0.0"
 more-itertools = "^10.2.0"
 numpy = "^1.26.3"
-polars = "^1.9.0"
+polars = "^1.25.2"
 # buf (which we use to generate protobuf code) doesn't yet have a plugin
 # that supports patch versions greater than 3.
 protobuf = ">=4.25.0,<4.25.4"
@@ -335,6 +335,7 @@ known-third-party = ["proto", "rockset"]
 "pyarrow.parquet" = "pq" # this alias is used in pyarrow examples
 "sqlalchemy" = "sa" # common way to refer to sqlalchemy
 "sqlalchemy.orm" = "sa_orm" # shortening to disambiguate from corvic.orm
+"polars.selectors" = "cs" # this alias is used in polars examples
 
 [tool.pyright]
 include = ["python", "tests"]

{corvic_engine-0.3.0rc54 → corvic_engine-0.3.0rc56}/python/corvic/embed/node2vec.py

@@ -176,7 +176,7 @@ class Space:
                     left_on=start_id_column_names,
                     right_on=self._node_ids,
                     how="left",
-                    join_nulls=True,
+                    nulls_equal=True,
                 )
                 .select(["index"])
                 .rename({"index": "start"})
@@ -187,7 +187,7 @@ class Space:
                     left_on=end_id_column_names,
                     right_on=self._node_ids,
                     how="left",
-                    join_nulls=True,
+                    nulls_equal=True,
                 )
                 .select(["index"])
                 .rename({"index": "end"})

{corvic_engine-0.3.0rc54 → corvic_engine-0.3.0rc56}/python/corvic/model/__init__.py

@@ -2,7 +2,13 @@
 
 import corvic.model._feature_type as feature_type
 from corvic.model._agent import Agent, AgentID
-from corvic.model._base_model import BaseModel, HasProtoSelf, UsesOrmID
+from corvic.model._base_model import (
+    BaseModel,
+    BelongsToOrgModel,
+    BelongsToRoomModel,
+    HasProtoSelf,
+    UsesOrmID,
+)
 from corvic.model._completion_model import (
     CompletionModel,
     CompletionModelID,
@@ -25,6 +31,14 @@ from corvic.model._pipeline import (
 )
 from corvic.model._proto_orm_convert import (
     UNCOMMITTED_ID_PREFIX,
+    OrmBelongsToOrgObj,
+    OrmBelongsToRoomObj,
+    OrmObj,
+    ProtoBelongsToOrgObj,
+    ProtoBelongsToRoomObj,
+    ProtoObj,
+    add_orm_org_mixin_to_session,
+    add_orm_room_mixin_to_session,
     space_orm_to_proto,
     timestamp_orm_to_proto,
 )
@@ -60,6 +74,8 @@ __all__ = [
     "Agent",
     "AgentID",
     "BaseModel",
+    "BelongsToOrgModel",
+    "BelongsToRoomModel",
     "ChunkPdfsPipeline",
     "Column",
     "CompletionModel",
@@ -76,8 +92,14 @@ __all__ = [
     "ImageSpace",
     "Node2VecParameters",
     "OcrPdfsPipeline",
+    "OrmObj",
+    "OrmBelongsToOrgObj",
+    "OrmBelongsToRoomObj",
     "Pipeline",
     "PipelineID",
+    "ProtoObj",
+    "ProtoBelongsToOrgObj",
+    "ProtoBelongsToRoomObj",
     "RelationalSpace",
     "Resource",
     "ResourceID",
@@ -96,6 +118,8 @@ __all__ = [
     "UnknownTransformationPipeline",
     "UnknownSpace",
     "UsesOrmID",
+    "add_orm_org_mixin_to_session",
+    "add_orm_room_mixin_to_session",
     "embedding_model_proto_to_name",
     "feature_type",
     "image_model_proto_to_name",

{corvic_engine-0.3.0rc54 → corvic_engine-0.3.0rc56}/python/corvic/model/_agent.py

@@ -10,7 +10,7 @@ from typing import TypeAlias
 from sqlalchemy import orm as sa_orm
 
 from corvic import orm, system
-from corvic.model._base_model import BaseModel
+from corvic.model._base_model import BelongsToRoomModel
 from corvic.model._defaults import Defaults
 from corvic.model._proto_orm_convert import (
     agent_delete_orms,
@@ -27,7 +27,7 @@ FeatureViewID: TypeAlias = orm.FeatureViewID
 AgentID: TypeAlias = orm.AgentID
 
 
-class Agent(BaseModel[AgentID, models_pb2.Agent, orm.Agent]):
+class Agent(BelongsToRoomModel[AgentID, models_pb2.Agent, orm.Agent]):
     """A corvic agent represents a named agent that can produce embeddings."""
 
