Add valid word characters as a configuration option

docs/source/history.rst

@@ -6,6 +6,27 @@
 
    unmaintained
 
+Unreleased
+==========
+
+Bug Fixes
+---------
+
+None.
+
+New Features
+------------
+
+- `#147 <https://github.com/sphinx-contrib/spelling/pull/147>`__
+  Adds the ability to pass in valid word characters as the
+  configuration option ``spelling_valid_word_characters``, where its
+  value is a tuple of strings. For example,
+  ``spelling_valid_word_characters = ("'", "-",)``. By default
+  PyEnchant considers only ``'`` as a valid word character. This pull
+  request also adds the hyphen ``-`` as a valid word character to
+  prevent PyEnchant from tokenizing hyphenated words into two words
+  and marking them as misspellings.
+
 7.3.2
 =====
 

docs/source/spelling_wordlist.txt

@@ -6,6 +6,7 @@ pypi
 reStructuredText
 sphinxcontrib
 tokenizer
+tokenizing
 txt
 wikis
 wordfiles

sphinxcontrib/spelling/__init__.py

@@ -54,6 +54,8 @@ def setup(app):
     app.add_config_value('spelling_filters', [], 'env')
     # Set a user-provided list of files to ignore
     app.add_config_value('spelling_exclude_patterns', [], 'env')
+    # Set valid word characters
+    app.add_config_value('spelling_valid_word_characters', ("'", "-",), 'env')
     # Choose whether or not the misspelled output should be displayed
     # in the terminal
     app.add_config_value('spelling_verbose', True, 'env')

sphinxcontrib/spelling/builder.py

@@ -86,6 +86,7 @@ def init(self):
             word_list_filename=word_list,
             filters=f,
             context_line=self.config.spelling_show_whole_line,
+            valid_word_chars=self.config.spelling_valid_word_characters,
         )
 
     def _load_filter_classes(self, filters):

sphinxcontrib/spelling/checker.py

@@ -21,7 +21,8 @@ class SpellingChecker:
     """
 
     def __init__(self, lang, suggest, word_list_filename,
-                 tokenizer_lang='en_US', filters=None, context_line=False):
+                 tokenizer_lang='en_US', filters=None, context_line=False,
+                 valid_word_chars=None):
         if enchant_import_error is not None:
             raise RuntimeError(
                 'Cannot instantiate SpellingChecker '