Improve buffer-accepting hashes and more

CHANGELOG.md

@@ -10,30 +10,43 @@ This project has adhered to
 
 ### Added
 
-- Add `digest` functions that accept a non-immutable buffer as input
-  and process it without internal copying
-  ([#75](https://github.com/hajimes/mmh3/issues/75)).
-- Slightly improve the performance of the `hash_bytes` function.
 - Add support for Python 3.13.
+- Add `digest` functions that support the new buffer protocol
+  ([PEP 688](https://peps.python.org/pep-0688/)) as input
+  ([#75](https://github.com/hajimes/mmh3/pull/75)).
+  These functions are implemented with
+  [METH_FASTCALL](https://docs.python.org/3/c-api/structures.html#c.METH_FASTCALL),
+  offering improved performance over legacy functions.
+- Slightly improve the performance of the `hash_bytes()` function.
 - Add Read the Docs documentation
   ([#54](https://github.com/hajimes/mmh3/issues/54)).
 - (planned: Document benchmark results
   ([#53](https://github.com/hajimes/mmh3/issues/53))).
 
 ### Changed
 
+- **Backward-incompatible**: The `seed` argument is now strictly validated to
+  ensure it falls within the range [0, 0xFFFFFFFF]. A `ValueError` is raised
+  if the seed is out of range.
+- **Backward-incompatible**: Change the constructors of hasher classes to
+  accept a buffer as the first argument
+  ([#83](https://github.com/hajimes/mmh3/pull/83)).
+- The type of flag argumens has been changed from `bool` to `Any`.
 - Change the format of CHANGELOG.md to conform to the
   [Keep a Changelog](https://keepachangelog.com/en/1.1.0/) standard
-  ([#63](https://github.com/hajimes/mmh3/issues/63)).
-- **Backward-incompatible**: Change the constructors of hasher classes to
-  accept a buffer as the first argument.
+  ([#63](https://github.com/hajimes/mmh3/pull/63)).
+
+### Deprecated
+
+- Deprecate the `hash_from_buffer()` function.
+  Use `mmh3_32_sintdigest()` or `mmh3_32_uintdigest()` as alternatives.
 
 ### Fixed
 
 - Fix a reference leak in the `hash_from_buffer()` function
-  ([#75](https://github.com/hajimes/mmh3/issues/75)).
-- Fix type hints ([#76](https://github.com/hajimes/mmh3/issues/76),
-  [#77](https://github.com/hajimes/mmh3/issues/77)).
+  ([#75](https://github.com/hajimes/mmh3/pull/75)).
+- Fix type hints ([#76](https://github.com/hajimes/mmh3/pull/76),
+  [#77](https://github.com/hajimes/mmh3/pull/77)).
 
 ## [4.1.0] - 2024-01-09
 
@@ -47,7 +60,7 @@ This project has adhered to
   ([#50](https://github.com/hajimes/mmh3/issues/50)).
 - Fix incorrect type hints ([#51](https://github.com/hajimes/mmh3/issues/51)).
 - Fix invalid results on s390x when the arg `x64arch` of `hash64` or
-  `hash_bytes` is set to `False`
+  `hash_bytes()` is set to `False`
   ([#52](https://github.com/hajimes/mmh3/issues/52)).
 
 ## [4.0.1] - 2023-07-14
@@ -97,8 +110,8 @@ This project has adhered to
   [wouter bolsterlee](https://github.com/wbolster) and
   [Dušan Nikolić](https://github.com/n-dusan)!
 - Add support for 32-bit architectures such as `i686` and `armv7l`. From now on,
-  `hash` and `hash_from_buffer` on these architectures will generate the same
-  hash values as those on other environments. Thanks
+  `hash()` and `hash_from_buffer()` on these architectures will generate the
+  same hash values as those on other environments. Thanks
   [Danil Shein](https://github.com/dshein-alt)!
 - In relation to the above, `manylinux2014_i686` wheels are now available.
 - Support for hashing huge data (>16GB). Thanks
@@ -134,13 +147,13 @@ This project has adhered to
 
 ### Fixed
 
-- Bugfix for `hash_bytes`. Thanks [doozr](https://github.com/doozr)!
+- Bugfix for `hash_bytes()`. Thanks [doozr](https://github.com/doozr)!
 