     @classmethod
@@ -133,10 +133,6 @@ class Agent(BaseModel[AgentID, models_pb2.Agent, orm.Agent]):
     def name(self) -> str:
         return self.proto_self.name
 
-    @property
-    def room_id(self) -> RoomID:
-        return RoomID(self.proto_self.room_id)
-
     @property
     def parameters(self) -> AgentParameters:
         return self.proto_self.agent_parameters

{corvic_engine-0.3.0rc54 → corvic_engine-0.3.0rc56}/python/corvic/model/_base_model.py

@@ -5,29 +5,29 @@ import datetime
 import functools
 import uuid
 from collections.abc import Callable, Iterable, Iterator, Sequence
-from typing import Final, Generic, Protocol, TypeVar
+from typing import Final, Generic
 
 import sqlalchemy as sa
 import sqlalchemy.orm as sa_orm
 import structlog
-from google.protobuf import timestamp_pb2
 from typing_extensions import Self
 
 from corvic import orm, system
-from corvic.model._proto_orm_convert import ID, UNCOMMITTED_ID_PREFIX, OrmObj
+from corvic.model._proto_orm_convert import (
+    ID,
+    UNCOMMITTED_ID_PREFIX,
+    OrmBelongsToOrgObj,
+    OrmBelongsToRoomObj,
+    OrmObj,
+    ProtoBelongsToOrgObj,
+    ProtoBelongsToRoomObj,
+    ProtoObj,
+)
 from corvic.result import InvalidArgumentError, NotFoundError, Ok
 
 _logger = structlog.get_logger()
 
 
-class _ModelProto(Protocol):
-    id: str
-    created_at: timestamp_pb2.Timestamp
-
-
-_ProtoObj = TypeVar("_ProtoObj", bound=_ModelProto)
-
-
 def _generate_uncommitted_id_str():
     return f"{UNCOMMITTED_ID_PREFIX}{uuid.uuid4()}"
 
@@ -43,11 +43,11 @@ def _create_or_join_session(
             yield session
 
 
-class HasProtoSelf(Generic[_ProtoObj], abc.ABC):
+class HasProtoSelf(Generic[ProtoObj], abc.ABC):
     client: Final[system.Client]
-    proto_self: Final[_ProtoObj]
+    proto_self: Final[ProtoObj]
 
-    def __init__(self, client: system.Client, proto_self: _ProtoObj):
+    def __init__(self, client: system.Client, proto_self: ProtoObj):
         self.proto_self = proto_self
         self.client = client
 
@@ -58,8 +58,8 @@ class HasProtoSelf(Generic[_ProtoObj], abc.ABC):
         return None
 
 
-class UsesOrmID(Generic[ID, _ProtoObj], HasProtoSelf[_ProtoObj]):
-    def __init__(self, client: system.Client, proto_self: _ProtoObj):
+class UsesOrmID(Generic[ID, ProtoObj], HasProtoSelf[ProtoObj]):
+    def __init__(self, client: system.Client, proto_self: ProtoObj):
         if not proto_self.id:
             proto_self.id = _generate_uncommitted_id_str()
         super().__init__(client, proto_self)
@@ -73,7 +73,7 @@ class UsesOrmID(Generic[ID, _ProtoObj], HasProtoSelf[_ProtoObj]):
         return self.id_class().from_str(self.proto_self.id)
 
 
-class BaseModel(Generic[ID, _ProtoObj, OrmObj], UsesOrmID[ID, _ProtoObj]):
+class BaseModel(Generic[ID, ProtoObj, OrmObj], UsesOrmID[ID, ProtoObj]):
     """Base for orm wrappers providing a unified update mechanism."""
 
