Fix D401 and D404 checks not working for docstrings containing only one word and ending with non-alpha character (#421) (#422)

* Fix D401 and D404 checks not working for docstrings containing only one word and ending with non-alpha character (#421)

* Add test and information in release notes. (#421)

* Fix too long test documentation. (#421)

* Move regex for stripping non-alphanumeric characters outside of strip_non_alphanumeric function. (#421)

Author: Strus

docs/release_notes.rst

@@ -31,6 +31,7 @@ Bug Fixes
 * Fix decorator parsing (#411).
 * Google-style sections no longer cause false errors when used with
   Numpy-style sections (#388, #424).
+* Fix D401 and D404 checks not working for docstrings containing only one word and ending with non-alpha character (#421)
 
 4.0.1 - August 14th, 2019
 -------------------------

src/pydocstyle/checker.py

@@ -14,7 +14,7 @@
 from .parser import (Package, Module, Class, NestedClass, Definition, AllError,
                      Method, Function, NestedFunction, Parser, StringIO,
                      ParseError)
-from .utils import log, is_blank, pairwise, common_prefix_length
+from .utils import log, is_blank, pairwise, common_prefix_length, strip_non_alphanumeric
 from .wordlists import IMPERATIVE_VERBS, IMPERATIVE_BLACKLIST, stem
 
 
@@ -437,7 +437,7 @@ def check_imperative_mood(self, function, docstring):  # def context
         if docstring and not function.is_test:
             stripped = ast.literal_eval(docstring).strip()
             if stripped:
-                first_word = stripped.split()[0]
+                first_word = strip_non_alphanumeric(stripped.split()[0])
                 check_word = first_word.lower()
 
                 if check_word in IMPERATIVE_BLACKLIST:
@@ -499,10 +499,16 @@ def check_starts_with_this(self, function, docstring):
         with "This class is [..]" or "This module contains [..]".
 
         """
-        if docstring:
-            first_word = ast.literal_eval(docstring).split()[0]
-            if first_word.lower() == 'this':
-                return violations.D404()
+        if not docstring:
+            return
+
+        stripped = ast.literal_eval(docstring).strip()
+        if not stripped:
+            return
+
+        first_word = strip_non_alphanumeric(stripped.split()[0])
+        if first_word.lower() == 'this':
+            return violations.D404()
 
     @staticmethod
     def _is_docstring_section(context):

src/pydocstyle/utils.py

@@ -1,5 +1,7 @@
 """General shared utilities."""
+import ast
 import logging
+import re
 from typing import Iterable, Any, Tuple
 from itertools import tee, zip_longest
 
@@ -8,6 +10,9 @@
 __version__ = '4.0.1'
 log = logging.getLogger(__name__)
 
+#: Regular expression for stripping non-alphanumeric characters
+NON_ALPHANUMERIC_STRIP_RE = re.compile(r'[\W_]+')
+
 
 def is_blank(string: str) -> bool:
     """Return True iff the string contains only whitespaces."""
@@ -37,4 +42,9 @@ def common_prefix_length(a: str, b: str) -> int:
     for common, (ca, cb) in enumerate(zip(a, b)):
         if ca != cb:
             return common
-    return min(len(a), len(b))
+    return min(len(a), len(b))
+
+
+def strip_non_alphanumeric(string: str) -> str:
+    """Strip string from any non-alphanumeric characters."""
+    return NON_ALPHANUMERIC_STRIP_RE.sub('', string)

src/tests/test_cases/test.py

@@ -309,6 +309,12 @@ def sdgfsdg23245():
     """Constructor for a foo."""
 
 
+@expect("D401: First line should be in imperative mood; try rephrasing "
+        "(found 'Constructor')")
+def sdgfsdg23245777():
+    """Constructor."""
+
+
 @expect('D402: First line should not be the function\'s "signature"')
 def foobar():
     """Signature: foobar()."""

src/tests/test_utils.py

@@ -26,3 +26,8 @@ def test_differ_length():
 def test_empty_string():
     """Test common prefix length of two strings, one of them empty."""
     assert utils.common_prefix_length('abcd', '') == 0
+
+
+def test_strip_non_alphanumeric():
+    """Test strip of a string leaves only alphanumeric characters."""
+    assert utils.strip_non_alphanumeric("  1abcd1...") == "1abcd1"