Commit Graph

38 Commits

Author SHA1 Message Date
Arel Cordero fd4289dae0 Adding `does_not_raise` to documentation only 2019-01-27 16:42:10 +00:00
Arel Cordero 977adf1354 Improving sphinx docs based on feedback 2019-01-27 16:41:23 +00:00
Arel Cordero afe9fd5ffd Adds `does_not_raise` context manager
Addressing issues #4324 and #1830
2019-01-27 16:40:35 +00:00
Bruno Oliveira 42bb0b3904 Use a.item() instead of deprecated np.asscalar(a)
np.asscalar() has been deprecated in numpy 1.16:

https://github.com/numpy/numpy/blob/master/doc/release/1.16.0-notes.rst#new-deprecations
2019-01-14 08:59:09 -02:00
Ronny Pfannschmidt e63c7a13ff
Merge pull request #4327 from ndevenish/approx
Let approx() work on more generic sequences
2018-12-17 15:21:12 +01:00
Daniel Hahler 8e287c5c77 Merge master into features 2018-12-14 16:44:43 +01:00
Bruno Oliveira 5b83417afc Deprecate the 'message' parameter of pytest.raises
Fix #3974
2018-12-12 22:26:30 -02:00
Anthony Sottile a254ad0436 Raise `TypeError` for `with raises(..., match=<non-None falsey value>)`. 2018-12-12 14:59:22 -08:00
Anthony Sottile 1bba0a9714 Deprecate `raises(..., 'code(as_a_string)')` / `warns(..., 'code(as_a_string)') 2018-11-29 09:34:51 -08:00
Daniel Hahler 5bee396e4b Merge master into merge-master-into-features 2018-11-24 13:47:10 +01:00
Anthony Sottile 0d5298475d Fix `raises(..., "code(string)")` frame filename. 2018-11-23 15:01:35 -08:00
Ronny Pfannschmidt 88bf01a31e fix #4386 - restructure construction and partial state of ExceptionInfo 2018-11-22 21:20:32 +01:00
Nicholas Devenish 1a8d9bf254 Let approx() work on more generic sequences
approx() was updated in 9f3122fe to work better with numpy arrays,
however at the same time the requirements were tightened from
requiring an Iterable to requiring a Sequence - the former being
tested only on interface, while the latter requires subclassing or
registration with the abc.

Since the ApproxSequence only used __iter__ and __len__ this commit
reduces the requirement to only what's used, and allows unregistered
Sequence-like containers to be used.

Since numpy arrays qualify for the new criteria, reorder the checks so
that generic sequences are checked for after numpy arrays.
2018-11-18 20:01:11 +00:00
Anthony Sottile 2368fbb63c Apply reorder-python-imports to all files 2018-10-25 00:01:29 -07:00
Anthony Sottile c2cd337886 py.builtin.exec_ => six.exec_ 2018-08-22 23:00:06 -07:00
Bruno Oliveira 5db2e6c7a1
Merge pull request #3761 from nicoddemus/numpy-recursion-error
Fix recursion in pytest.approx() with arrays in numpy<1.13
2018-08-01 23:40:30 -03:00
Kale Kundert b8255308d6
Make the infinite-recusrion fix more explicit.
So we remember what happened and don't accidentally regress in the
future.
2018-08-01 12:11:03 -07:00
Bruno Oliveira a5c0fb7f6b Rename recursive_map -> _recursive_list_map as requested in review 2018-08-01 15:17:58 -03:00
Bruno Oliveira 7d13599ba1 Fix recursion in pytest.approx() with arrays in numpy<1.13 2018-08-01 08:04:09 -03:00
Bruno Oliveira 2a2f888909 Move recursive_map from local to free function 2018-08-01 07:30:40 -03:00
Bruno Oliveira ad5ddaf55a Simplify is_numpy_array as suggested in review 2018-08-01 07:28:39 -03:00
Bruno Oliveira 5003bae0de Fix 'at' string for non-numeric messages in approx() 2018-08-01 07:07:37 -03:00
Bruno Oliveira 611d254ed5 Improve error checking messages: add position and use pprint 2018-08-01 07:01:00 -03:00
Bruno Oliveira 098dca3a9f Use {!r} for a few other messages as well 2018-07-31 21:14:51 -03:00
Bruno Oliveira bf7c188cc0 Improve error message for invalid types passed to pytest.approx()
* Hide the internal traceback
* Use !r representation instead of !s (the default for {} formatting)
2018-07-31 21:08:24 -03:00
Kale Kundert d02491931a
Fix the unused import. 2018-07-31 11:33:46 -07:00
Kale Kundert 032db159c9
Let black reformat the code... 2018-07-31 11:23:23 -07:00
Kale Kundert cd2085ee71
approx(): Detect type errors earlier. 2018-07-31 00:26:35 -07:00
Kale Kundert ad305e71d7
Improve docstrings for Approx classes. 2018-07-30 23:26:57 -07:00
Kale Kundert 7d8688d54b
Reflect dimension in approx repr for numpy arrays. 2018-07-30 23:23:17 -07:00
Kale Kundert bf127a63b2
Need to iterate over the flattened array. 2018-07-27 11:24:42 -07:00
Bruno Oliveira d0ba242c46 Implement change suggested by @kalekundert in PR 2018-07-27 15:07:20 -03:00
abrammer f0db64ac2e drop the duplicate approx call
update test to include both np.array(actual) and np.array(expected)
2018-07-24 21:18:44 -04:00
Alan 7bff5866b1 bugfix in ApproxNumpy initialisation, use keywords for arguments now 2018-07-18 17:29:55 -04:00
Anthony Sottile cbaa7dd56a Upgrade pre-commit hooks except pyupgrade 2018-06-26 06:35:27 -07:00
Marcelo Duarte Trevisani 0030ceb11c Bug fix #3593 - approx method in a single element of numpy array
If the user pass as a expected value a numpy array created like
numpy.array(5); it will creates an array with one element without shape,
when used with approx it will raise an error
'TypeError: iteration over a 0-d array'
2018-06-25 22:55:16 +02:00
Hynek Schlawack acb8f23311 Fix 2 DeprecationWarnings about invalid escape seq 2018-06-16 10:37:33 +02:00
Ronny Pfannschmidt eaa882f3d5 switch to src layout 2018-05-26 09:10:38 +02:00