ormsgpack 1.6.0__tar.gz → 1.7.0__tar.gz

{ormsgpack-1.6.0 → ormsgpack-1.7.0}/CHANGELOG.md

@@ -1,5 +1,20 @@
 # Changelog
 
+## 1.7.0 29/11/2024
+
+### Fixed
+
+- Detect Pydantic 2.10 models [#311](/../../issues/311)
+- Fix serialization of dataclasses without `__slots__` and with a field defined
+  with a descriptor object as default value, a field defined with `init=False`
+  and a default value, or a cached property
+
+### Changed
+
+- Drop support for Python 3.8
+- Support `OPT_SORT_KEYS` also for Pydantic models [#312](/../../issues/312)
+- Improve deserialization performance
+
 ## 1.6.0 18/10/2024
 
 ### Fixed

{ormsgpack-1.6.0 → ormsgpack-1.7.0}/Cargo.lock

@@ -14,21 +14,6 @@ dependencies = [
  "zerocopy",
 ]
 
-[[package]]
-name = "any_all_workaround"
-version = "0.1.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "88fea40735f2cc320a5133ce772d39c571bd6c9b0d4c1a326926eecdd5af2e86"
-dependencies = [
- "cfg-if",
-]
-
-[[package]]
-name = "associative-cache"
-version = "2.0.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b993cd767a2bc7307dd87622311ca22c44329cc7a21366206bfa0896827b2bad"
-
 [[package]]
 name = "autocfg"
 version = "1.3.0"
@@ -68,16 +53,6 @@ version = "0.2.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "7a81dae078cea95a014a339291cec439d2f232ebe854a9d672b796c6afafa9b7"
 
-[[package]]
-name = "encoding_rs"
-version = "0.8.34"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "b45de904aa0b010bce2ab45264d0631681847fa7b6f2eaa7dab7619943bc4f59"
-dependencies = [
- "any_all_workaround",
- "cfg-if",
-]
-
 [[package]]
 name = "half"
 version = "2.4.1"
@@ -90,9 +65,9 @@ dependencies = [
 
 [[package]]
 name = "itoa"
-version = "1.0.11"
+version = "1.0.13"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "49f1f14873335454500d59611f1cf4a4b0f786f9ac11f4312a78e4cf2566695b"
+checksum = "540654e97a3f4470a492cd30ff187bc95d89557a903a2bbf112e2fae98104ef2"
 
 [[package]]
 name = "libc"
@@ -126,22 +101,21 @@ checksum = "1261fe7e33c73b354eab43b1273a57c8f967d0391e80353e51f764ac02cf6775"
 
 [[package]]
 name = "ormsgpack"
-version = "1.6.0"
+version = "1.7.0"
 dependencies = [
  "ahash",
- "associative-cache",
  "bytecount",
+ "byteorder",
  "chrono",
- "encoding_rs",
  "half",
  "itoa",
  "once_cell",
  "pyo3",
  "pyo3-build-config",
  "rmp",
- "rmp-serde",
  "serde",
  "serde_bytes",
+ "simdutf8",
  "smallvec",
 ]
 
@@ -159,18 +133,18 @@ checksum = "7170ef9988bc169ba16dd36a7fa041e5c4cbeb6a35b76d4c03daded371eae7c0"
 
 [[package]]
 name = "proc-macro2"
-version = "1.0.82"
+version = "1.0.89"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "8ad3d49ab951a01fbaafe34f2ec74122942fe18a3f9814c3268f1bb72042131b"
+checksum = "f139b0662de085916d1fb67d2b4169d1addddda1919e696f3252b740b629986e"
 dependencies = [
  "unicode-ident",
 ]
 
 [[package]]
 name = "pyo3"
-version = "0.22.5"
+version = "0.22.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "3d922163ba1f79c04bc49073ba7b32fd5a8d3b76a87c955921234b8e77333c51"
+checksum = "f402062616ab18202ae8319da13fa4279883a2b8a9d9f83f20dbade813ce1884"
 dependencies = [
  "cfg-if",
  "libc",
@@ -183,9 +157,9 @@ dependencies = [
 
 [[package]]
 name = "pyo3-build-config"
-version = "0.22.5"
+version = "0.22.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "bc38c5feeb496c8321091edf3d63e9a6829eab4b863b4a6a65f26f3e9cc6b179"
+checksum = "b14b5775b5ff446dd1056212d778012cbe8a0fbffd368029fd9e25b514479c38"
 dependencies = [
  "once_cell",
  "target-lexicon",
@@ -193,9 +167,9 @@ dependencies = [
 
