corvic-engine 0.3.0rc83__cp38-abi3-win_amd64.whl → 0.3.0rc85__cp38-abi3-win_amd64.whl

corvic/system/_embedder.py

@@ -39,10 +39,6 @@ class EmbedTextResult:
 class TextEmbedder(Protocol):
     """Use a model to embed text."""
 
-    def embed(
-        self, context: EmbedTextContext
-    ) -> Ok[EmbedTextResult] | InvalidArgumentError | InternalError: ...
-
     async def aembed(
         self,
         context: EmbedTextContext,
@@ -74,10 +70,6 @@ class ImageEmbedder(Protocol):
     @classmethod
     def model_name(cls) -> str: ...
 
-    def embed(
-        self, context: EmbedImageContext
-    ) -> Ok[EmbedImageResult] | InvalidArgumentError | InternalError: ...
-
     async def aembed(
         self,
         context: EmbedImageContext,

corvic/system/_image_embedder.py

@@ -3,7 +3,7 @@ import asyncio
 import dataclasses
 from concurrent.futures import ThreadPoolExecutor
 from io import BytesIO
-from typing import TYPE_CHECKING, Any, cast
+from typing import TYPE_CHECKING, Any, Protocol, cast
 
 import numpy as np
 import polars as pl
@@ -17,8 +17,7 @@ from corvic.system._embedder import (
 
 if TYPE_CHECKING:
     from PIL import Image
-    from transformers.models.auto.modeling_auto import AutoModel
-    from transformers.models.auto.processing_auto import AutoProcessor
+    from torch import FloatTensor
 
 
 class RandomImageEmbedder(ImageEmbedder):
@@ -77,10 +76,23 @@ def image_from_bytes(
         return InvalidArgumentError("invalid image format")
 
 
+class TransformersImageModel(Protocol):
+    """Generic class for a Model from transformers."""
+
+    def eval(self): ...
+    def get_image_features(*, pixel_values: "FloatTensor") -> "FloatTensor": ...
+
+
+class TransformersProcessor(Protocol):
+    """Generic class for a Processor from transformers."""
+
+    def __call__(self, *, images: list["Image.Image"], return_tensors: str): ...
+
+
 @dataclasses.dataclass
 class LoadedModels:
-    model: "AutoModel"
-    processor: "AutoProcessor"
+    model: TransformersImageModel
+    processor: TransformersProcessor
 
 
 class HFModelImageEmbedder(ImageEmbedder):
@@ -91,7 +103,7 @@ class HFModelImageEmbedder(ImageEmbedder):
     def model_revision(cls) -> str: ...
 
     @abc.abstractmethod
-    def _load_models(self) -> LoadedModels: ...
+    def load_models(self) -> LoadedModels: ...
 
     def embed(
         self, context: EmbedImageContext
@@ -122,19 +134,19 @@ class HFModelImageEmbedder(ImageEmbedder):
                 )
             )
 
-        models = self._load_models()
+        models = self.load_models()
         model = models.model
         processor = models.processor
-        model.eval()  # type: ignore[reportAttributeAccess]
+        model.eval()
 
         import torch
 
         with torch.no_grad():
             inputs = cast(
                 dict[str, torch.FloatTensor],
-                processor(images=images, return_tensors="pt"),  # type: ignore[reportAttributeAccess]
+                processor(images=images, return_tensors="pt"),
             )
-            image_features = model.get_image_features(  # type: ignore[reportAttributeAccess]
+            image_features = model.get_image_features(
                 pixel_values=inputs["pixel_values"]
             )
 
@@ -181,21 +193,21 @@ class Clip(HFModelImageEmbedder):
     def model_revision(cls) -> str:
         return "5812e510083bb2d23fa43778a39ac065d205ed4d"
 
