Merge pull request #4 from SuffolkLITLab/spellchecking

Add a basic spellchecking functionality to information page

docassemble/ALLinter/data/questions/interview_wide_linter.yml

@@ -57,6 +57,9 @@ code: |
 code: |
   warnings = text_violations(interview_texts_no_mako)
 ---
+code: |
+  misspelled = get_misspelled_words(paragraph)
+---
 event: results
 question: |
   Interview Suggestions
@@ -67,6 +70,16 @@ subquestion: |
   * ${ key } : ${ val }
   % endfor
   
+  % if misspelled:
+  ### Misspelled words
+  
+  It looks like these words are misspelled:
+  
+  % for misspelled_word in misspelled:
+  * ${ misspelled_word }
+  % endfor  
+  % endif
+  
   % if headings_warnings or warnings:
   ### Warnings:
   
@@ -83,6 +96,8 @@ subquestion: |
   % endfor
   
   % endif
+  
+  
 help:
   label: |
     Text used

docassemble/ALLinter/linter.py

@@ -6,7 +6,33 @@
 import mako.exceptions
 import docassemble.webapp.screenreader
 import docassemble.base.filter
-from typing import List, Tuple
+from typing import List, Tuple, Dict, Mapping, Set, Union
+from spellchecker import SpellChecker
+
+__all__ = [
+    "get_misspelled_words",
+    "get_corrections",
+    "load_interview",
+    "remove_mako",
+    "get_all_headings",
+    "get_heading_width",
+    "headings_violations",
+    "text_violations",
+    "get_all_text",
+]
+
+def get_misspelled_words(text:str, language:str="en")->Set[str]:
+    spell = SpellChecker(language=language)
+    return spell.unknown(re.findall(r'\w+', text))
+
+def get_corrections(misspelled:Union[Set[str], List[str]], language:str="en")->Mapping[str, Set[str]]:
+    spell = SpellChecker(language=language)
+    return [
+        {
+            misspelled_word: spell.corrections(misspelled_word)
+        }
+       for misspelled_word in misspelled
+    ]
 
 mega_list = ('default', 'input type', 'using', 'keep for training',
       'validation messages', 'validate', 'rows', 'maximum image size',

setup.py

@@ -53,7 +53,7 @@ def find_package_data(where='.', package='', exclude=standard_exclude, exclude_d
       url='https://docassemble.org',
       packages=find_packages(),
       namespace_packages=['docassemble'],
-      install_requires=['Mako>=1.1.4', 'ruamel.yaml>=0.17.4', 'plumbum>=1.7.2',  'textstat>=0.7.0'],
+      install_requires=['Mako>=1.1.4', 'pyspellchecker>=0.6.3', 'ruamel.yaml>=0.17.4', 'plumbum>=1.7.2',  'textstat>=0.7.0'],
       zip_safe=False,
       package_data=find_package_data(where='docassemble/ALLinter/', package='docassemble.ALLinter'),
      )