Non-XML dataclass/ignored fields (#2)

Author: tobywf

README.md

@@ -20,6 +20,7 @@ Requires Python 3.7 or higher.
 * Inheritance does work, but has the same limitations as dataclasses. Inheriting from base classes with required fields and declaring optional fields doesn't work due to field order. This isn't recommended
 * Namespace support is decent as long as correctly declared. I've tried on several real-world examples, although they were known to be valid. `lxml` does a great job at expanding namespace information when loading and simplifying it when saving
 * Post-load validation hook `xml_validate`
+* Fields not required in the constructor are ignored by this library (via `ignored()` or `init=False`)
 
 ## Patterns
 
@@ -84,6 +85,14 @@ def xml_validate(self) -> None:
 
 If defined, the `load` function will call it after all values have been loaded and assigned to the XML dataclass. You can validate the fields you want inside this method. Return values are ignored; instead raise and catch exceptions.
 
+### Ignored fields
+
+Fields not required in the constructor are ignored by this library (new in version 0.0.6). This is useful if you want to populate a field via post-load validation.
+
+You can simply set `init=False`, although you may also want to exclude the field from comparisons. The `ignored` function does this, and can also be used.
+
+The name doesn't matter, but it might be useful to use the `_` prefix as a convention.
+
 ## Example (fully type hinted)
 
 (This is a simplified real world example - the container can also include optional `links` child elements.)
@@ -192,10 +201,27 @@ This makes sense in many cases, but possibly not every case.
 Most of these limitations/assumptions are enforced. They may make this project unsuitable for your use-case.
 
 * If you need to pass any parameters to the wrapped `@dataclass` decorator, apply it before the `@xml_dataclass` decorator
-* Setting the `init` parameter of a dataclass' `field` will lead to bad things happening, this isn't supported.
 * Deserialisation is strict; missing required attributes and child elements will cause an error. I want this to be the default behaviour, but it should be straightforward to add a parameter to `load` for lenient operation
 * Dataclasses must be written by hand, no tools are provided to generate these from, DTDs, XML schema definitions, or RELAX NG schemas
 
+## Changelog
+
+### [0.0.6] - 2020-03-25
+
+* Allow ignored fields via `init=false` or the `ignored` function
+
+### [0.0.5] - 2020-02-18
+
+* Fixed type hinting for consumers. While the library passed mypy validation, it was hard to get XML dataclasses in a codebase to pass mypy validation
+
+### [0.0.4] - 2020-02-16
+
+* Improved type resolving. This lead to easier field definitions, as `attr` and `child` are no longer needed because the type of the field is inferred
+
+### [0.0.3] - 2020-02-16
+
+* Added support for union types on children
+
 ## Development
 
 This project uses [pre-commit](https://pre-commit.com/) to run some linting hooks when committing. When you first clone the repo, please run:

functional/container_test.py

@@ -5,7 +5,15 @@
 import pytest  # type: ignore
 from lxml import etree  # type: ignore
 
-from xml_dataclasses import NsMap, XmlDataclass, dump, load, rename, xml_dataclass
+from xml_dataclasses import (
+    NsMap,
+    XmlDataclass,
+    dump,
+    ignored,
+    load,
+    rename,
+    xml_dataclass,
+)
 
 from .utils import lmxl_dump
 
@@ -36,10 +44,12 @@ class Container(XmlDataclass):
     version: str
     rootfiles: RootFiles
     # WARNING: this is an incomplete implementation of an OPF container
+    _version: int = ignored()
 
     def xml_validate(self) -> None:
         if self.version != "1.0":
             raise ValueError(f"Unknown container version '{self.version}'")
+        self._version = 1
 
 
 @pytest.mark.parametrize("remove_blank_text", [True, False])

pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "xml_dataclasses"
-version = "0.0.5"
+version = "0.0.6"
 description = "(De)serialize XML documents into specially-annotated dataclasses"
 authors = ["Toby Fleming <[email protected]>"]
 license = "MPL-2.0"

src/xml_dataclasses/__init__.py

@@ -2,7 +2,7 @@
 
 logging.getLogger(__name__).addHandler(logging.NullHandler())
 
-from .modifiers import rename, text  # isort:skip
+from .modifiers import rename, text, ignored  # isort:skip
 from .resolve_types import (  # isort:skip
     is_xml_dataclass,
     xml_dataclass,
@@ -23,4 +23,5 @@
     "xml_dataclass",
     "NsMap",
     "XmlDataclass",
+    "ignored",
 ]

src/xml_dataclasses/modifiers.py

@@ -46,3 +46,10 @@ def text(
     metadata["xml:text"] = True
     f.metadata = metadata
     return f  # type: ignore
+
+
+# NOTE: Actual return type is 'Field[_T]', but we want to help type checkers
+# to understand the magic that happens at runtime.
+# see https://github.com/python/typeshed/blob/master/stdlib/3.7/dataclasses.pyi
+def ignored() -> _T:
+    return field(init=False, compare=False)  # type: ignore

src/xml_dataclasses/resolve_types.py

@@ -257,6 +257,10 @@ def xml_dataclass(cls: Type[Any]) -> Type[XmlDataclassInstance]:
     children: List[ChildInfo] = []
     text_field = None
     for f in fields(cls):
+        # ignore fields not required in the constructor
+        if not f.init:
+            continue
+
         field_info = _resolve_field_type(f)
         if isinstance(field_info, TextInfo):
             if text_field is not None:

tests/modifiers_test.py

@@ -5,7 +5,7 @@
 
 import pytest
 
-from xml_dataclasses.modifiers import rename, text
+from xml_dataclasses.modifiers import ignored, rename, text
 
 
 def dict_comb(items, r=2):
@@ -82,3 +82,9 @@ def test_text_has_field_default_ignored(default):
     assert actual_field is expected_field
     assert actual_field.default is MISSING
     assert actual_field.metadata == expected_md
+
+
+def test_ignored_field():
+    actual_field = ignored()
+    assert not actual_field.init
+    assert not actual_field.compare

tests/resolve_types_resolve_test.py

@@ -1,4 +1,4 @@
-from dataclasses import dataclass
+from dataclasses import dataclass, field, fields
 from itertools import product
 from typing import List, Optional, Union, _GenericAlias
 
@@ -118,3 +118,20 @@ class Foo:
     # resolve_types_resolve_test.XmlDt1]], NoneType]
     assert set(bar.base_types) == types
     assert bar.is_list is is_list
+
+
+def test_non_ctor_field_is_ignored():
+    @xml_dataclass
+    class Foo:
+        __ns__ = None
+        bar: str = field(init=False)
+
+    assert not Foo.__attributes__
+    assert not Foo.__text_field__
+    assert not Foo.__children__
+
+    dt_fields = fields(Foo)
+    assert len(dt_fields) == 1
+    dt_field = dt_fields[0]
+    assert dt_field.name == "bar"
+    assert not dt_field.init