Serpyco-rs

Bug Fix:

**Memory Leak Fix**: We have resolved a memory leak issue occurring during the loading and dumping of objects of type `dict`. This fix ensures that memory is properly managed, preventing any unnecessary memory consumption during these operations.

**Full Changelog**: https://github.com/ermakov-oleg/serpyco-rs/compare/v1.1.2...v1.1.3

Library Updates:

* Bump speedate from 0.12.0 to 0.13.0
* Bump dyn-clone from 1.0.14 to 1.0.16

Platform Support:

* **Expanded wheel builds:** We are excited to introduce expanded wheel builds for a variety of platforms. With this release, we now offer wheel builds for Linux on the following architectures:
- aarch64
- armv7
- s390x
- ppc64le
- ppc64

These wheel builds aim to provide broader compatibility and accessibility for users on diverse hardware architectures.

Bug Fix:

**Resolved segfault when loading int/float as Decimal**: We have fixed a critical issue that was causing a segmentation fault (segfault) when attempting to load integer or float values as Decimal.

Enhancements:

Support for frozen dataclasses/attrs: With this release, we've added support for frozen dataclasses and attrs. You can now use our library seamlessly with these data structures.

Bug Fixes:

Fix for unexpected KeyError: We've addressed an issue where an unexpected KeyError occurred when validating a tagged union. This problem has been resolved, ensuring smoother validation processes.

We are thrilled to announce the release of version 1.0.1 of our library! This update brings notable optimizations and performance enhancements.

Enhancements:

* **Reduced unnecessary memory allocation:** We've eliminated unnecessary memory allocations when invoking the `load` method for data classes. This optimization results in a speed boost of approximately 10%.

* **Preallocated enum values:** We now use preallocated enum values when calling the `load` method. This change significantly accelerates the process, with a remarkable speed improvement of approximately 330%.

We are excited to announce the release of version 1.0.0 of our library! In this update, we have introduced significant changes and improvements. Please take a moment to review the following release notes.

Enhancements:

* **Improved validation mechanism:** We have replaced the use of `jsonschema` for data validation with custom validators. Previously, data had to be converted into Rust structures for validation, but now validation is performed directly on Python objects. This change has resulted in a substantial performance boost for the `Serializer.load` method, [improving its speed by 25% to 350%](https://github.com/ermakov-oleg/serpyco-rs/pull/104#issuecomment-1768784012).

Breaking Changes:

* **Removal of `pass_through_bytes` argument:** The `pass_through_bytes` argument in the `Serializer` has been removed. Bytes are now handled in the same way as other data types.

* **Elimination of `validate` parameter:** The `validate` parameter in the `Serializer.load` method has been removed, as validation is now always enforced.

* **Deletion of `Serializer.load_json` method:** The `Serializer.load_json` method has been removed. We recommend using `orjson` for JSON deserialization.

* **Removal of `schema_path` attribute:** The `schema_path` attribute has been removed from `ErrorItem`.

* **Change in validation behavior:** Validation will no longer return all errors encountered but only the first encountered error.

* **Correction of typo:** An error in the class name `metadata.FieldFormat` has been fixed.


**Full Changelog**: https://github.com/ermakov-oleg/serpyco-rs/compare/v0.15.6...v1.0.0