     @classmethod
@@ -82,12 +82,12 @@ class BaseModel(Generic[ID, _ProtoObj, OrmObj], UsesOrmID[ID, _ProtoObj]):
 
     @classmethod
     @abc.abstractmethod
-    def orm_to_proto(cls, orm_obj: OrmObj) -> _ProtoObj: ...
+    def orm_to_proto(cls, orm_obj: OrmObj) -> ProtoObj: ...
 
     @classmethod
     @abc.abstractmethod
     def proto_to_orm(
-        cls, proto_obj: _ProtoObj, session: orm.Session
+        cls, proto_obj: ProtoObj, session: orm.Session
     ) -> Ok[OrmObj] | InvalidArgumentError: ...
 
     @classmethod
@@ -102,7 +102,7 @@ class BaseModel(Generic[ID, _ProtoObj, OrmObj], UsesOrmID[ID, _ProtoObj]):
         obj_id: ID,
         client: system.Client,
         existing_session: sa_orm.Session | None = None,
-    ) -> Ok[_ProtoObj] | NotFoundError:
+    ) -> Ok[ProtoObj] | NotFoundError:
         """Create a model object by loading it from the database."""
         with _create_or_join_session(client, existing_session) as session:
             orm_self = session.get(cls.orm_class(), obj_id)
@@ -146,7 +146,7 @@ class BaseModel(Generic[ID, _ProtoObj, OrmObj], UsesOrmID[ID, _ProtoObj]):
         ]
         | None = None,
         existing_session: sa_orm.Session | None = None,
-    ) -> Ok[list[_ProtoObj]] | NotFoundError | InvalidArgumentError:
+    ) -> Ok[list[ProtoObj]] | NotFoundError | InvalidArgumentError:
         """List sources that exist in storage."""
         orm_class = cls.orm_class()
         with _create_or_join_session(client, existing_session) as session:
@@ -231,3 +231,25 @@ class BaseModel(Generic[ID, _ProtoObj, OrmObj], UsesOrmID[ID, _ProtoObj]):
                     proto_self=new_proto_self,
                 )
             )
+
+
+class BelongsToOrgModel(
+    Generic[ID, ProtoBelongsToOrgObj, OrmBelongsToOrgObj],
+    BaseModel[ID, ProtoBelongsToOrgObj, OrmBelongsToOrgObj],
+):
+    """Base for orm wrappers with org mixin providing a unified update mechanism."""
+
+    @property
+    def org_id(self) -> orm.OrgID:
+        return orm.OrgID().from_str(self.proto_self.org_id)
+
+
+class BelongsToRoomModel(
+    Generic[ID, ProtoBelongsToRoomObj, OrmBelongsToRoomObj],
+    BelongsToOrgModel[ID, ProtoBelongsToRoomObj, OrmBelongsToRoomObj],
+):
+    """Base for orm wrappers with room mixin providing a unified update mechanism."""
+
+    @property
+    def room_id(self) -> orm.RoomID:
+        return orm.RoomID().from_str(self.proto_self.room_id)

{corvic_engine-0.3.0rc54 → corvic_engine-0.3.0rc56}/python/corvic/model/_completion_model.py

@@ -7,10 +7,11 @@ import datetime
 from collections.abc import Iterable, Sequence
 from typing import Literal, TypeAlias
 
+from google.protobuf import timestamp_pb2
 from sqlalchemy import orm as sa_orm
 
 from corvic import orm, system
-from corvic.model._base_model import BaseModel
+from corvic.model._base_model import BelongsToOrgModel
 from corvic.model._defaults import Defaults
 from corvic.model._proto_orm_convert import (
     completion_model_delete_orms,
@@ -24,9 +25,13 @@ from corvic_generated.orm.v1 import completion_model_pb2
 CompletionModelID: TypeAlias = orm.CompletionModelID
 OrgID: TypeAlias = orm.OrgID
 
+UNIX_TIMESTAMP_START_DATETIME = timestamp_pb2.Timestamp(seconds=0, nanos=0)
+
 
 class CompletionModel(
-    BaseModel[CompletionModelID, models_pb2.CompletionModel, orm.CompletionModel]
+    BelongsToOrgModel[
+        CompletionModelID, models_pb2.CompletionModel, orm.CompletionModel
+    ]
 ):
     """Completion Models."""
 
