X-Git-Url: https://git.mdrn.pl/wolnelektury.git/blobdiff_plain/73ce961f14509aabfa26536f847afd28111029c6..96a96c16513a98378e2f75ead26a22c358795182:/apps/search/views.py diff --git a/apps/search/views.py b/apps/search/views.py index d056a1826..00391f1be 100644 --- a/apps/search/views.py +++ b/apps/search/views.py @@ -31,7 +31,6 @@ def match_word_re(word): def did_you_mean(query, tokens): change = {} 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 @@ -120,34 +119,44 @@ 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() - - results.sort(reverse=True) + 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, snippets=True, book=False)) + + everywhere = SearchResult.aggregate(srch.search_everywhere(toks, fuzzy=fuzzy, tokens_cache=tokens_cache), author_title_rest) - for r in results: - print "-----" - for h in r.hits: - print "- %s" % h + for res in [author_results, title_results, text_phrase, everywhere]: + res.sort(reverse=True) - # Did you mean? suggestion = did_you_mean(query, srch.get_tokens(toks, field="SIMPLE")) + results = author_results + title_results + text_phrase + everywhere + results.sort(reverse=True) + if len(results) == 1: - if len(results[0].hits) == 0: - return HttpResponseRedirect(results[0].book.get_absolute_url()) - elif len(results[0].hits) == 1 and results[0].hits[0] is not None: - frag = Fragment.objects.get(anchor=results[0].hits[0]) + fragment_hits = filter(lambda h: 'fragment' in h, results[0].hits) + if len(fragment_hits) == 1: + anchor = fragment_hits[0]['fragment'] + frag = Fragment.objects.get(anchor=anchor) return HttpResponseRedirect(frag.get_absolute_url()) + return HttpResponseRedirect(results[0].book.get_absolute_url()) elif len(results) == 0: form = PublishingSuggestForm(initial={"books": query + ", "}) return render_to_response('catalogue/search_no_hits.html', @@ -160,6 +169,9 @@ def main(request): return render_to_response('catalogue/search_multiple_hits.html', {'tags': tag_list, 'prefix': query, - 'results': results, + 'results': { 'author': author_results, + 'title': title_results, + 'content': text_phrase, + 'other': everywhere}, 'did_you_mean': suggestion}, context_instance=RequestContext(request))