Add disable_skip_errors flag to ConventionChecker API (#485)

* Add disable_skip_errors flag to ConventionChecker API

Added to ConventionChecker.check_source() and ConventionChecker.check().
This allows other systems that integrate pydocstyle,
such as flake8-docstyle, to request errors that are otherwise
skipped due to # noqa comments.

Added test for skip_errors feature in general and with this flag.

Fixes #484

* Update release notes

* rename argument to ignore inline noqa comments

* Add docstring description of new argument

* Move release note

Co-authored-by: Sambhav Kothari <sambhavs.email@gmail.com>

Author: plinss

docs/release_notes.rst

@@ -8,6 +8,9 @@ Release Notes
 Current Development Version
 ---------------------------
 
+New Features
+
+* Add flag to disable `# noqa` comment processing in API (#485).
 
 5.1.1 - August 29th, 2020
 ---------------------------

src/pydocstyle/checker.py

@@ -104,7 +104,7 @@ class ConventionChecker:
         ".+"            # Followed by 1 or more characters - which is the docstring for the parameter
     )
 
-    def check_source(self, source, filename, ignore_decorators=None):
+    def check_source(self, source, filename, ignore_decorators=None, ignore_inline_noqa=False):
         module = parse(StringIO(source), filename)
         for definition in module:
             for this_check in self.checks:
@@ -114,15 +114,15 @@ def check_source(self, source, filename, ignore_decorators=None):
                     decorator_skip = ignore_decorators is not None and any(
                         len(ignore_decorators.findall(dec.name)) > 0
                         for dec in definition.decorators)
-                    if not skipping_all and not decorator_skip:
+                    if (ignore_inline_noqa or not skipping_all) and not decorator_skip:
                         error = this_check(self, definition,
                                            definition.docstring)
                     else:
                         error = None
                     errors = error if hasattr(error, '__iter__') else [error]
                     for error in errors:
-                        if error is not None and error.code not in \
-                                definition.skipped_error_codes:
+                        if error is not None and (ignore_inline_noqa or error.code not in \
+                                definition.skipped_error_codes):
                             partition = this_check.__doc__.partition('.\n')
                             message, _, explanation = partition
                             error.set_context(explanation=explanation,
@@ -932,7 +932,7 @@ def check_docstring_sections(self, definition, docstring):
 parse = Parser()
 
 
-def check(filenames, select=None, ignore=None, ignore_decorators=None):
+def check(filenames, select=None, ignore=None, ignore_decorators=None, ignore_inline_noqa=False):
     """Generate docstring errors that exist in `filenames` iterable.
 
     By default, the PEP-257 convention is checked. To specifically define the
@@ -949,6 +949,8 @@ def check(filenames, select=None, ignore=None, ignore_decorators=None):
     convenience, you may use `pydocstyle.violations.conventions.pep257` as
     a base set to add or remove errors from.
 
+    `ignore_inline_noqa` controls if `# noqa` comments are respected or not.
+
     Examples
     ---------
     >>> check(['pydocstyle.py'])
@@ -978,7 +980,8 @@ def check(filenames, select=None, ignore=None, ignore_decorators=None):
             with tk.open(filename) as file:
                 source = file.read()
             for error in ConventionChecker().check_source(source, filename,
-                                                          ignore_decorators):
+                                                          ignore_decorators,
+                                                          ignore_inline_noqa):
                 code = getattr(error, 'code', None)
                 if code in checked_codes:
                     yield error

src/tests/test_integration.py

@@ -190,6 +190,36 @@ def function_with_bad_docstring(foo):
         assert error_codes == expected_error_codes - ignored
 
 
+def test_skip_errors():
+    """Test that `ignore`d errors are not reported in the API."""
+    function_to_check = textwrap.dedent('''
+        def function_with_bad_docstring(foo):  # noqa: D400, D401, D403, D415
+            """ does spacinwithout a period in the end
+            no blank line after one-liner is bad. Also this - """
+            return foo
+    ''')
+    expected_error_codes = {'D100', 'D205', 'D209', 'D210', 'D213'}
+    mock_open = mock.mock_open(read_data=function_to_check)
+    from pydocstyle import checker
+    with mock.patch.object(
+            checker.tk, 'open', mock_open, create=True):
+        # Passing a blank ignore here explicitly otherwise
+        # checkers takes the pep257 ignores by default.
+        errors = tuple(checker.check(['filepath'], ignore={}))
+        error_codes = {error.code for error in errors}
+        assert error_codes == expected_error_codes
+
+    skipped_error_codes = {'D400', 'D401', 'D403', 'D415'}
+    # We need to recreate the mock, otherwise the read file is empty
+    mock_open = mock.mock_open(read_data=function_to_check)
+    with mock.patch.object(
+            checker.tk, 'open', mock_open, create=True):
+        errors = tuple(checker.check(['filepath'], ignore={},
+                                     ignore_inline_noqa=True))
+        error_codes = {error.code for error in errors}
+        assert error_codes == expected_error_codes | skipped_error_codes
+
+
 def test_run_as_named_module():
     """Test that pydocstyle can be run as a "named module".