new: drop python3.9, replace optional and union with | (#574)

* new: drop python3.9, replace optional and union with |

* new: remove python 3.9 from pyproject

* refactor: replace remaining union and optional with |

* new: remove optional and union in dataclasses

* fix: add typealias to numpy type

* new: replace union with | in token count

Author: joein

.github/workflows/python-publish.yml

@@ -25,7 +25,7 @@ jobs:
     - name: Set up Python
       uses: actions/setup-python@v2
       with:
-        python-version: '3.9.x'
+        python-version: '3.10.x'
     - name: Install dependencies
       run: |
         python -m pip install poetry

.github/workflows/python-tests.yml

@@ -15,7 +15,6 @@ jobs:
     strategy:
       matrix:
         python-version:
-          - '3.9.x'
           - '3.10.x'
           - '3.11.x'
           - '3.12.x'

.github/workflows/type-checkers.yml

@@ -8,7 +8,7 @@ jobs:
     strategy:
       fail-fast: true
       matrix:
-        python-version: ["3.9", "3.10", "3.11", "3.12", "3.13"]
+        python-version: ["3.10", "3.11", "3.12", "3.13"]
         os: [ubuntu-latest]
 
     name: Python ${{ matrix.python-version }} test

fastembed/common/model_description.py

@@ -1,12 +1,12 @@
 from dataclasses import dataclass, field
 from enum import Enum
-from typing import Optional, Any
+from typing import Any
 
 
 @dataclass(frozen=True)
 class ModelSource:
-    hf: Optional[str] = None
-    url: Optional[str] = None
+    hf: str | None = None
+    url: str | None = None
     _deprecated_tar_struct: bool = False
 
     @property
@@ -33,17 +33,17 @@ class BaseModelDescription:
 
 @dataclass(frozen=True)
 class DenseModelDescription(BaseModelDescription):
-    dim: Optional[int] = None
-    tasks: Optional[dict[str, Any]] = field(default_factory=dict)
+    dim: int | None = None
+    tasks: dict[str, Any] | None = field(default_factory=dict)
 
     def __post_init__(self) -> None:
         assert self.dim is not None, "dim is required for dense model description"
 
 
 @dataclass(frozen=True)
 class SparseModelDescription(BaseModelDescription):
-    requires_idf: Optional[bool] = None
-    vocab_size: Optional[int] = None
+    requires_idf: bool | None = None
+    vocab_size: int | None = None
 
 
 class PoolingType(str, Enum):

fastembed/common/model_management.py

@@ -5,7 +5,7 @@
 import tarfile
 from copy import deepcopy
 from pathlib import Path
-from typing import Any, Optional, Union, TypeVar, Generic
+from typing import Any, TypeVar, Generic
 
 import requests
 from huggingface_hub import snapshot_download, model_info, list_repo_tree
@@ -180,8 +180,8 @@ def _verify_files_from_metadata(
 
         def _collect_file_metadata(
             model_dir: Path, repo_files: list[RepoFile]
-        ) -> dict[str, dict[str, Union[int, str]]]:
-            meta: dict[str, dict[str, Union[int, str]]] = {}
+        ) -> dict[str, dict[str, int | str]]:
+            meta: dict[str, dict[str, int | str]] = {}
             file_info_map = {f.path: f for f in repo_files}
             for file_path in model_dir.rglob("*"):
                 if file_path.is_file() and file_path.name != cls.METADATA_FILE:
@@ -193,9 +193,7 @@ def _collect_file_metadata(
                         }
             return meta
 
-        def _save_file_metadata(
-            model_dir: Path, meta: dict[str, dict[str, Union[int, str]]]
-        ) -> None:
+        def _save_file_metadata(model_dir: Path, meta: dict[str, dict[str, int | str]]) -> None:
             try:
                 if not model_dir.exists():
                     model_dir.mkdir(parents=True, exist_ok=True)
@@ -397,7 +395,7 @@ def download_model(cls, model: T, cache_dir: str, retries: int = 3, **kwargs: An
             Path: The path to the downloaded model directory.
         """
         local_files_only = kwargs.get("local_files_only", False)
-        specific_model_path: Optional[str] = kwargs.pop("specific_model_path", None)
+        specific_model_path: str | None = kwargs.pop("specific_model_path", None)
         if specific_model_path:
             return Path(specific_model_path)
         retries = 1 if local_files_only else retries

fastembed/common/onnx_model.py

@@ -1,7 +1,7 @@
 import warnings
 from dataclasses import dataclass
 from pathlib import Path
-from typing import Any, Generic, Iterable, Optional, Sequence, Type, TypeVar
+from typing import Any, Generic, Iterable, Sequence, Type, TypeVar
 
 import numpy as np
 import onnxruntime as ort
@@ -19,8 +19,8 @@
 @dataclass
 class OnnxOutputContext:
     model_output: NumpyArray
-    attention_mask: Optional[NDArray[np.int64]] = None
-    input_ids: Optional[NDArray[np.int64]] = None
+    attention_mask: NDArray[np.int64] | None = None
+    input_ids: NDArray[np.int64] | None = None
 
 
 class OnnxModel(Generic[T]):
@@ -43,8 +43,8 @@ def _post_process_onnx_output(self, output: OnnxOutputContext, **kwargs: Any) ->
         raise NotImplementedError("Subclasses must implement this method")
 