 ## [2.5] - 2017-10-28
 
 ### Added
 
-- Add `hash_from_buffer`. Thanks [Dimitri Vorona](https://github.com/alendit)!
+- Add `hash_from_buffer()`. Thanks [Dimitri Vorona](https://github.com/alendit)!
 - Add a keyword argument `signed`.
 
 ## [2.4] - 2017-05-27
@@ -175,7 +188,7 @@ Thanks!
 
 ### Added
 
-- Add `hash128`, which returns a 128-bit signed integer.
+- Add `hash128()`, which returns a 128-bit signed integer.
 
 ### Fixed
 

README.md

@@ -128,37 +128,50 @@ b'\x82_n\xdd \xac\xb6j\xef\x99\xb1e\xc4\n\xc9\xfd'
 
 ## Changelog
 
-See [Changelog](https://mmh3.readthedocs.io/en/latest/changelog_link.html) for the
+See [Changelog](https://mmh3.readthedocs.io/en/latest/changelog.html) for the
 complete changelog.
 
 ### [Unreleased]
 
 #### Added
 
-- Add `digest` functions that accept a non-immutable buffer as input
-  and process it without internal copying
-  ([#75](https://github.com/hajimes/mmh3/issues/75)).
-- Slightly improve the performance of the `hash_bytes` function.
 - Add support for Python 3.13.
+- Add `digest` functions that support the new buffer protocol
+  ([PEP 688](https://peps.python.org/pep-0688/)) as input
+  ([#75](https://github.com/hajimes/mmh3/pull/75)).
+  These functions are implemented with
+  [METH_FASTCALL](https://docs.python.org/3/c-api/structures.html#c.METH_FASTCALL),
+  offering improved performance over legacy functions.
+- Slightly improve the performance of the `hash_bytes()` function.
 - Add Read the Docs documentation
   ([#54](https://github.com/hajimes/mmh3/issues/54)).
 - (planned: Document benchmark results
   ([#53](https://github.com/hajimes/mmh3/issues/53))).
 
 #### Changed
 
+- **Backward-incompatible**: The `seed` argument is now strictly validated to
+  ensure it falls within the range [0, 0xFFFFFFFF]. A `ValueError` is raised
+  if the seed is out of range.
+- **Backward-incompatible**: Change the constructors of hasher classes to
+  accept a buffer as the first argument
+  ([#83](https://github.com/hajimes/mmh3/pull/83)).
+- The type of flag argumens has been changed from `bool` to `Any`.
 - Change the format of CHANGELOG.md to conform to the
   [Keep a Changelog](https://keepachangelog.com/en/1.1.0/) standard
-  ([#63](https://github.com/hajimes/mmh3/issues/63)).
-- **Backward-incompatible**: Change the constructors of hasher classes to
-  accept a buffer as the first argument.
+  ([#63](https://github.com/hajimes/mmh3/pull/63)).
+
+#### Deprecated
+
+- Deprecate the `hash_from_buffer()` function.
+  Use `mmh3_32_sintdigest()` or `mmh3_32_uintdigest()` as alternatives.
 
 #### Fixed
 
 - Fix a reference leak in the `hash_from_buffer()` function
-  ([#75](https://github.com/hajimes/mmh3/issues/75)).
-- Fix type hints ([#76](https://github.com/hajimes/mmh3/issues/76),
-  [#77](https://github.com/hajimes/mmh3/issues/77)).
+  ([#75](https://github.com/hajimes/mmh3/pull/75)).
+- Fix type hints ([#76](https://github.com/hajimes/mmh3/pull/76),
+  [#77](https://github.com/hajimes/mmh3/pull/77)).
 