-    def _load_models(self) -> LoadedModels:
+    def load_models(self) -> LoadedModels:
         from transformers.models.clip import (
             CLIPModel,
             CLIPProcessor,
         )
 
         model = cast(
-            AutoModel,
+            TransformersImageModel,
             CLIPModel.from_pretrained(  # pyright: ignore[reportUnknownMemberType]
                 pretrained_model_name_or_path=self.model_name(),
                 revision=self.model_revision(),
             ),
         )
         processor = cast(
-            AutoProcessor,
+            TransformersProcessor,
             CLIPProcessor.from_pretrained(  # pyright: ignore[reportUnknownMemberType]
                 pretrained_model_name_or_path=self.model_name(),
                 revision=self.model_revision(),
@@ -216,12 +228,12 @@ class SigLIP2(HFModelImageEmbedder):
     def model_revision(cls) -> str:
         return "a89f5c5093f902bf39d3cd4d81d2c09867f0724b"
 
-    def _load_models(self):
+    def load_models(self):
         from transformers.models.auto.modeling_auto import AutoModel
         from transformers.models.auto.processing_auto import AutoProcessor
 
         model = cast(
-            AutoModel,
+            TransformersImageModel,
             AutoModel.from_pretrained(  # pyright: ignore[reportUnknownMemberType]
                 pretrained_model_name_or_path=self.model_name(),
                 revision=self.model_revision(),
@@ -229,7 +241,7 @@ class SigLIP2(HFModelImageEmbedder):
             ),
         )
         processor = cast(
-            AutoProcessor,
+            TransformersProcessor,
             AutoProcessor.from_pretrained(  # pyright: ignore[reportUnknownMemberType]
                 pretrained_model_name_or_path=self.model_name(),
                 revision=self.model_revision(),

corvic/system/in_memory_executor.py

@@ -1539,8 +1539,7 @@ class InMemoryExecutor(OpGraphExecutor):
                 case Ok(query):
                     pass
             return (
-                await context.run_on_worker(
-                    self._staging_db.run_select_query,
+                await self._staging_db.run_select_query(
                     query,
                     expected_schema,
                     context.current_slice_args,

corvic/system/staging.py

@@ -45,7 +45,7 @@ class StagingDB(Protocol):
         similarity_metric: VectorSimilarityMetric,
     ) -> "sqlglot.exp.Query": ...
 
-    def run_select_query(
+    async def run_select_query(
         self,
         query: "sqlglot.exp.Query",
         expected_schema: pa.Schema,

corvic/system_sqlite/staging.py

@@ -393,7 +393,7 @@ class DuckDBStaging(StagingDB):
 
         return vector_search_query.with_(vector_union_table, as_=union)
 
-    def run_select_query(
+    async def run_select_query(
         self,
         query: sqlglot.exp.Query,
         expected_schema: pa.Schema,

{corvic_engine-0.3.0rc83.dist-info → corvic_engine-0.3.0rc85.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: corvic-engine
-Version: 0.3.0rc83
+Version: 0.3.0rc85
 Classifier: Environment :: Console
 Classifier: License :: Other/Proprietary License
 Classifier: Programming Language :: Python :: Implementation :: CPython
@@ -27,6 +27,7 @@ Requires-Dist: pillow>=10.0.0 ; extra == 'ml'
 Requires-Dist: scikit-learn>=1.4.0 ; extra == 'ml'
 Requires-Dist: transformers[torch]>=4.45.0 ; extra == 'ml'
 Requires-Dist: sentencepiece>=0.2.0 ; extra == 'ml'
+Requires-Dist: torchvision==0.22.1 ; extra == 'ml'
 Requires-Dist: opentelemetry-api>=1.20.0 ; extra == 'telemetry'
 Requires-Dist: opentelemetry-sdk>=1.20.0 ; extra == 'telemetry'
 Provides-Extra: ml

