Numpy now lazily imported. For typing, removed numpy extra. justfile now uses with

Author: caseyclements

bson/binary.py

@@ -65,16 +65,9 @@
     from array import array as _array
     from mmap import mmap as _mmap
 
-
-_NUMPY_AVAILABLE = False
-try:
     import numpy as np
     import numpy.typing as npt
 
-    _NUMPY_AVAILABLE = True
-except ImportError:
-    np = None  # type: ignore
-
 
 class UuidRepresentation:
     UNSPECIFIED = 0
@@ -492,9 +485,7 @@ def from_vector(
             )
         metadata = struct.pack("<sB", dtype.value, padding)
 
-        if _NUMPY_AVAILABLE and isinstance(vector, np.ndarray):
-            data = _numpy_vector_to_bytes(vector, dtype)
-        else:
+        if isinstance(vector, list):
             if dtype == BinaryVectorDtype.INT8:  # pack ints in [-128, 127] as signed int8
                 format_str = "b"
                 if padding:
@@ -511,7 +502,36 @@ def from_vector(
                     raise ValueError(f"padding does not apply to {dtype=}")
             else:
                 raise NotImplementedError("%s not yet supported" % dtype)
-            data = struct.pack(f"<{len(vector)}{format_str}", *vector)  # type: ignore
+            data = struct.pack(f"<{len(vector)}{format_str}", *vector)
+        else:  # vector is numpy array or incorrect type.
+            try:
+                import numpy as np
+            except ImportError as exc:
+                raise ImportError(
+                    "Failed to create binary from vector. Check type. If numpy array, numpy must be installed."
+                ) from exc
+            if not isinstance(vector, np.ndarray):
+                raise TypeError("Vector must be a numpy array.")
+            if vector.ndim != 1:
+                raise ValueError(
+                    "from_numpy_vector only supports 1D arrays as it creates a single vector."
+                )
+
+            if dtype == BinaryVectorDtype.FLOAT32:
+                vector = vector.astype(np.dtype("float32"), copy=False)
+            elif dtype == BinaryVectorDtype.INT8:
+                if vector.min() >= -128 and vector.max() <= 127:
+                    vector = vector.astype(np.dtype("int8"), copy=False)
+                else:
+                    raise ValueError("Values found outside INT8 range.")
+            elif dtype == BinaryVectorDtype.PACKED_BIT:
+                if vector.min() >= 0 and vector.max() <= 127:
+                    vector = vector.astype(np.dtype("uint8"), copy=False)
+                else:
+                    raise ValueError("Values found outside UINT8 range.")
+            else:
+                raise NotImplementedError("%s not yet supported" % dtype)
+            data = vector.tobytes()
 
         if padding and len(vector) and not (data[-1] & ((1 << padding) - 1)) == 0:
             raise ValueError(
@@ -596,8 +616,13 @@ def as_numpy_vector(self) -> BinaryVector:
         """
         if self.subtype != VECTOR_SUBTYPE:
             raise ValueError(f"Cannot decode subtype {self.subtype} as a vector")
-        if not _NUMPY_AVAILABLE:
-            raise ImportError("Converting binary to numpy.ndarray requires numpy to be installed.")
+        try:
+            import numpy as np
+        except ImportError as exc:
+            raise ImportError(
+                "Converting binary to numpy.ndarray requires numpy to be installed."
+            ) from exc
+
         dtype, padding = struct.unpack_from("<sB", self, 0)
         dtype = BinaryVectorDtype(dtype)
 
@@ -637,32 +662,3 @@ def __repr__(self) -> str:
             return f"<Binary(REDACTED, {self.__subtype})>"
         else:
             return f"Binary({bytes.__repr__(self)}, {self.__subtype})"
-
-
-def _numpy_vector_to_bytes(
-    vector: npt.NDArray[np.number],
-    dtype: BinaryVectorDtype,
-) -> bytes:
-    if not _NUMPY_AVAILABLE:
-        raise ImportError("Converting numpy.ndarray to binary requires numpy to be installed.")
-
-    if not isinstance(vector, np.ndarray):
-        raise TypeError("Vector must be a numpy array.")
-    if vector.ndim != 1:
-        raise ValueError("from_numpy_vector only supports 1D arrays as it creates a single vector.")
-
-    if dtype == BinaryVectorDtype.FLOAT32:
-        vector = vector.astype(np.dtype("float32"), copy=False)
-    elif dtype == BinaryVectorDtype.INT8:
-        if vector.min() >= -128 and vector.max() <= 127:
-            vector = vector.astype(np.dtype("int8"), copy=False)
-        else:
-            raise ValueError("Values found outside INT8 range.")
-    elif dtype == BinaryVectorDtype.PACKED_BIT:
-        if vector.min() >= 0 and vector.max() <= 127:
-            vector = vector.astype(np.dtype("uint8"), copy=False)
-        else:
-            raise ValueError("Values found outside UINT8 range.")
-    else:
-        raise NotImplementedError("%s not yet supported" % dtype)
-    return vector.tobytes()

