Validation using dataclasses

.github/workflows/ci.yml

@@ -13,7 +13,7 @@ jobs:
     runs-on: ${{ matrix.os }}
     strategy:
       matrix:
-        python-version: [3.7, 3.8, 3.9, "3.10", 3.11-dev]
+        python-version: [3.8, 3.9, "3.10", "3.11", "3.12"]
         os: [ubuntu-latest]
 
     steps:

docs/nested.rst

@@ -12,46 +12,45 @@ Pipfile and Pipfile.lock, such as individual entries in ``[packages]``,
 ``[dev-packages]``, and ``lockfile['_meta']``.
 
 
-The Data View
-=============
+Base Model 
+===========
 
 Every non-scalar value you get from Plette (e.g. sequence, mapping) is
-represented as a `DataView`, or one of its subclasses. This class is simply a
-wrapper around the basic collection class, and you can access the underlying
-data structure via the ``_data`` attribute::
+represented inherits from `models.BaseModel`, which is a Python `dataclass`::
 
     >>> import plette.models
-    >>> source = plette.models.Source({
+    >>> source = plette.models.Source(**{
     ...     'name': 'pypi',
     ...     'url': 'https://pypi.org/simple',
     ...     'verify_ssl': True,
     ... })
     ...
-    >>> source._data
-    {'name': 'pypi', 'url': 'https://pypi.org/simple', 'verify_ssl': True}
-
-
-Data View Collections
-=====================
-
-There are two special collection classes, ``DataViewMapping`` and
-``DataViewSequence``, that hold homogeneous ``DataView`` members. They are
-also simply wrappers to ``dict`` and ``list``, respectively, but have specially
-implemented magic methods to automatically coerce contained data into a
-``DataView`` subclass::
-
-    >>> sources = plette.models.SourceCollection([source._data])
-    >>> sources._data
-    [{'name': 'pypi', 'url': 'https://pypi.org/simple', 'verify_ssl': True}]
-    >>> type(sources[0])
-    <class 'plette.models.sources.Source'>
-    >>> sources[0] == source
-    True
-    >>> sources[0] = {
-    ...     'name': 'devpi',
-    ...     'url': 'http://localhost/simple',
-    ...     'verify_ssl': True,
-    ... }
-    ...
-    >>> sources._data
-    [{'name': 'devpi', 'url': 'http://localhost/simple', 'verify_ssl': True}]
+    >>> source
+    Source(name='pypi', verify_ssl=True, url='https://pypi.org/simple')
+
+
+Collections
+===========
+
+There a few special collection classes, which can be I dentified by the
+suffix ``Collection`` or ``Specifiers``.
+They group attributes and behave like ``list`` or ``mappings``.
+These classes accept a list of dictionaries as input,
+and convert them to the correct object type::
+
+    >>> SourceCollection([{'name': 'r-pi', 'url': '192.168.1.129:8000', 'verify_ssl': False}, {'name': 'pypi', 'url': 'https://pypi.org/simple', 'verify_ssl': True}])
+    SourceCollection(sources=[Source(name='r-pi', verify_ssl=False, url='192.168.1.129:8000'), Source(name='pypi', verify_ssl=True, url='https://pypi.org/simple')])
+
+
+In addition, they can also accept a list of items of the correct type::
+
+    >>> rpi = models.Source(**{'name': 'r-pi', 'url': '192.168.1.129:8000', 'verify_ssl': False})
+    >>> pypi = models.Source(**{'name': 'pypi', 'url': 'https://pypi.org/simple', 'verify_ssl': True})
+    >>> SourceCollection([rpi, pypi])
+    SourceCollection(sources=[Source(name='r-pi', verify_ssl=False, url='192.168.1.129:8000'), Source(name='pypi', verify_ssl=True, url='https://pypi.org/simple')])
+
+They can also be indexed by name, and can be iterated over::
+
+    >>> sc = SourceCollection([{'name': 'r-pi', 'url': '192.168.1.129:8000', 'verify_ssl': False}, {'name': 'pypi', 'url': 'https://pypi.org/simple', 'verify_ssl': True}])
+    >>> sc[0]
+    Source(name='r-pi', verify_ssl=False, url='192.168.1.129:8000')

docs/validation.rst

@@ -2,56 +2,4 @@
 Validating Data
 ===============
 
-Plette provides optional validation for input data. This chapter discusses
-how validation works.
-
-
-Setting up Validation
-=====================
-
-Validation is provided by the Cerberus_ library. You can install it along with
-Plette manually, or by specifying the “validation” extra when installing
-Plette:
-
-.. code-block:: none
-
-    pip install plette[validation]
-
-Plette automatically enables validation when Cerberus is available.
-
-.. _Cerberus: http://docs.python-cerberus.org/
-
-
-Validating Data
-===============
-
-Data is validated on input (or when a model is loaded). ``ValidationError`` is
-raised when validation fails::
-
-    >>> plette.models.Source({})
-    Traceback (most recent call last):
-      File "<stdin>", line 1, in <module>
-      File "plette/models/base.py", line 37, in __init__
-        self.validate(data)
-      File "plette/models/base.py", line 67, in validate
-        return validate(cls, data)
-      File "plette/models/base.py", line 27, in validate
-        raise ValidationError(data, v)
-    plette.models.base.ValidationError: {}
-
-This exception class has a ``validator`` member to allow you to access the
-underlying Cerberus validator, so you can know what exactly went wrong::
-
-    >>> try:
-    ...     plette.models.Source({'verify_ssl': True})
-    ... except plette.models.ValidationError as e:
-    ...     for error in e.validator._errors:
-    ...         print(error.schema_path)
-    ...
-    ('name', 'required')
-    ('url', 'required')
-
-See `Ceberus’s error handling documentation`_ to know how the errors are
-represented and reported.
-
-.. _`Ceberus’s error handling documentation`: http://docs.python-cerberus.org/en/stable/errors.html
+Validation is no longer an optional feature. All data models are validated.

examples/Pipfile.invalid.extras-list

@@ -1,4 +1,4 @@
-# package specifiers should be a dict
+# package extras should be a list
 [packages]
 msal = {version= "==1.20.0", extras = "broker"}
 

examples/Pipfile.invalid.with-categories

@@ -1,3 +1,5 @@
+# this file contains a section with one package with wrong extras
+# extras should be passed as a list
 [packages]
 plette = { path = '.', extras = ['validation'], editable = true }
 

examples/extras-list/Pipfile

@@ -0,0 +1,6 @@
+
+# extras should be a list
+[packages]
+msal = {version="==1.20.0", extras=["broker"]}
+parver = '*'
+