 ### [4.1.0] - 2024-01-09
 
@@ -172,7 +185,7 @@ complete changelog.
   ([#50](https://github.com/hajimes/mmh3/issues/50)).
 - Fix incorrect type hints ([#51](https://github.com/hajimes/mmh3/issues/51)).
 - Fix invalid results on s390x when the arg `x64arch` of `hash64` or
-  `hash_bytes` is set to `False`
+  `hash_bytes()` is set to `False`
   ([#52](https://github.com/hajimes/mmh3/issues/52)).
 
 ## License
@@ -201,29 +214,6 @@ For compatibility with
 [murmur3 (Go)](https://pkg.go.dev/github.com/spaolacci/murmur3), see
 <https://github.com/hajimes/mmh3/issues/46>.
 
-### Unexpected results when given non 32-bit seeds
-
-In version 2.4, the type of a seed was changed from a signed 32-bit integer to
-an unsigned 32-bit integer. However, the resulting values for signed seeds
-remain unchanged from previous versions, as long as they are 32-bit.
-
-```pycon
->>> mmh3.hash("aaaa", -1756908916) # signed representation for 0x9747b28c
-1519878282
->>> mmh3.hash("aaaa", 2538058380) # unsigned representation for 0x9747b28c
-1519878282
-```
-
-Be careful so that these seeds do not exceed 32-bit. Unexpected results may
-happen with invalid values.
-
-```pycon
->>> mmh3.hash("foo", 2 ** 33)
--156908512
->>> mmh3.hash("foo", 2 ** 34)
--156908512
-```
-
 ## Contributing Guidelines
 
 See [Contributing](https://mmh3.readthedocs.io/en/latest/CONTRIBUTING.html).

benchmark/plot_graph.py

@@ -124,8 +124,8 @@ def ordered_intersection(list1: list[T], list2: list[T]) -> list[T]:
     plt.savefig(os.path.join(args.output_dir, BANDWIDTH_SMALL_FILE_NAME))
 
     df_latency_all = df_latency * 1000
-    df_latency_all.index = df_latency_all.index / (1024 * 1024)
-    df_latency_all.plot(xlabel="Input size (MiB)", ylabel="Latency (ms)")
+    df_latency_all.index = df_latency_all.index / 1024
+    df_latency_all.plot(xlabel="Input size (KiB)", ylabel="Latency (ms)")
     plt.savefig(os.path.join(args.output_dir, LATENCY_FILE_NAME))
 
     df_latency_small = df_latency * 1000 * 1000 * 1000

docs/CONTRIBUTING.md

@@ -129,13 +129,13 @@ The idea of the subproject directory loosely follows the
 
 ### Updating mmh3 core C code
 
-Run `tox -e build-cfiles`. This will fetch Appleby's original SMHasher project
+Run `tox -e build_cfiles`. This will fetch Appleby's original SMHasher project
 as a git submodule and then generate PEP 7-compliant C code from the original
 project.
 
 To perform further edits, add transformation code to the `refresh.py` script,
 instead of editing `murmurhash3.*` files manually.
-Then, run `tox -e build-cfiles` again to update the `murmurhash3.*` files.
+Then, run `tox -e build_cfiles` again to update the `murmurhash3.*` files.
 
 ### Local files
 
@@ -153,8 +153,7 @@ Then, run `tox -e build-cfiles` again to update the `murmurhash3.*` files.
 To run benchmarks locally, try the following command:
 
 ```shell
-pip install ".[benchmark]"
-python benchmark/benchmark.py -o OUTPUT_FILE \
+tox -e benchmark -- -o OUTPUT_FILE \
             --test-hash HASH_NAME --test-buffer-size-max HASH_SIZE
 ```
 