{corvic_engine-0.3.0rc83.dist-info → corvic_engine-0.3.0rc85.dist-info}/RECORD

@@ -25,7 +25,7 @@ corvic/emodel/_source.py,sha256=yWZeYHX_jFIwqRZ1eUHIMN9Gu8EIYEZ8fmXDO6QpfyQ,9799
 corvic/emodel/_space.py,sha256=ewAGHHVY4tkoMRjyKwAvLq_3Zky8Zaa7k3DEbhkhM7E,38549
 corvic/emodel/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 corvic/engine/__init__.py,sha256=XL4Vg7rNcBi29ccVelpeFizR9oJtGYXDn84W9zok9d4,975
-corvic/engine/_native.pyd,sha256=uHSk2hJH4b8Rxhbj3lb8hLPBsEYqvMsPtjfI18kIMSY,438272
+corvic/engine/_native.pyd,sha256=P11yaiVhGrz1RBR_GKA93caLSLzHA6hNHTY03-ajVnU,438272
 corvic/engine/_native.pyi,sha256=KYMPtvXqHZ-jMgZohLf4se3rr-rBpCihmjANcr6s8ag,1390
 corvic/engine/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 corvic/eorm/__init__.py,sha256=b4dFnu4fW7wj3Y0SMNVXOp8KoKOp_HAL4GyDN-S8fOY,13704
@@ -71,22 +71,22 @@ corvic/sql/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 corvic/system/__init__.py,sha256=PlsjTRklExrPA83n-PM5CqVmNcIQdIRhFHn3tZ0qh1Q,2765
 corvic/system/_column_encoding.py,sha256=feSWIv4vKstVq-aavWPk53YucUiq7rZvuyofqTicXBE,7574
 corvic/system/_dimension_reduction.py,sha256=2tg5SIHY4P480DJQj6PSjW1VgAJCAVJAH8D3BY-ZYXA,2964
-corvic/system/_embedder.py,sha256=3TJayoEKMOaooY3M2CZml0znwjPCguD-pbrGOXwamIg,6921
-corvic/system/_image_embedder.py,sha256=SiyoITnMfHXoLmtROdNJ1JlY6sBhA_kTxvbKYvZnRqY,12760
+corvic/system/_embedder.py,sha256=5_dXBxfeXnjuCRexTLhGWVpZ9OW9CLZDcYzxZMHgiD0,6653
+corvic/system/_image_embedder.py,sha256=O2hOEtSIoQRrf_xoRnPpSNGn_-aGZC-X1DPfuWykWPU,13003
 corvic/system/_planner.py,sha256=ecL-HW8PVz5eWJ1Ktf-RAD2IdZkHu3GuBtXdqElo4ts,8210
 corvic/system/_text_embedder.py,sha256=NDi--3_tzwIWVImjhFWmp8dHmydGGXNu6GYH8qODsIc,4000
 corvic/system/client.py,sha256=JcA-fPraqDkl9f8BiClS0qeGY6wzKcEDPymutWrJo54,812
-corvic/system/in_memory_executor.py,sha256=H51fob98zZ8weZtyaWd5TpZHv5KPIW3buTixYqCsOqU,66454
+corvic/system/in_memory_executor.py,sha256=4Wyi5pTCBVlsXJFansx1LT7OmfubzN7Hg6j1Uw8_8wo,66410
 corvic/system/op_graph_executor.py,sha256=tSKro-yb_y1_sgajZluM-6FCvDqO1oUPsiWw2DRxyMQ,3641
 corvic/system/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-corvic/system/staging.py,sha256=K5P5moiuAMfPx7lxK4mArxeURBwKoyB6x9HGu9JJ16E,1846
+corvic/system/staging.py,sha256=8XasqXY887n--lrcvp0E_nYnAz_D5_sqAZzgZP5PRoQ,1852
 corvic/system/storage.py,sha256=bp9llPmE6PwFta_9bhZ6d785ba3wbEysaL_0EzdAjPs,5575
 corvic/system_sqlite/__init__.py,sha256=F4UN9vFsXiDY2AKk1jYZPuWWJpSugKHS7ghXeZYlbZs,390
