src/search/management/commands/snippets.py

   1 # -*- coding: utf-8 -*-
   2 # This file is part of Wolnelektury, licensed under GNU Affero GPLv3 or later.
   3 # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
   4 #
   5 from django.core.management.base import BaseCommand
   6
   7 from glob import glob
   8 from optparse import make_option
   9 from os import path
  10 from sys import stdout
  11 from django.conf import settings
  12
  13
  14 class Command(BaseCommand):
  15     help = 'Reindex everything.'
  16     args = ''
  17
  18     option_list = BaseCommand.option_list + (
  19         make_option('-C', '--check-just-read', action='store_true', dest='check', default=False,
  20                     help='Check snippets utf-8'),
  21         make_option('-c', '--check', action='store_true', dest='check2', default=False,
  22                     help='Check snippets utf-8 by walking through index'),
  23         )
  24
  25     def handle(self, *args, **opts):
  26         from search.index import Search, Snippets
  27
  28         if opts['check']:
  29             sfn = glob(settings.SEARCH_INDEX+'snippets/*')
  30             print sfn
  31             for fn in sfn:
  32                 print fn
  33                 bkid = int(path.basename(fn))
  34                 with open(fn) as f:
  35                     cont = f.read()
  36                     try:
  37                         uc = cont.decode('utf-8')
  38                     except UnicodeDecodeError, ude:
  39                         print "error in snippets %d" % bkid
  40         if opts['check2']:
  41             s = Search()
  42             reader = s.searcher.getIndexReader()
  43             numdocs = reader.numDocs()
  44             for did in range(numdocs):
  45                 doc = reader.document(did)
  46                 if doc and doc.get('book_id'):
  47                     bkid = int(doc.get('book_id'))
  48                     # import pdb; pdb.set_trace()
  49                     stdout.write("\r%d / %d" % (did, numdocs))
  50                     stdout.flush()
  51                     ss = doc.get('snippet_position')
  52                     sl = doc.get('snippet_length')
  53                     if ss and sl:
  54                         # WTF (nie było zaimportowane)
  55                         snips = Snippets(bkid)
  56                         try:
  57                             txt = snips.get((ss, sl))
  58                             assert len(txt) == sl
  59                         except UnicodeDecodeError, ude:
  60                             stdout.write("\nerror in snippets %d\n" % bkid)
  61                             raise ude
  62
  63             stdout.write("\ndone.\n")