fix #832 : implemented CheckedSession, CheckedParameters and ArrayConstraint classes (based on preliminary code written by gdementen)

Author: alixdamman

README.rst

@@ -129,6 +129,12 @@ For plotting
 - `matplotlib <http://matplotlib.org/>`__:
   required for plotting.
 
+Miscellaneous
+~~~~~~~~~~~~~
+
+- `pydantic <https://github.com/samuelcolvin/pydantic>`__:
+  required to use `CheckedSession`.
+
 .. _start-documentation:
 
 Documentation

doc/source/api.rst

@@ -795,7 +795,6 @@ Modifying
 
    Session.add
    Session.update
-   Session.get
    Session.apply
    Session.transpose
 
@@ -821,6 +820,30 @@ Load/Save
    Session.to_hdf
    Session.to_pickle
 
+ArrayConstraint
+===============
+
+.. autosummary::
+   :toctree: _generated/
+
+   ArrayConstraint
+
+CheckedSession
+==============
+
+.. autosummary::
+   :toctree: _generated/
+
+   CheckedSession
+
+CheckedParameters
+=================
+
+.. autosummary::
+   :toctree: _generated/
+
+   CheckedParameters
+
 .. _api-editor:
 
 Editor

doc/source/changes/version_0_33.rst.inc

@@ -20,30 +20,21 @@ New features
 
 * added official support for Python 3.9 (0.32.3 already supports it even though it was not mentioned).
 
-* added a feature (see the :ref:`miscellaneous section <misc>` for details). It works on :ref:`api-axis` and
-  :ref:`api-group` objects.
+* added :py:obj:`CheckedSession`, :py:obj:`CheckedParameters` and :py:obj:`ArrayConstraint` objects.
 
-  Here is an example of the new feature:
+  `CheckedSession` is intended to be inherited by user defined classes in which the variables of a model
+  are declared. By declaring variables, users will speed up the development of their models using the auto-completion
+  (the feature in which development tools like PyCharm try to predict the variable or function a user intends
+  to enter after only a few characters have been typed). All user defined classes inheriting from `CheckedSession`
+  will have access to the same methods as `Session` objects.
 
-    >>> arr = ndtest((2, 3))
-    >>> arr
-    a\b  b0  b1  b2
-     a0   0   1   2
-     a1   3   4   5
+  `CheckedParameters` is the same as `CheckedSession` but the declared variables cannot be
+  modified after initialization.
 
-  And it can also be used like this:
+  The special :py:funct:`ArrayConstraint` type represents an Array object with fixed axes and/or dtype.
+  It is intended to be only used along with :py:class:`CheckedSession`.
 
-    >>> arr = ndtest("a=a0..a2")
-    >>> arr
-    a  a0  a1  a2
-        0   1   2
-
-* added another feature in the editor (closes :editor_issue:`1`).
-
-  .. note::
-
-     - It works for foo bar !
-     - It does not work for foo baz !
+  Closes :issue:`832`.
 
 
 .. _misc:

environment.yml

@@ -12,4 +12,5 @@ dependencies:
   - pytest>=3.5
   - flake8
   - pip:
-    - pytest-flake8
+    - pytest-flake8
+    - pydantic==1.5

larray/__init__.py

@@ -8,6 +8,7 @@
                                eye, all, any, sum, prod, cumsum, cumprod, min, max, mean, ptp, var,
                                std, median, percentile, stack, zip_array_values, zip_array_items)
 from larray.core.session import Session, local_arrays, global_arrays, arrays
+from larray.core.checked import ArrayConstraint, CheckedSession, CheckedParameters
 from larray.core.constants import nan, inf, pi, e, euler_gamma
 from larray.core.metadata import Metadata
 from larray.core.ufuncs import wrap_elementwise_array_func, maximum, minimum, where
@@ -55,6 +56,8 @@
     'median', 'percentile', 'stack', 'zip_array_values', 'zip_array_items',
     # session
     'Session', 'local_arrays', 'global_arrays', 'arrays',
+    # constrained
+    'ArrayConstraint', 'CheckedSession', 'CheckedParameters',
     # constants
     'nan', 'inf', 'pi', 'e', 'euler_gamma',
     # metadata

larray/core/axis.py

@@ -839,7 +839,7 @@ def __getitem__(self, key):
         -----
         key is label-based (slice and fancy indexing are supported)
         """
-        # if isinstance(key, basestring):
+        # if isinstance(key, str):
         #     key = to_keys(key)
 
         def isscalar(k):
@@ -862,7 +862,7 @@ def isscalar(k):
             and key.name in self
         ):
             return LGroup(key.name, None, self)
-        # elif isinstance(key, basestring) and key in self:
+        # elif isinstance(key, str) and key in self:
             # TODO: this is an awful workaround to avoid the "processing" of string keys which exist as is in the axis
             #       (probably because the string was used in an aggregate function to create the label)
             # res = LGroup(slice(None), None, self)