NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
PyPI
PyPI, #2517 most downloaded on PyPI
Common interface for data container classes
Latest stable
0.14.0
Last release 24 days ago
29 Oct 2026
Release timing varies
gaps range from 2 weeks to 1.1 years
Nearly every release is documented
notes for 26 of 26 stable releases
Nothing withdrawn
no release was ever pulled
7 years old
26 releases · first in 2020
Drag across the ruler to compare, or tap a release to open it.
Bump version: 0.13.1 → 0.14.0
Bump version: 0.13.1 → 0.14.0
Removed support for Python 3.9 and 3.10, and added support for Python 3.15 (#112, #130, #137).
Raised the minimum supported versions of optional dependencies: attrs 22.1.0
(was 20.1.0), pydantic 2.7 (was 1.8) and scrapy 2.7 (was 2.2)
(#137,
#139,
#140).
Removed support for Pydantic 1 models, including pydantic.v1 models from
Pydantic 2, and for scrapy.item.BaseItem subclasses that are not
scrapy.Item subclasses
(#139).
ItemAdapter.ADAPTER_CLASSES is now a tuple instead of a
collections.deque, so it no longer supports appendleft() and other
in-place changes. To register an adapter class, reassign it instead, e.g.
ItemAdapter.ADAPTER_CLASSES = (MyAdapter, *ItemAdapter.ADAPTER_CLASSES)
(#118,
#128).
Which adapter class handles each item class is now cached, making
ItemAdapter() and is_item() faster. As a result, the is_item_class()
class method of adapter classes must return the same result every time it is
called with the same item class
(#128).
Overriding the is_item() class method in an adapter class is deprecated,
override is_item_class() instead. Adapter classes that override is_item()
still work, but trigger a DeprecationWarning and disable the cache above
(#128).
Added support for
TypedDict
item classes, including typing_extensions.TypedDict ones: field names, field
metadata from the first mapping in
Annotated
type hints, and get_json_schema() support, which honors Required and
NotRequired (#123).
get_json_schema() is much faster on repeated calls, as field docstrings read
from source code are now cached
(#141).
get_json_schema() no longer fails on unions of 2 or more non-simple types,
e.g. 2 item classes (#123).
get_json_schema() now maps int | float to number regardless of the order
of the union members; float | int already did
(#133).
get_json_schema() now handles attrs.validators.matches_re() validators
better: patterns are anchored to match their match function (re.fullmatch()
by default), and patterns with flags that JSON Schema cannot express, e.g.
re.IGNORECASE, are left out
(#122).
Improved type hints (#138).
Your coding agent can read these notes before it upgrades. Set up the MCP server →
Bump version: 0.13.0 → 0.13.1
Bump version: 0.13.0 → 0.13.1
Fixed get_json_schema() to properly format multiline field docstrings using
inspect.cleandoc(), ensuring consistent indentation and whitespace handling
(#115).
Bump version: 0.12.2 → 0.13.0
Bump version: 0.12.2 → 0.13.0
Added support for Python 3.14 and removed support for PyPy 3.10.
get_json_schema() no longer reports all scrapy.Item fields as required.
Bump version: 0.12.1 → 0.12.2
Bump version: 0.12.1 → 0.12.2
The return value of get_json_schema() is now deterministic (deterministic
order of dict keys and list items).
Bump version: 0.12.0 → 0.12.1
Bump version: 0.12.0 → 0.12.1
get_json_schema() now supports inherited field docstrings.
Added support for PyPy 3.11 (https://github.com/scrapy/itemadapter/pull/97).
Added support for PyPy 3.11 (https://github.com/scrapy/itemadapter/pull/97).
Added a new get_json_schema() class method to ItemAdapter and all built-in adapters to output a JSON Schema for a given item class (https://github.com/scrapy/itemadapter/pull/101).
Modernized the code base, now making full use of pyproject.toml and using ruff (https://github.com/scrapy/itemadapter/pull/96, https://github.com/scrapy/itemadapter/pull/98).
Added support for PyPy 3.11 (#97).
Added a new get_json_schema() class method to ItemAdapter and all built-in
adapters to output a JSON Schema for a given item
class (#101).
Modernized the code base, now making full use of pyproject.toml and using ruff (#96, #98).
