source: ValBot/Python/check_interwiki_links.py@ 1192

# Check Interwiki Links
# by iritscen@yahoo.com
# Looks at each link on a page (or in all the pages in a category) which uses a registered
# interwiki prefix and loads the linked page, verifying that it exists and that any section
# link, if present, is valid as well. The output will use the word "ERROR" when it cannot
# validate the interwiki link.
# Recommended viewing width:
# |---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---|

import os

from urllib.parse import urljoin

import pywikibot
import bs4
import re
import requests # for listing members with dir()

from pywikibot.bot import QuitKeyboardInterrupt
from pywikibot import pagegenerators
from pywikibot.tools.formatter import color_format
from pywikibot.comms.http import fetch
from pywikibot.specialbots import UploadRobot
from bs4 import BeautifulSoup

# Parallel arrays based on https://wiki.oni2.net/Special:Interwiki
interwiki_prefixes = ('acronym', 'cache', 'commons', 'dictionary', 'google', 'metawikimedia', 'mw', 'wikibooks', 'wikidata', 'wikimedia', 'wikinews', 'wikipedia', 'wikiquote', 'wikisource', 'wikispecies', 'wikiversity', 'wikivoyage', 'wikt', 'wiktionary', 'wp')

interwiki_urls = ('http://www.acronymfinder.com/~/search/af.aspx?string=exact&Acronym=', 'http://www.google.com/search?q=cache:', 'https://commons.wikimedia.org/wiki/', 'http://www.dict.org/bin/Dict?Database=*&Form=Dict1&Strategy=*&Query=', 'http://www.google.com/search?q=', 'https://meta.wikimedia.org/wiki/', 'https://www.mediawiki.org/wiki/', 'https://en.wikibooks.org/wiki/', 'https://www.wikidata.org/wiki/', 'https://foundation.wikimedia.org/wiki/', 'https://en.wikinews.org/wiki/', 'https://en.wikipedia.org/wiki/', 'https://en.wikiquote.org/wiki/', 'https://wikisource.org/wiki/', 'https://species.wikimedia.org/wiki/', 'https://en.wikiversity.org/wiki/', 'https://en.wikivoyage.org/wiki/', 'https://en.wiktionary.org/wiki/', 'https://en.wiktionary.org/wiki/', 'https://en.wikipedia.org/wiki/')

# Initialize globals
debug = 0
pages_checked = 0
iw_found = 0
errors_issued = 0
name_printed = 0

# Prints the name of a page on which something occurred, if it has not been printed before
def possibly_print(page_name):
   global debug
   global name_printed
   
   if not name_printed and not debug:
      pywikibot.stdout('')
      pywikibot.stdout('From page "{}":'.format(page_name))
      name_printed = 1

# Search a page for the section specified in the link
def find_section(page_text, page_name, page_slug, prefix, print_result):
   global errors_issued

   # Isolate section link
   target_page_name, anchor_name = page_slug.split('#')
   target_page_name_human = target_page_name.replace('_', ' ')
   
   # Convert dot-notation hex entities to proper characters
   anchor_name = anchor_name.replace('.22', '"')
   anchor_name = anchor_name.replace('.27', '\'')
   anchor_name = anchor_name.replace('.28', '(')
   anchor_name = anchor_name.replace('.29', ')')
   
   # Read linked page to see if it really has this anchor link
   soup = BeautifulSoup(page_text, 'html.parser')
   found_section = False
   for the_tag in soup.findAll('span'): # search for span with ID matching the section name
      tag_name = the_tag.get('id', None)
      if tag_name == anchor_name:
         found_section = True
         break
   if found_section == False:
      for the_tag in soup.findAll('div'): # search for div with ID matching the section name
         tag_name = the_tag.get('id', None)
         if tag_name == anchor_name:
            found_section = True
            break
   if found_section == False:
      for the_tag in soup.findAll('h2'): # search for h2 with ID matching the section name
         tag_name = the_tag.get('id', None)
         if tag_name == anchor_name:
            found_section = True
            break
   if found_section == False:
      for the_tag in soup.findAll('h3'): # search for h3 with ID matching the section name
         tag_name = the_tag.get('id', None)
         if tag_name == anchor_name:
            found_section = True
            break
   if found_section == False:
      for the_tag in soup.findAll('h4'): # search for h4 with ID matching the section name
         tag_name = the_tag.get('id', None)
         if tag_name == anchor_name:
            found_section = True
            break
   if found_section == False:
      possibly_print(page_name)
      pywikibot.stdout('   ERROR: Could not find section "{0}" on {1} page "{2}".'.format(anchor_name, prefix, target_page_name_human))
      errors_issued = errors_issued + 1
   elif print_result == True:
      pywikibot.stdout('   The section "{0}" was found on {1} page "{2}".'.format(anchor_name, prefix, target_page_name_human))