 [[package]]
 name = "pyo3-ffi"
-version = "0.22.5"
+version = "0.22.6"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "94845622d88ae274d2729fcefc850e63d7a3ddff5e3ce11bd88486db9f1d357d"
+checksum = "9ab5bcf04a2cdcbb50c7d6105de943f543f9ed92af55818fd17b660390fc8636"
 dependencies = [
  "libc",
  "pyo3-build-config",
@@ -221,22 +195,11 @@ dependencies = [
  "paste",
 ]
 
-[[package]]
-name = "rmp-serde"
-version = "1.3.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "52e599a477cf9840e92f2cde9a7189e67b42c57532749bf90aea6ec10facd4db"
-dependencies = [
- "byteorder",
- "rmp",
- "serde",
-]
-
 [[package]]
 name = "serde"
-version = "1.0.210"
+version = "1.0.215"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c8e3592472072e6e22e0a54d5904d9febf8508f65fb8552499a1abc7d1078c3a"
+checksum = "6513c1ad0b11a9376da888e3e0baa0077f1aed55c17f50e7b2397136129fb88f"
 dependencies = [
  "serde_derive",
 ]
@@ -252,15 +215,21 @@ dependencies = [
 
 [[package]]
 name = "serde_derive"
-version = "1.0.210"
+version = "1.0.215"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "243902eda00fad750862fc144cea25caca5e20d615af0a81bee94ca738f1df1f"
+checksum = "ad1e866f866923f252f05c889987993144fb74e722403468a4ebd70c3cd756c0"
 dependencies = [
  "proc-macro2",
  "quote",
  "syn",
 ]
 
+[[package]]
+name = "simdutf8"
+version = "0.1.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e3a9fe34e3e7a50316060351f37187a3f546bce95496156754b601a5fa71b76e"
+
 [[package]]
 name = "smallvec"
 version = "1.13.2"
@@ -269,9 +238,9 @@ checksum = "3c5e1a9a646d36c3599cd173a41282daf47c44583ad367b8e6837255952e5c67"
 
 [[package]]
 name = "syn"
-version = "2.0.62"
+version = "2.0.82"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "9f660c3bfcefb88c538776b6685a0c472e3128b51e74d48793dc2a488196e8eb"
+checksum = "83540f837a8afc019423a8edb95b52a8effe46957ee402287f4292fae35be021"
 dependencies = [
  "proc-macro2",
  "quote",

{ormsgpack-1.6.0 → ormsgpack-1.7.0}/Cargo.toml

@@ -1,12 +1,13 @@
 [package]
 name = "ormsgpack"
-version = "1.6.0"
+version = "1.7.0"
 authors = [
     "Aviram Hassan <aviramyhassan@gmail.com>",
     "Emanuele Giaquinta <emanuele.giaquinta@gmail.com>",
 ]
 description = "Fast, correct Python msgpack library supporting dataclasses, datetimes, and numpy"
 edition = "2021"
+rust-version = "1.70"
 license = "Apache-2.0 OR MIT"
 repository = "https://github.com/aviramha/ormsgpack"
 homepage = "https://github.com/aviramha/ormsgpack"
@@ -33,23 +34,21 @@ default = ["unstable-simd"]
 # Use SIMD intrinsics. This requires Rust on the nightly channel.
 unstable-simd = [
     "bytecount/generic-simd",
-    "encoding_rs/simd-accel",
 ]
 
 [dependencies]
 ahash = { version = "0.8", default-features = false }
-associative-cache = { version = "2", default-features = false }
 bytecount = { version = "^0.6.7", default-features = false, features = ["runtime-dispatch-simd"] }
+byteorder = { version = "1.5.0", default-features = false, features = ["std"] }
 chrono = { version = "0.4.38", default-features = false }
-encoding_rs = { version = "0.8", default-features = false }
 half = { version = "2.4.1", default-features = false }
 itoa = { version = "1", default-features = false }
 once_cell = { version = "1", default-features = false, features = ["race"] }
-pyo3 = { version = "^0.22.5", default-features = false, features = ["extension-module"] }
+pyo3 = { version = "^0.22.6", default-features = false, features = ["extension-module"] }
 rmp = { version = "^0.8.14", default-features = false, features = ["std"] }
-rmp-serde = { version = "1", default-features = false }
 serde = { version = "1", default-features = false }
 serde_bytes = { version = "0.11.15", default-features = false, features = ["std"] }
+simdutf8 = { version = "0.1.5", default-features = false, features = ["std"] }
 smallvec = { version = "^1.13", default-features = false, features = ["union", "write"] }
 
