Revert "fix(utils): fix the regexp matching with problematic unicode casing check in extract_relevant_sections()"

This reverts commit c11b338.

Related to #2684.

Author: BoboTiG

wikidict/utils.py

@@ -1025,30 +1025,24 @@ def extract_keywords_from(parts: list[str]) -> defaultdict[str, str]:
 
 
 def extract_relevant_sections(wikitext: str, locale: str) -> str:
-    """Extract relevant sections for the chosen locale from a given wikitext.
-
-    Process: we check all lines in lower case but keep relevant ones in their original casing.
-    This is to workaround unicode diacritics being lost when working on Turkish, for instance.
-    See https://stackoverflow.com/q/79169550/1117028 for more details.
-    """
+    """Extract relevant sections for the chosen locale from a given wikitext."""
     level = lang.section_level[locale]
     equals = "=" * level
 
     interesting_sections = [
-        re.compile(rf"{equals}[ ]*{re.escape(section)}[ ]*{equals}") for section in lang.head_sections[locale]
+        re.compile(rf"{equals}[ ]*{re.escape(section)}[ ]*{equals}", flags=re.IGNORECASE)
+        for section in lang.head_sections[locale]
     ]
 
     cleaned: list[str] = []
     in_expected_section = False
-    raw_lines = wikitext.splitlines()
-    raw_lines_lower = wikitext.lower().splitlines()
-    for idx, raw_line_lower in enumerate(raw_lines_lower):
-        if not (line := raw_line_lower.strip()):
+    for raw_line in wikitext.splitlines():
+        if not (line := raw_line.strip()):
             continue
         if line.startswith(equals) and line[level] != "=":
             in_expected_section = any(pattern.match(line) for pattern in interesting_sections)
         if in_expected_section:
-            cleaned.append(raw_lines[idx])
+            cleaned.append(line)
     return "\n".join(cleaned) if cleaned else ""