@@ -165,9 +164,8 @@ in bytes.
 For example,
 
 ```shell
-pip install ".[benchmark]"
-mkdir results
-python benchmark/benchmark.py -o results/mmh3_128.json \
+mkdir -p _results
+tox -e benchmark -- -o _results/mmh3_128.json \
             --test-hash mmh3_128 --test-buffer-size-max 262144
 ```
 
@@ -182,8 +180,7 @@ After obtaining the benchmark results, you can plot graphs by `plot_graph.py`.
 The following is an example of how to run the script:
 
 ```shell
-pip install ".[benchmark,plot]"
-python benchmark/plot_graph.py --output-dir docs/_static RESULT_DIR/*.json
+tox -e plot -- --output-dir docs/_static RESULT_DIR/*.json
 ```
 
 where `RESULT_DIR` is the directory containing the benchmark results.

docs/api.md

@@ -22,9 +22,8 @@ UTF-8 encoding before hashing.
 
 The following functions are used to hash types that implement the buffer
 protocol such as `bytes`, `bytearray`, `memoryview`, and `numpy` arrays.
-String inputs are also supported and are automatically converted to `bytes`
-using UTF-8 encoding before hashing.
 
+```{seealso}
 The buffer protocol,
 [originally implemented as a part of Python/C API](https://docs.python.org/3/c-api/buffer.html),
 was formally defined as a Python-level API in
@@ -37,6 +36,7 @@ type hint
 which is itself an alias for
 [typing_extensions.Buffer](https://typing-extensions.readthedocs.io/en/latest/#typing_extensions.Buffer),
 the backported type hint for `collections.abc.Buffer`.
+```
 
 ```{eval-rst}
 .. autofunction:: mmh3.hash_from_buffer

docs/index.rst

@@ -6,9 +6,9 @@ mmh3 is a Python extension for `MurmurHash (MurmurHash3) <https://en.wikipedia.o
    :maxdepth: 2
    :caption: User Guideline
 
-   Quickstart<readme_link>
+   Quickstart<quickstart>
    api
-   Changelog<changelog_link>
+   Changelog<changelog>
 
 .. toctree::
    :maxdepth: 2
@@ -21,5 +21,4 @@ Indices and tables
 ==================
 
 * :ref:`genindex`
-* :ref:`modindex`
 * :ref:`search`

src/mmh3/__init__.pyi

@@ -2,26 +2,24 @@
 from __future__ import annotations
 
 import sys
-from typing import Union, final
+from typing import Any, Union, final
 
 if sys.version_info >= (3, 12):
     from collections.abc import Buffer
 else:
     from _typeshed import ReadableBuffer as Buffer
 
-def hash(key: Union[bytes, str], seed: int = 0, signed: bool = True) -> int: ...
+def hash(key: Union[bytes, str], seed: int = 0, signed: Any = True) -> int: ...
 def hash_from_buffer(
-    key: Union[Buffer, str], seed: int = 0, signed: bool = True
+    key: Union[Buffer, str], seed: int = 0, signed: Any = True
 ) -> int: ...
 def hash64(
-    key: Union[bytes, str], seed: int = 0, x64arch: bool = True, signed: bool = True
+    key: Union[bytes, str], seed: int = 0, x64arch: Any = True, signed: Any = True
 ) -> tuple[int, int]: ...
 def hash128(
-    key: Union[bytes, str], seed: int = 0, x64arch: bool = True, signed: bool = False
+    key: Union[bytes, str], seed: int = 0, x64arch: Any = True, signed: Any = False
 ) -> int: ...
-def hash_bytes(
-    key: Union[bytes, str], seed: int = 0, x64arch: bool = True
-) -> bytes: ...
+def hash_bytes(key: Union[bytes, str], seed: int = 0, x64arch: Any = True) -> bytes: ...
 def mmh3_32_digest(key: Union[Buffer, str], seed: int = 0) -> bytes: ...
 def mmh3_32_sintdigest(key: Union[Buffer, str], seed: int = 0) -> int: ...
 def mmh3_32_uintdigest(key: Union[Buffer, str], seed: int = 0) -> int: ...