X-Git-Url: https://git.mdrn.pl/wolnelektury.git/blobdiff_plain/3c12104d064098096eb5fe9c9ee122d0293e10f2..8fc17a12a92f649f2b9dd8c884a85d1a7232f248:/apps/search/views.py diff --git a/apps/search/views.py b/apps/search/views.py index 166011cc3..e9b25649c 100644 --- a/apps/search/views.py +++ b/apps/search/views.py @@ -1,5 +1,6 @@ # -*- coding: utf-8 -*- +from django.conf import settings from django.shortcuts import render_to_response, get_object_or_404 from django.template import RequestContext from django.contrib.auth.decorators import login_required @@ -20,11 +21,20 @@ import enchant dictionary = enchant.Dict('pl_PL') +def match_word_re(word): + if 'sqlite' in settings.DATABASES['default']['ENGINE']: + return r"\b%s\b" % word + elif 'mysql' in settings.DATABASES['default']['ENGINE']: + return "[[:<:]]%s[[:>:]]" % word + + def did_you_mean(query, tokens): change = {} - # sprawdzić, czy słowo nie jest aby autorem - proste szukanie termu w author! for t in tokens: - print("%s ok? %s, sug: %s" %(t, dictionary.check(t), dictionary.suggest(t))) + authors = Tag.objects.filter(category='author', name__iregex=match_word_re(t)) + if len(authors) > 0: + continue + if not dictionary.check(t): try: change[t] = dictionary.suggest(t)[0] @@ -109,25 +119,39 @@ def main(request): hint.books(book) toks = StringReader(query) + tokens_cache = {} fuzzy = 'fuzzy' in request.GET if fuzzy: fuzzy = 0.7 - results = SearchResult.aggregate(srch.search_perfect_book(toks, fuzzy=fuzzy, hint=hint), - srch.search_book(toks, fuzzy=fuzzy, hint=hint), - srch.search_perfect_parts(toks, fuzzy=fuzzy, hint=hint), - srch.search_everywhere(toks, fuzzy=fuzzy, hint=hint)) - - for r in results: - r.process_hits() - + author_results = srch.search_phrase(toks, 'authors', fuzzy=fuzzy, tokens_cache=tokens_cache) + title_results = srch.search_phrase(toks, 'title', fuzzy=fuzzy, tokens_cache=tokens_cache) + + # Boost main author/title results with mixed search, and save some of its results for end of list. + # boost author, title results + author_title_mixed = srch.search_some(toks, ['authors', 'title', 'tags'], fuzzy=fuzzy, tokens_cache=tokens_cache) + author_title_rest = [] + for b in author_title_mixed: + bks = filter(lambda ba: ba.book_id == b.book_id, author_results + title_results) + for b2 in bks: + b2.boost *= 1.1 + if bks is []: + author_title_rest.append(b) + + text_phrase = SearchResult.aggregate(srch.search_phrase(toks, 'content', fuzzy=fuzzy, tokens_cache=tokens_cache)) + [r.process_hits() for r in text_phrase] + + everywhere = SearchResult.aggregate(srch.search_everywhere(toks, fuzzy=fuzzy, tokens_cache=tokens_cache), author_title_rest) + [r.process_hits() for r in everywhere] + + for res in [author_results, title_results, text_phrase, everywhere]: + res.sort(reverse=True) + + suggestion = did_you_mean(query, srch.get_tokens(toks, field="SIMPLE")) + + results = author_results + title_results + text_phrase + everywhere results.sort(reverse=True) - - for r in results: - print "-----" - for h in r.hits: - print "- %s" % h - + if len(results) == 1: if len(results[0].hits) == 0: return HttpResponseRedirect(results[0].book.get_absolute_url()) @@ -137,11 +161,18 @@ def main(request): elif len(results) == 0: form = PublishingSuggestForm(initial={"books": query + ", "}) return render_to_response('catalogue/search_no_hits.html', - {'tags': tag_list, 'prefix': query, - "form": form}, + {'tags': tag_list, + 'prefix': query, + "form": form, + 'did_you_mean': suggestion}, context_instance=RequestContext(request)) return render_to_response('catalogue/search_multiple_hits.html', - {'tags': tag_list, 'prefix': query, - 'results': results}, + {'tags': tag_list, + 'prefix': query, + 'results': { 'author': author_results, + 'title': title_results, + 'content': text_phrase, + 'other': everywhere}, + 'did_you_mean': suggestion}, context_instance=RequestContext(request))