# For a link that redirected us to another page, extract the name of the target page from
# the target page's source
def find_canonical_link(page_text, page_name, page_slug, prefix, prefix_url):
   # Extract link from this markup which contains name of redirected-to page:
   # <link rel="canonical" href="https://en.wikipedia.org/wiki/Page_name"/>
   canonical_name = page_text.split('<link rel="canonical" href="')[-1]
   prefix_length = len(prefix_url)
   canonical_name = canonical_name[prefix_length:]
   tag_end = canonical_name.find('">')
   
   if tag_end == -1:
      pywikibot.stdout('   ERROR: The {0} link "{1}" is a redirect page, but this script could not isolate the target page name.'.format(prefix, page_slug))
      errors_issued = errors_issued + 1
   else:
      canonical_name = canonical_name[:tag_end]
      if len(canonical_name) > 100:
         # Certain things can cause the trim to fail; report error and avoid slamming the
         # output with massive page source from a failed trim
         pywikibot.stdout('   ERROR: The {0} link "{1}" is a redirect to "{2}…" (string overflow).'.format(prefix, page_slug, canonical_name[:100]))
         errors_issued = errors_issued + 1
      else:
         canonical_name = canonical_name.replace('_', ' ')
         if '#' in page_slug:
            _, anchor_name = page_slug.split('#')
            pywikibot.stdout('   The {0} link "{1}" is a redirect to "{2}#{3}", which is a valid page. Checking section link….'.format(prefix, page_slug, canonical_name, anchor_name))
            find_section(page_text, page_name, page_slug, prefix, True)
         else:
            pywikibot.stdout('   The {0} link "{1}" is a redirect to "{2}", which is a valid page.'.format(prefix, page_slug, canonical_name))

# Test an interwiki link and look for a section link if applicable
def test_interwiki_link(prefix, prefix_url, iw_url, page_name, page_slug):
   global errors_issued
   
   response = fetch(iw_url)

   # One way we tell that a redirect occurred is by checking fetch's history, as it
   # automatically follows redirects. This will catch formal redirects which come from pages
   # such as Special:PermanentLink.
   if response.history != []:
      possibly_print(page_name)
         
      if page_slug.startswith('WP:') and page_slug == page_slug.upper():
         pywikibot.stdout('   Got redirection code "{0}" for {1} link "{2}". This appears to be a deliberate use of a Wikipedia shortcut. Checking the target page….'.format(response.history[0], prefix, page_slug))
         find_canonical_link(response.text, page_name, page_slug, prefix, prefix_url)
      else:
         permalink1 = 'Special:PermanentLink/'.lower()
         permalink2 = 'Special:Permalink/'.lower()
         page_slug_lower = page_slug.lower()
         if page_slug_lower.startswith(permalink1) or page_slug_lower.startswith(permalink2):
            pywikibot.stdout('   Got redirection code "{0}" for {1} permanent revision link "{2}". Checking the target page….'.format(response.history[0], prefix, page_slug))
            find_canonical_link(response.text, page_name, page_slug, prefix, prefix_url)
         else:
            pywikibot.stdout('   ERROR: Unrecognized type of redirection (code "{0}") for {1} link "{2}". You should check the link manually.'.format(response.history[0], prefix, page_slug))
            errors_issued = errors_issued + 1
   elif response.status_code != 200:
      possibly_print(page_name)
      pywikibot.stdout('   ERROR: Got response code {0} for {1} link "{2}". The page may not exist.'.format(response.status_code, prefix, page_slug))
      errors_issued = errors_issued + 1
   # However the usual way that a redirect occurs is that MediaWiki redirects us sneakily
   # using JavaScript, while returning code OK 200 as if the link was correct; this happens
   # when a redirect page is accessed. We must detect these soft redirects by looking at the
   # page source to find the redirect note inserted at the top of the page for the reader.
   elif 'Redirected from <a' in response.text:
      possibly_print(page_name)
      pywikibot.stdout('   Got silently redirected by {0} link "{1}". Checking the target page….'.format(prefix, page_slug))
      find_canonical_link(response.text, page_name, page_slug, prefix, prefix_url)
   elif '#' in page_slug:
      find_section(response.text, page_name, page_slug, prefix, False)