 [build-dependencies]

{ormsgpack-1.6.0 → ormsgpack-1.7.0}/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.3
+Metadata-Version: 2.4
 Name: ormsgpack
-Version: 1.6.0
+Version: 1.7.0
 Classifier: Development Status :: 5 - Production/Stable
 Classifier: Intended Audience :: Developers
 Classifier: License :: OSI Approved :: Apache Software License
@@ -9,7 +9,6 @@ Classifier: Operating System :: MacOS
 Classifier: Operating System :: Microsoft :: Windows
 Classifier: Operating System :: POSIX :: Linux
 Classifier: Programming Language :: Python :: 3
-Classifier: Programming Language :: Python :: 3.8
 Classifier: Programming Language :: Python :: 3.9
 Classifier: Programming Language :: Python :: 3.10
 Classifier: Programming Language :: Python :: 3.11
@@ -27,7 +26,7 @@ Home-Page: https://github.com/aviramha/ormsgpack
 Author: Aviram Hassan <aviramyhassan@gmail.com>, Emanuele Giaquinta <emanuele.giaquinta@gmail.com>
 Author-email: Aviram Hassan <aviramyhassan@gmail.com>, Emanuele Giaquinta <emanuele.giaquinta@gmail.com>
 License: Apache-2.0 OR MIT
-Requires-Python: >=3.8
+Requires-Python: >=3.9
 Description-Content-Type: text/markdown; charset=UTF-8; variant=GFM
 Project-URL: Source Code, https://github.com/aviramha/ormsgpack
 
@@ -35,28 +34,13 @@ Project-URL: Source Code, https://github.com/aviramha/ormsgpack
 ![PyPI](https://img.shields.io/pypi/v/ormsgpack)
 ![PyPI - Downloads](https://img.shields.io/pypi/dm/ormsgpack)
 
-ormsgpack is a fast msgpack library for Python. It is a fork/reboot of [orjson](https://github.com/ijl/orjson)
-It serializes faster than [msgpack-python](https://github.com/msgpack/msgpack-python) and deserializes a bit slower (right now).
-It supports serialization of:
-[dataclass](#dataclass),
-[datetime](#datetime),
-[numpy](#numpy),
-[pydantic](#pydantic) and
-[UUID](#uuid) instances natively.
-
-Its features and drawbacks compared to other Python msgpack libraries:
-
-* serializes `dataclass` instances natively.
-* serializes `datetime`, `date`, and `time` instances to RFC 3339 format,
-e.g., "1970-01-01T00:00:00+00:00"
-* serializes `numpy.ndarray` instances natively and faster.
-* serializes `pydantic.BaseModel` instances natively
-* serializes arbitrary types using a `default` hook
-
-ormsgpack supports CPython 3.8, 3.9, 3.10, 3.11, 3.12 and 3.13.
-ormsgpack does not support PyPy. Releases follow semantic
-versioning and serializing a new object type without an opt-in flag is
-considered a breaking change.
+ormsgpack is a fast msgpack serialization library for Python derived
+from [orjson](https://github.com/ijl/orjson), with native support for
+various Python types.
+
+ormsgpack supports CPython 3.9, 3.10, 3.11, 3.12 and 3.13. Releases
+follow semantic versioning and serializing a new object type without
+an opt-in flag is considered a breaking change.
 
 ormsgpack is licensed under both the Apache 2.0 and MIT licenses. The
 repository and issue tracker is
@@ -139,7 +123,7 @@ serializes subclasses of `str`, `int`, `dict`, `list`,
 of `tuple` to avoid serializing `namedtuple` objects as arrays. To avoid
 serializing subclasses, specify the option `ormsgpack.OPT_PASSTHROUGH_SUBCLASS`.
 
-The output is a `bytes` object containing UTF-8.
+The output is a `bytes` object.
 
 The global interpreter lock (GIL) is held for the duration of the call.
 
@@ -443,8 +427,8 @@ Serialize `pydantic.BaseModel` instances.
 
 ##### OPT_SORT_KEYS
 
-Serialize `dict` keys in sorted order. The default is to serialize in an
-unspecified order.
+Serialize `dict` keys and pydantic model fields in sorted order. The default
+is to serialize in an unspecified order.
 