Removed functions deprecated in 0.5.0:
Removed functions deprecated in 0.5.0:
itemadapter.utils.is_attrs_instance()itemadapter.utils.is_dataclass_instance()itemadapter.utils.is_pydantic_instance()itemadapter.utils.is_scrapy_item()(#93).
Dropped Python 3.8 support, added official Python 3.13 and PyPy 3.10 support (#79, #87).
Dropped Python 3.7 support, added official Python 3.12 support (#75, #77).
ItemAdapter.ADAPTER_CLASSES to say that subclasses can use any iterable, not just collections.deque (#74).Pydantic >= 2 is not supported yet (#73).Add support for Python 3.11. by @wRAR in https://github.com/scrapy/itemadapter/pull/65
Full Changelog: https://github.com/scrapy/itemadapter/compare/v0.7.0...v0.8.0
Dropped Python 3.6 support, and made Python 3.11 support official (#65, #66, #69).
It is now possible to declare custom ItemAdapter subclasses with their own
ADAPTER_CLASSES attribute, allowing to support different item types in
different parts of the same code base
(#68).
Improved type hint support (#67).
ItemAdapter.get_field_names_from_class by @kmike in https://github.com/scrapy/itemadapter/pull/64
Full Changelog: https://github.com/scrapy/itemadapter/compare/v0.6.0...v0.7.0
Improve performance by @elacuesta in https://github.com/scrapy/itemadapter/pull/62
Full Changelog: https://github.com/scrapy/itemadapter/compare/v0.5.0...v0.6.0
Deprecate undocumented predicates from the utils module by @elacuesta in https://github.com/scrapy/itemadapter/pull/56
utils module by @elacuesta in https://github.com/scrapy/itemadapter/pull/56Full Changelog: https://github.com/scrapy/itemadapter/compare/v0.4.0...v0.5.0
Improve performance by removing imports inside functions (#60)
Added ItemAdapter.is_item_class and ItemAdapter.get_field_meta_from_class class methods
Added ItemAdapter.is_item_class and ItemAdapter.get_field_meta_from_class class methods (#54)
Added support for `pydantic` models
Highlights:
pydantic models (#53)Added the ability to extend the library to support arbitrary types, by implementing a MutableMapping-based interface (as a consequence of this change,
MutableMapping-based interface (as a consequence of this change, now any type can be used as a Scrapy item! :tada:)Adapter interface: added the ability to support arbitrary types, by implementing a MutableMapping-based interface. By way of this change, now any type can be used as a Scrapy item.
Dropped support for Python 3.5 (#38).
Dropped support for Python 3.5 (#38).
The new get_field_meta_from_class function offers the same functionality as
ItemAdapter.get_field_meta but for an item class, as opposed to an item
object (#34, #35).
ItemAdapter.__repr__ no longer raises exceptions caused by the underlying
item (#31, #41).
Minor improvement to the release process (#37), and cleanup of test warnings (#40).
Added ItemAdapter.asdict, which allows converting an item and all of its nested items into dict objects (#27, #29).
Added ItemAdapter.asdict, which allows converting an item and all of its
nested items into dict objects (#27, #29).
Improved ItemAdapter performance by reducing time complexity for lookups and
traversals for dataclass and attrs items (#28).
ItemAdapter.field_names now returns a KeysView instead of a list.
ItemAdapter.field_names now returns a KeysView instead of a list.
Minor CI and test changes.
ItemAdapter.get_field_meta now returns an empty MappingProxyType object for items without metadata support, instead of raising TypeError.
ItemAdapter.get_field_meta now returns an empty MappingProxyType object for
items without metadata support, instead of raising TypeError.
Improved the README and some docstrings.
Provided full test coverage, and refactored CI configuration, test configuration and tests themselves.
Added support for Scrapy’s BaseItem.
Added support for Scrapy’s BaseItem.
Refactored and extended tests.
Code style and documentation fixes.
Removed support for MutableMapping.
Removed support for MutableMapping.
Removed metadata support for arbitrary mutable mappings.
Removed metadata support for arbitrary mutable mappings.
Rebuild for the Python Package Index.
Rebuild for the Python Package Index.
Split the implementation into several files for better code organization, and without an impact on the existing API import paths.
Split the implementation into several files for better code organization, and without an impact on the existing API import paths.
Also improved the README.
Initial release.
Initial release.
One column per quarter.