@@ -58,10 +63,6 @@ class CompletionModel(
     def name(self) -> str:
         return self.proto_self.name
 
-    @property
-    def org_id(self) -> OrgID:
-        return OrgID(self.proto_self.org_id)
-
     @property
     def provider(self) -> Literal["openai-generic", "azure-openai"] | None:
         match self.proto_self.parameters.WhichOneof("params"):
@@ -112,6 +113,14 @@ class CompletionModel(
     def description(self) -> str:
         return self.proto_self.description
 
+    @property
+    def last_validation_time(self) -> datetime.datetime | None:
+        if self.proto_self.last_validation_time != UNIX_TIMESTAMP_START_DATETIME:
+            return self.proto_self.last_validation_time.ToDatetime(
+                tzinfo=datetime.timezone.utc
+            )
+        return None
+
     @classmethod
     def create(
         cls,
@@ -197,3 +206,10 @@ class CompletionModel(
         proto_self = copy.deepcopy(self.proto_self)
         proto_self.secret_api_key = secret_api_key
         return CompletionModel(self.client, proto_self)
+
+    def with_last_validation_time(
+        self, last_validation_time: datetime.datetime
+    ) -> CompletionModel:
+        proto_self = copy.deepcopy(self.proto_self)
+        proto_self.last_validation_time.FromDatetime(last_validation_time)
+        return CompletionModel(self.client, proto_self)

{corvic_engine-0.3.0rc54 → corvic_engine-0.3.0rc56}/python/corvic/model/_feature_view.py

@@ -18,7 +18,7 @@ from sqlalchemy.orm.interfaces import LoaderOption
 from typing_extensions import Self
 
 from corvic import op_graph, orm, system
-from corvic.model._base_model import BaseModel, UsesOrmID
+from corvic.model._base_model import BelongsToRoomModel, UsesOrmID
 from corvic.model._defaults import Defaults
 from corvic.model._proto_orm_convert import (
     feature_view_delete_orms,
@@ -642,7 +642,9 @@ class DeepGnnCsvUrlMetadata(DataclassAsTypedMetadataMixin):
     output_csv_url: str
 
 
-class FeatureView(BaseModel[FeatureViewID, models_pb2.FeatureView, orm.FeatureView]):
+class FeatureView(
+    BelongsToRoomModel[FeatureViewID, models_pb2.FeatureView, orm.FeatureView]
+):
     """FeatureViews describe how Sources should be modeled to create a feature space.
 
     Example:
@@ -738,10 +740,6 @@ class FeatureView(BaseModel[FeatureViewID, models_pb2.FeatureView, orm.FeatureVi
                 pass
         return Ok([cls.from_proto(proto, client) for proto in protos])
 
-    @property
-    def room_id(self):
-        return RoomID(self.proto_self.room_id)
-
     @property
     def source_ids(self):
         return [fv_source.source.id for fv_source in self.feature_view_sources]

{corvic_engine-0.3.0rc54 → corvic_engine-0.3.0rc56}/python/corvic/model/_pipeline.py

@@ -15,7 +15,7 @@ from typing_extensions import Self
 
 import corvic.table
 from corvic import op_graph, orm, system
-from corvic.model._base_model import BaseModel
+from corvic.model._base_model import BelongsToRoomModel
 from corvic.model._defaults import Defaults
 from corvic.model._proto_orm_convert import (
     pipeline_delete_orms,
@@ -31,7 +31,7 @@ PipelineID: TypeAlias = orm.PipelineID
 RoomID: TypeAlias = orm.RoomID
 
 
-class Pipeline(BaseModel[PipelineID, models_pb2.Pipeline, orm.Pipeline]):
+class Pipeline(BelongsToRoomModel[PipelineID, models_pb2.Pipeline, orm.Pipeline]):
     """Pipelines map resources to sources."""
 
     @classmethod
@@ -152,10 +152,6 @@ class Pipeline(BaseModel[PipelineID, models_pb2.Pipeline, orm.Pipeline]):
             case Ok(protos):
                 return Ok([cls.from_proto(proto, client) for proto in protos])
 
-    @property
-    def room_id(self):
-        return RoomID(self.proto_self.room_id)
-
     @property
     def name(self):
         return self.proto_self.name