     def __init__(self) -> None:
-        self.model: Optional[ort.InferenceSession] = None
-        self.tokenizer: Optional[Tokenizer] = None
+        self.model: ort.InferenceSession | None = None
+        self.tokenizer: Tokenizer | None = None
 
     def _preprocess_onnx_input(
         self, onnx_input: dict[str, NumpyArray], **kwargs: Any
@@ -58,11 +58,11 @@ def _load_onnx_model(
         self,
         model_dir: Path,
         model_file: str,
-        threads: Optional[int],
-        providers: Optional[Sequence[OnnxProvider]] = None,
+        threads: int | None,
+        providers: Sequence[OnnxProvider] | None = None,
         cuda: bool = False,
-        device_id: Optional[int] = None,
-        extra_session_options: Optional[dict[str, Any]] = None,
+        device_id: int | None = None,
+        extra_session_options: dict[str, Any] | None = None,
     ) -> None:
         model_path = model_dir / model_file
         # List of Execution Providers: https://onnxruntime.ai/docs/execution-providers

fastembed/common/types.py

@@ -1,25 +1,20 @@
 from pathlib import Path
-import sys
-from PIL import Image
-from typing import Any, Union
+
+from typing import Any, TypeAlias
 import numpy as np
 from numpy.typing import NDArray
-
-if sys.version_info >= (3, 10):
-    from typing import TypeAlias
-else:
-    from typing_extensions import TypeAlias
+from PIL import Image
 
 
-PathInput: TypeAlias = Union[str, Path]
-ImageInput: TypeAlias = Union[PathInput, Image.Image]
+PathInput: TypeAlias = str | Path
+ImageInput: TypeAlias = PathInput | Image.Image
 
-OnnxProvider: TypeAlias = Union[str, tuple[str, dict[Any, Any]]]
-NumpyArray = Union[
-    NDArray[np.float64],
-    NDArray[np.float32],
-    NDArray[np.float16],
-    NDArray[np.int8],
-    NDArray[np.int64],
-    NDArray[np.int32],
-]
+OnnxProvider: TypeAlias = str | tuple[str, dict[Any, Any]]
+NumpyArray: TypeAlias = (
+    NDArray[np.float64]
+    | NDArray[np.float32]
+    | NDArray[np.float16]
+    | NDArray[np.int8]
+    | NDArray[np.int64]
+    | NDArray[np.int32]
+)

fastembed/common/utils.py

@@ -5,7 +5,7 @@
 import unicodedata
 from pathlib import Path
 from itertools import islice
-from typing import Iterable, Optional, TypeVar
+from typing import Iterable, TypeVar
 
 import numpy as np
 from numpy.typing import NDArray
@@ -45,7 +45,7 @@ def iter_batch(iterable: Iterable[T], size: int) -> Iterable[list[T]]:
         yield b
 
 
-def define_cache_dir(cache_dir: Optional[str] = None) -> Path:
+def define_cache_dir(cache_dir: str | None = None) -> Path:
     """
     Define the cache directory for fastembed
     """

fastembed/embedding.py

@@ -1,4 +1,4 @@
-from typing import Optional, Any
+from typing import Any
 
 from loguru import logger
 
@@ -17,8 +17,8 @@ class JinaEmbedding(TextEmbedding):
     def __init__(
         self,
         model_name: str = "jinaai/jina-embeddings-v2-base-en",
-        cache_dir: Optional[str] = None,
-        threads: Optional[int] = None,
+        cache_dir: str | None = None,
+        threads: int | None = None,
         **kwargs: Any,
     ):
         super().__init__(model_name, cache_dir, threads, **kwargs)

fastembed/image/image_embedding.py

@@ -1,4 +1,4 @@
-from typing import Any, Iterable, Optional, Sequence, Type, Union
+from typing import Any, Iterable, Sequence, Type
 from dataclasses import asdict
 
 from fastembed.common.types import NumpyArray
@@ -48,11 +48,11 @@ def _list_supported_models(cls) -> list[DenseModelDescription]:
     def __init__(
         self,
         model_name: str,
-        cache_dir: Optional[str] = None,
-        threads: Optional[int] = None,
-        providers: Optional[Sequence[OnnxProvider]] = None,
+        cache_dir: str | None = None,
+        threads: int | None = None,
+        providers: Sequence[OnnxProvider] | None = None,
         cuda: bool = False,
-        device_ids: Optional[list[int]] = None,
+        device_ids: list[int] | None = None,
         lazy_load: bool = False,
         **kwargs: Any,
     ):
@@ -98,7 +98,7 @@ def get_embedding_size(cls, model_name: str) -> int:
             ValueError: If the model name is not found in the supported models.
         """
         descriptions = cls._list_supported_models()
-        embedding_size: Optional[int] = None
+        embedding_size: int | None = None
         for description in descriptions:
             if description.model.lower() == model_name.lower():
                 embedding_size = description.dim
@@ -113,9 +113,9 @@ def get_embedding_size(cls, model_name: str) -> int:
 
     def embed(
         self,
-        images: Union[ImageInput, Iterable[ImageInput]],
+        images: ImageInput | Iterable[ImageInput],
         batch_size: int = 16,
-        parallel: Optional[int] = None,
+        parallel: int | None = None,
         **kwargs: Any,
     ) -> Iterable[NumpyArray]:
         """