Merge pull request #287 from hakonhagland/setkw_status

Use more precise regex
OPM · Jun 6, 2024 · c902de6 · c902de6
2 parents 3d40a6f + 633ad63
commit c902de6
Show file tree

Hide file tree

Showing 3 changed files with 7 additions and 4 deletions.
diff --git a/scripts/python/src/fodt/add_keyword.py b/scripts/python/src/fodt/add_keyword.py
@@ -11,7 +11,7 @@
 
 import click
 
-from fodt.constants import ClickOptions, Directories, FileExtensions, KeywordStatus
+from fodt.constants import ClickOptions, Directories, FileExtensions, KeywordStatus, Regex
 from fodt.create_subdocument import CreateSubDocument3
 from fodt.helpers import Helpers
 from fodt.remove_subsections import RemoveSubSections
@@ -118,7 +118,7 @@ def extract_keyword_name(self, href: str) -> str:
         # Assume href starts with "#xxx.yyy.zzz.KEYWORD_NAME<space>"
         # or "#xxx.yyy.zzz.KEYWORD_NAME|outline"
         # KEYWORD_NAME can contain letters, numbers, and optionally a trailing hyphen or en-dash
-        if m:= re.match(r"#\d+\.\d+\.\d+\.(\w+[\-–]?)(?:\s+|$|\|outline$)", href):
+        if m:= re.match(Regex.href_keyword_name, href):
             return m.group(1)
         else:
             return '<NOT FOUND>'

diff --git a/scripts/python/src/fodt/add_keyword_status.py b/scripts/python/src/fodt/add_keyword_status.py
@@ -17,7 +17,7 @@
 
 import click
 
-from fodt.constants import ClickOptions, Directories, FileExtensions, KeywordStatus
+from fodt.constants import ClickOptions, Directories, FileExtensions, KeywordStatus, Regex
 from fodt.xml_helpers import XMLHelper
 
 class AppendixKeywordHandler(xml.sax.handler.ContentHandler):
@@ -132,7 +132,7 @@ def handle_table_row(
                 href = attrs.getValue("xlink:href")
                 # the href value is on the form "#1.2.1.ACTDIMS – ACTION Keyword Dimensions"
                 # we want to extract the keyword name from this string
-                if match := re.match(r"#\d+.\d+.\d+.(\w+)\s+", href):
+                if match := re.match(Regex.href_keyword_name, href):
                     self.current_keyword = match.group(1)
         elif self.in_table_cell and name == 'text:p':
             if self.found_table_cell:

diff --git a/scripts/python/src/fodt/constants.py b/scripts/python/src/fodt/constants.py
@@ -76,6 +76,9 @@ class MetaSections():
         'office:master-styles',
     ]
 
+class Regex:
+    href_keyword_name = r"#\d+\.\d+\.\d+\.(\w+[\-–]?)(?:\s+|$|\|outline$)"
+
 class TagEvent():
     NONE = 0
     START = 1