justfile

@@ -2,7 +2,7 @@
 set shell := ["bash", "-c"]
 
 # Commonly used command segments.
-typing_run := "uv run --group typing --extra aws --extra encryption --extra numpy --extra ocsp --extra snappy --extra test --extra zstd"
+typing_run := "uv run --group typing --extra aws --extra encryption --with numpy --extra ocsp --extra snappy --extra test --extra zstd"
 docs_run := "uv run --extra docs"
 doc_build := "./doc/_build"
 mypy_args := "--install-types --non-interactive"
@@ -39,14 +39,14 @@ typing: && resync
 
 [group('typing')]
 typing-mypy: && resync
-    {{typing_run}} mypy {{mypy_args}} bson gridfs tools pymongo
-    {{typing_run}} mypy {{mypy_args}} --config-file mypy_test.ini test
-    {{typing_run}} mypy {{mypy_args}} test/test_typing.py test/test_typing_strict.py
+    {{typing_run}} python -m mypy {{mypy_args}} bson gridfs tools pymongo
+    {{typing_run}} python -m mypy {{mypy_args}} --config-file mypy_test.ini test
+    {{typing_run}} python -m mypy {{mypy_args}} test/test_typing.py test/test_typing_strict.py
 
 [group('typing')]
 typing-pyright: && resync
-    {{typing_run}} pyright test/test_typing.py test/test_typing_strict.py
-    {{typing_run}} pyright -p strict_pyrightconfig.json test/test_typing_strict.py
+    {{typing_run}} python -m pyright test/test_typing.py test/test_typing_strict.py
+    {{typing_run}} python -m pyright -p strict_pyrightconfig.json test/test_typing_strict.py
 
 [group('lint')]
 lint *args="": && resync
@@ -58,7 +58,13 @@ lint-manual *args="": && resync
 
 [group('test')]
 test *args="-v --durations=5 --maxfail=10": && resync
-    uv run --extra test pytest {{args}}
+    uv run --extra test python -m pytest {{args}}
+
+[group('test')]
+test-bson *args="-v --durations=5 --maxfail=10": && resync
+    uv run --extra test --with numpy python -m pytest test/test_bson.py
+
+
 
 [group('test')]
 run-tests *args: && resync

pyproject.toml

@@ -87,7 +87,6 @@ ocsp = ["requirements/ocsp.txt"]
 snappy = ["requirements/snappy.txt"]
 test = ["requirements/test.txt"]
 zstd = ["requirements/zstd.txt"]
-numpy = ["requirements/numpy.txt"]
 
 [tool.pytest.ini_options]
 minversion = "7"

requirements/numpy.txt

@@ -19,6 +19,7 @@
 import array
 import collections
 import datetime
+import importlib.util
 import mmap
 import os
 import pickle
@@ -71,13 +72,7 @@
 from bson.timestamp import Timestamp
 from bson.tz_util import FixedOffset, utc
 
-_NUMPY_AVAILABLE = False
-try:
-    import numpy as np
-
-    _NUMPY_AVAILABLE = True
-except ImportError:
-    np = None  # type: ignore
+_NUMPY_AVAILABLE = importlib.util.find_spec("numpy") is not None
 
 
 class NotADict(abc.MutableMapping):
@@ -883,6 +878,8 @@ def test_binaryvector_equality(self):
     def test_vector_from_numpy(self):
         """Follows test_vector except for input type numpy.ndarray"""
         # Simple data values could be treated as any of our BinaryVectorDtypes
+        import numpy as np
+
         arr = np.array([2, 3])
         # INT8
         binary_vector_int8 = Binary.from_vector(arr, BinaryVectorDtype.INT8)