Fixing Python 3.13 Regex Error | Generated by AI

Home 2025.11

The error you’re seeing is caused by Python 3.13 being stricter about regular expression syntax than previous versions.

Specifically, this line in wikiextractor/extract.py (around line 378) is now invalid in Python 3.13:

ExtLinkBracketedRegex = re.compile(
    '\\[(((?i)' + '|'.join(wgUrlProtocols) + ')' + EXT_LINK_URL_CLASS + r'+)\s*([^\\]\x00-\x08\x0a-\x1F]*?)\\]',
    re.S | re.U)

The problem is the inline (?i) flag placed in the middle of the string after string concatenation. Starting with Python 3.12/3.13, inline flags like (?i) are no longer allowed anywhere except at the very beginning of the regex pattern.

Quick Fix (works on Python 3.13)

Edit the file:

~/.local/lib/python3.13/site-packages/wikiextractor/extract.py

Go to line ~378 and replace the broken block with this corrected version:

# Old broken way (fails on Python 3.13)
# ExtLinkBracketedRegex = re.compile(
#     '\\[(((?i)' + '|'.join(wgUrlProtocols) + ')' + EXT_LINK_URL_CLASS + r'+)\s*([^\\]\x00-\x08\x0a-\x1F]*?)\\]',
#     re.S | re.U)

# Fixed version – move the case-insensitivity flag to the re.compile flags
protocol_pattern = '|'.join(wgUrlProtocols)
ExtLinkBracketedRegex = re.compile(
    r'\\['
    r'(((' + protocol_pattern + r')' + EXT_LINK_URL_CLASS + r'+)'
    r'\s*([^\\]\x00-\x08\x0a-\x1F]*?)\\])',
    re.IGNORECASE | re.DOTALL | re.UNICODE
)

Or even cleaner (recommended):

ExtLinkBracketedRegex = re.compile(
    r'\\['
    r'(((' + '|'.join(wgUrlProtocols) + r')' + EXT_LINK_URL_CLASS + r'+)'
    r'\s*([^\\]\x00-\x08\x0a-\x1F]*?)\\])',
    re.I | re.S | re.U
)

re.I is the same as re.IGNORECASE, and it can be passed as a flag instead of embedding (?i) inside the pattern.

Alternative solutions

  1. Use Python 3.11 or 3.12 (easiest if you don’t want to patch) 
    python3.11 -m pip install wikiextractor
# then run with python3.11 wikiextractor ...
  2. Use the maintained fork that already fixes Python 3.13 compatibility: 
    pip uninstall wikiextractor
pip install git+https://github.com/attardi/wikiextractor.git

    Giuseppe Attardi’s fork has been updated for Python 3.12+ and works perfectly on 3.13.

  3. Use the very fast modern alternative (highly recommended for large dumps): 
    pip install wikiextractor-ng   # or "wiki-dump-tools", "wikitextprocessor", etc.

After applying the patch or switching to the fixed fork, your command will work again:

wikiextractor *.bz2

Let me know if you want a one-liner patch or the exact sed command to fix it automatically.

Back

x-ai/grok-4-fast

Donate