 This can be used to ensure the order is deterministic for hashing or tests.
 It has a substantial performance penalty and is not recommended in general.
@@ -797,7 +781,11 @@ b'\xd9$886313e1-3b8a-5372-9b90-0c9aee199e5d'
 ```
 
 ### Pydantic
-ormsgpack serializes `pydantic.BaseModel` instances natively.
+ormsgpack serializes `pydantic.BaseModel` instances natively, with
+[duck-typing](https://docs.pydantic.dev/2.10/concepts/serialization/#serializing-with-duck-typing).
+This is equivalent to serializing
+`model.model_dump(serialize_as_any=True)` with Pydantic V2 or
+`model.dict()`with Pydantic V1.
 
 #### Performance
 ![alt text](doc/pydantic.svg "pydantic")
@@ -906,7 +894,7 @@ level above this.
 
 ## Packaging
 
-To package ormsgpack requires [Rust](https://www.rust-lang.org/) 1.65
+To package ormsgpack requires [Rust](https://www.rust-lang.org/) 1.70
 or newer and the [maturin](https://github.com/PyO3/maturin) build
 tool. The default feature `unstable-simd` enables the usage of SIMD
 operations and requires nightly Rust. The recommended build command

{ormsgpack-1.6.0 → ormsgpack-1.7.0}/README.md

@@ -2,28 +2,13 @@
 ![PyPI](https://img.shields.io/pypi/v/ormsgpack)
 ![PyPI - Downloads](https://img.shields.io/pypi/dm/ormsgpack)
 
-ormsgpack is a fast msgpack library for Python. It is a fork/reboot of [orjson](https://github.com/ijl/orjson)
-It serializes faster than [msgpack-python](https://github.com/msgpack/msgpack-python) and deserializes a bit slower (right now).
-It supports serialization of:
-[dataclass](#dataclass),
-[datetime](#datetime),
-[numpy](#numpy),
-[pydantic](#pydantic) and
-[UUID](#uuid) instances natively.
-
-Its features and drawbacks compared to other Python msgpack libraries:
-
-* serializes `dataclass` instances natively.
-* serializes `datetime`, `date`, and `time` instances to RFC 3339 format,
-e.g., "1970-01-01T00:00:00+00:00"
-* serializes `numpy.ndarray` instances natively and faster.
-* serializes `pydantic.BaseModel` instances natively
-* serializes arbitrary types using a `default` hook
-
-ormsgpack supports CPython 3.8, 3.9, 3.10, 3.11, 3.12 and 3.13.
-ormsgpack does not support PyPy. Releases follow semantic
-versioning and serializing a new object type without an opt-in flag is
-considered a breaking change.
+ormsgpack is a fast msgpack serialization library for Python derived
+from [orjson](https://github.com/ijl/orjson), with native support for
+various Python types.
+
+ormsgpack supports CPython 3.9, 3.10, 3.11, 3.12 and 3.13. Releases
+follow semantic versioning and serializing a new object type without
+an opt-in flag is considered a breaking change.
 
 ormsgpack is licensed under both the Apache 2.0 and MIT licenses. The
 repository and issue tracker is
@@ -106,7 +91,7 @@ serializes subclasses of `str`, `int`, `dict`, `list`,
 of `tuple` to avoid serializing `namedtuple` objects as arrays. To avoid
 serializing subclasses, specify the option `ormsgpack.OPT_PASSTHROUGH_SUBCLASS`.
 
-The output is a `bytes` object containing UTF-8.
+The output is a `bytes` object.
 
 The global interpreter lock (GIL) is held for the duration of the call.
 
@@ -410,8 +395,8 @@ Serialize `pydantic.BaseModel` instances.
 
 ##### OPT_SORT_KEYS
 
-Serialize `dict` keys in sorted order. The default is to serialize in an
-unspecified order.
+Serialize `dict` keys and pydantic model fields in sorted order. The default
+is to serialize in an unspecified order.
 
 This can be used to ensure the order is deterministic for hashing or tests.
 It has a substantial performance penalty and is not recommended in general.
@@ -764,7 +749,11 @@ b'\xd9$886313e1-3b8a-5372-9b90-0c9aee199e5d'
 ```
 
 ### Pydantic
-ormsgpack serializes `pydantic.BaseModel` instances natively.
+ormsgpack serializes `pydantic.BaseModel` instances natively, with
+[duck-typing](https://docs.pydantic.dev/2.10/concepts/serialization/#serializing-with-duck-typing).
+This is equivalent to serializing
+`model.model_dump(serialize_as_any=True)` with Pydantic V2 or
+`model.dict()`with Pydantic V1.
 