# Searches the given page text for interwiki links
def scan_for_interwiki_links(page_text, page_name):
   global debug
   global pages_checked
   global iw_found
   global name_printed
   pages_checked = pages_checked + 1
   cur_prefix = 0
   name_printed = 0

   for prefix in interwiki_prefixes:
      # Isolate strings that start with "[[prefix:" and end with "|" or "]"
      iw_link = r"\[\[" + prefix + r":[^|\]]*(\||\])"
      for match in re.finditer(iw_link, page_text):
         # Extract just the page title from this regex match
         s = match.start() + 2 + len(prefix) + 1
         e = match.end() - 1

         # Commonly we use spaces instead of underscores, so fix that before querying
         page_slug = page_text[s:e].replace(' ', '_')

         # But use spaces for title when printing it
         page_title_human = page_slug.replace('_', ' ')
         if debug: pywikibot.stdout('   Validating {0} link "{1}"'.format(prefix, page_title_human))
         iw_found = iw_found + 1

         # Construct full URL for the particular wiki
         iw_url = interwiki_urls[cur_prefix] + page_slug

         # Adjust URL if this is a foreign-language WP link
         if re.match("^[a-zA-Z]{2}:", page_slug):
            lang_code = page_slug[0:2] + "."
            # "wp:" is the Wikipedia: namespace, not a language
            if lang_code != "wp." and lang_code != "WP.":
               iw_url = iw_url.replace('en.', lang_code)
               iw_url = iw_url.replace(page_slug[0:3], '')

         # Test the URL
         test_interwiki_link(prefix, interwiki_urls[cur_prefix], iw_url, page_name, page_slug)
      cur_prefix = cur_prefix + 1

# Print a wrap-up message
def print_summary():
   global pages_checked
   global iw_found
   global errors_issued

   page_str = "pages"
   if pages_checked == 1:
      page_str = "page"

   link_str = "links"
   if iw_found == 1:
      link_str = "link"

   pywikibot.stdout('Checked {0} {1} and found {2} interwiki {3}.'.format(pages_checked, page_str, iw_found, link_str))

   error_str = "errors were"
   if errors_issued == 1:
      error_str = "error was"

   pywikibot.stdout('{0} {1} encountered in validating these links.'.format(errors_issued, error_str))

# Main function
def main(*args):
   global debug
   search_cat = ''
   search_page = ''

   # Process arguments
   local_args = pywikibot.handle_args(args)
   for arg in local_args:
      if arg.startswith('-cat:'):
         search_cat = arg[5:]
      elif arg.startswith('-page:'):
         search_page = arg[6:]
      elif arg == '-dbg':
         debug = 1
      else:
         pywikibot.stdout('Unknown argument "{}". Exiting.'.format(arg))
         return

   #pywikibot.stdout('The members of the requests.models.Response class are:')
   #pywikibot.stdout(format(dir(requests.models.Response)))
   #return
   
   # Check specified page or loop through specified category and check all pages
   site = pywikibot.Site()
   if search_cat != '':
      cat_obj = pywikibot.Category(site, search_cat)
      generator = pagegenerators.CategorizedPageGenerator(cat_obj, recurse=True)
      for page in pagegenerators.PreloadingGenerator(generator, 100):
         if debug: pywikibot.stdout('Checking page "{}"'.format(page.title()))
         scan_for_interwiki_links(page.text, page.title())
   elif search_page != '':
      page = pywikibot.Page(site, search_page)
      if debug: pywikibot.stdout('Checking page "{}"'.format(page.title()))
      scan_for_interwiki_links(page.text, page.title())

   # Print the results
   print_summary()

if __name__ == '__main__':
   main()