 corvic/system_sqlite/client.py,sha256=OqX2-3sjYzzkuZXRDTBDjfiDzzLx1jNSud7fXp8CO0A,7486
 corvic/system_sqlite/fs_blob_store.py,sha256=NTLzLFd56QNqA-iCxNjFAC-YePfXqWWTO9i_o1dJRr0,8563
 corvic/system_sqlite/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 corvic/system_sqlite/rdbms_blob_store.py,sha256=gTP_tQfTVb3wzZkzo8ys1zaz0rSrERzb57rqMHVpuBA,10563
-corvic/system_sqlite/staging.py,sha256=9WjXT_XT3wGCFgx9dieOGjHcuu2ZEaFWMChQaISMEIs,17332
+corvic/system_sqlite/staging.py,sha256=KoSIWV0xA6n8PnsSWSbITFZnYKc_UjYqW-Vx2o16jW4,17338
 corvic/table/__init__.py,sha256=Gj0IR8BQF5PZK92Us7PP0ZigMsVyrfWJupzH8TgzRQk,588
 corvic/table/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 corvic/table/table.py,sha256=fvt6G20Jz1Cfhf9zrmYIDPjUR4lGAWKjlx4eWKvoqtk,25815
@@ -98,9 +98,9 @@ corvic/version/__init__.py,sha256=JlkRLvKXsu3zIxhdynO_0Ub5NfQOvGjfwCRkNnaOu9U,11
 corvic/version/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 corvic/well_known_types/__init__.py,sha256=Btbeqieik2AcmijeOXeqBptzueBpgNitvH9J5VNm12w,1289
 corvic/well_known_types/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-corvic_engine-0.3.0rc83.dist-info/METADATA,sha256=--8vmdHNbuoemxu-cAytJy-VzCMw1avbGlbtNlA0eY4,1828
-corvic_engine-0.3.0rc83.dist-info/WHEEL,sha256=U8GAVPGuk_i_fNfXAys1dgH7mS1u7uA5UI-r13T9uEs,94
-corvic_engine-0.3.0rc83.dist-info/licenses/LICENSE,sha256=DSS1OD0oIgssKOmAzkMRBv5jvvVuZQbrIv8lpl9DXY8,1035
+corvic_engine-0.3.0rc85.dist-info/METADATA,sha256=b7VKq6Fo9qmnMARWRMYQvtnZs_jvOvYVxlF-j4qaUYE,1879
+corvic_engine-0.3.0rc85.dist-info/WHEEL,sha256=U8GAVPGuk_i_fNfXAys1dgH7mS1u7uA5UI-r13T9uEs,94
+corvic_engine-0.3.0rc85.dist-info/licenses/LICENSE,sha256=DSS1OD0oIgssKOmAzkMRBv5jvvVuZQbrIv8lpl9DXY8,1035
 corvic_generated/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 corvic_generated/algorithm/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 corvic_generated/algorithm/graph/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -211,4 +211,4 @@ corvic_generated/status/v1/service_pb2.py,sha256=CKXPX2ahq8O4cFhPpt6wo6l--6VZcgj
 corvic_generated/status/v1/service_pb2.pyi,sha256=iXLR2FOKQJpBgvBzpD2kVwcYOCksP2aRwK4JYaI9CBw,558
 corvic_generated/status/v1/service_pb2_grpc.py,sha256=y-a5ldrphWlNJW-yKswyjNmXokK4-5bbEEfczjagJHo,2736
 corvic_generated/status/v1/service_pb2_grpc.pyi,sha256=OoAnaZ64FD0UTzPoRhYvQU8ecoilhHj3ySjSfHbVDaU,1501
-corvic_engine-0.3.0rc83.dist-info/RECORD,,
+corvic_engine-0.3.0rc85.dist-info/RECORD,,