 #### Performance
 ![alt text](doc/pydantic.svg "pydantic")
@@ -873,7 +862,7 @@ level above this.
 
 ## Packaging
 
-To package ormsgpack requires [Rust](https://www.rust-lang.org/) 1.65
+To package ormsgpack requires [Rust](https://www.rust-lang.org/) 1.70
 or newer and the [maturin](https://github.com/PyO3/maturin) build
 tool. The default feature `unstable-simd` enables the usage of SIMD
 operations and requires nightly Rust. The recommended build command

{ormsgpack-1.6.0 → ormsgpack-1.7.0}/pyproject.toml

@@ -1,7 +1,7 @@
 [project]
 name = "ormsgpack"
 repository = "https://github.com/aviramha/ormsgpack"
-requires-python = ">=3.8"
+requires-python = ">=3.9"
 classifiers = [
     "Development Status :: 5 - Production/Stable",
     "Intended Audience :: Developers",
@@ -11,7 +11,6 @@ classifiers = [
     "Operating System :: Microsoft :: Windows",
     "Operating System :: POSIX :: Linux",
     "Programming Language :: Python :: 3",
-    "Programming Language :: Python :: 3.8",
     "Programming Language :: Python :: 3.9",
     "Programming Language :: Python :: 3.10",
     "Programming Language :: Python :: 3.11",
@@ -41,7 +40,7 @@ ignore_missing_imports = true
 
 [tool.ruff]
 line-length = 88
-target-version = "py38"
+target-version = "py39"
 
 [tool.ruff.lint]
 select = [
@@ -52,6 +51,8 @@ select = [
 ]
 ignore = [
   "E501",
+  "RUF022",
+  "RUF023",
 ]
 
 [tool.ruff.lint.isort]

ormsgpack-1.7.0/src/deserialize/cache.rs

@@ -0,0 +1,72 @@
+// SPDX-License-Identifier: (Apache-2.0 OR MIT)
+
+use crate::typeref::*;
+use crate::unicode::*;
+use once_cell::unsync::OnceCell;
+use simdutf8::basic::{from_utf8, Utf8Error};
+use std::hash::BuildHasher;
+use std::hash::Hasher;
+use std::ptr::NonNull;
+
+#[repr(transparent)]
+struct CachedKey {
+    ptr: *mut pyo3::ffi::PyObject,
+}
+
+unsafe impl Send for CachedKey {}
+unsafe impl Sync for CachedKey {}
+
+impl CachedKey {
+    fn new(ptr: *mut pyo3::ffi::PyObject) -> CachedKey {
+        CachedKey { ptr: ptr }
+    }
+
+    fn get(&mut self) -> *mut pyo3::ffi::PyObject {
+        ffi!(Py_INCREF(self.ptr));
+        self.ptr
+    }
+}
+
+impl Drop for CachedKey {
+    fn drop(&mut self) {
+        ffi!(Py_DECREF(self.ptr));
+    }
+}
+
+pub struct KeyMap<const C: usize> {
+    entries: Vec<Option<CachedKey>>,
+}
+
+impl<const C: usize> KeyMap<C> {
+    pub fn new() -> Self {
+        let mut entries = Vec::with_capacity(C);
+        for _ in 0..C {
+            entries.push(None);
+        }
+        KeyMap { entries: entries }
+    }
+
+    pub fn get(&mut self, key: &[u8]) -> Result<NonNull<pyo3::ffi::PyObject>, Utf8Error> {
+        let mut hasher = unsafe { HASH_BUILDER.get().unwrap().build_hasher() };
+        let hash = {
+            hasher.write(key);
+            hasher.finish()
+        } as usize;
+        let index = hash % C;
+        let entry = match &mut self.entries[index] {
+            Some(v) if unicode_to_str(v.ptr).unwrap().as_bytes() == key => v,
+            _ => {
+                let pykey = unicode_from_str(from_utf8(key)?);
+                hash_str(pykey);
+                self.entries[index] = Some(CachedKey::new(pykey));
+                match &mut self.entries[index] {
+                    Some(v) => v,
+                    _ => unreachable!(),
+                }
+            }
+        };
+        Ok(nonnull!(entry.get()))
+    }
+}
+
+pub static mut KEY_MAP: OnceCell<KeyMap<512>> = OnceCell::new();