Changelog

This page has changelogs for all releases of Oblate.

v1.2.1

• Fix missing oblate.fields package during installation of v1.2.0.

v1.2.0

New features

• Add support for frozen schemas and fields.

• Add FieldNotSet exception to be raised on accessing fields without values.

• Add Schema.copy() method for copying schema instances.

• Add Schema.preprocess_data() for preprocessing of input data.

• Add oblate.validate_types() function for type validation.

Improvements

• All fields now have relevant attributes to access values passed to constructor.

• Field accessing now raises a special exception.

• Schema constructor now raises TypeError if non-mapping type is passed as data.

Bug fixes

• Fix SyntaxError in various Union and Literal fields in older Python versions than 3.11.

v1.1.0

New features

• Add fields.List for accepting list structure.

• Add fields.Set for accepting set structures.

• Add fields.Dict and fields.TypedDict for handling dictionaries.

• Add fields.TypeExpr for validating types using simple type expressions.

• Add fields.Any field as a no validation field.

• Add fields.Literal field to accept pre-defined literal values.

• Add fields.Union for basic union type validation.

• Add validate.Range validator for validating integer ranges.

• Add validate.Length validator for validating length of sized types.

• Add validate.Regex validator for validating values using regular expression.

• Add validate.Exclude validator for disallowing/excluding specific values.

• Add validate.Or validator for OR’ing the result of multiple validators.

• Add Schema.__schema_post_init__() method as a post initialization hook.

• Add ErrorContext.metadata for storing extra error information.

• Add GlobalConfig.warn_unsupported_types to control behaviour on using unsupported types.

• Add Field.name and Field.schema properties.

• Add init_kwargs parameter in fields.Object to support passing initialization arguments.

Improvements

• fields.Object field now supports passing Schema instances directly instead of raw data.

• fields.Field.format_error() no longer requires super call and default error messages are resolved automatically.

v1.0.0

This version brings an entire rewrite of the library. Many features have been removed along with many additions and breaking changes. 0.2.0 and former are not compatible with this version.

For an overview of changes, see this pull request.

v0.2.0

New features

• Add support for partial schemas

  • Add fields.Partial field.

  • Add Schema.is_partial() helper method.

• Add ValidationError.is_field_error() helper to check whether exception relates to a field.

• Add Schema.after_init_hook() method that gets called when schema is finished initializing.

• Add load_key and dump_key parameters to fields.Field to control the field names during serialization/deserialization of data.

• Add ValidationError.state to allow storing state to an error that can later be accessed.

• Add include and exclude parameters to Schema.dump().

• Add raw parameter to Field.validate() and Field.add_validator() to validate raw value.

Improvements

• Object fields now support raw data serialization upon setting after initialization.

• Schema and all fields now have __slots__ defined for performance sake.

• User defined validators are now ran after field value_* methods by default. Such that, validators now take serialized value instead of raw one. For previous behaviour, raw parameter must be used.

Bug fixes

• Fix Schema.dump() returning empty dictionary.

• Fix fields returning improper values when initializing multiple schemas.

• Fix unable to set None on nullable fields after initialization.

• Fix field getting the new value using the setter even when the validations fail.

Documentation

• Add “Tutorials” section

v0.1.0

• Initial release