X-Git-Url: https://git.mdrn.pl/wolnelektury.git/blobdiff_plain/8b808f24709efb16f6b6eff6abb05b41341573c7..d454277a61e2f34e34ee87777b0d35724edf877e:/apps/search/views.py diff --git a/apps/search/views.py b/apps/search/views.py index 11ce8e8a7..dc9e27b51 100644 --- a/apps/search/views.py +++ b/apps/search/views.py @@ -1,9 +1,10 @@ # -*- coding: utf-8 -*- - +# This file is part of Wolnelektury, licensed under GNU Affero GPLv3 or later. +# Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information. +# 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 from django.views.decorators import cache from django.http import HttpResponse, HttpResponseRedirect, Http404, HttpResponsePermanentRedirect from django.utils.translation import ugettext as _ @@ -13,9 +14,7 @@ from catalogue.models import Book, Tag, Fragment from pdcounter.models import Author as PDCounterAuthor, BookStub as PDCounterBook from catalogue.views import JSONResponse from search import Search, SearchResult -from lucene import StringReader from suggest.forms import PublishingSuggestForm -from time import sleep import re #import enchant import json @@ -28,6 +27,13 @@ def match_word_re(word): return "[[:<:]]%s[[:>:]]" % word +query_syntax_chars = re.compile(r"[\\/*:(){}]") + + +def remove_query_syntax_chars(query, replace=' '): + return query_syntax_chars.sub(' ', query) + + def did_you_mean(query, tokens): return query # change = {} @@ -59,6 +65,8 @@ def hint(request): if len(prefix) < 2: return JSONResponse([]) + prefix = remove_query_syntax_chars(prefix) + search = Search() # tagi beda ograniczac tutaj # ale tagi moga byc na ksiazce i na fragmentach @@ -114,22 +122,30 @@ def main(request): return render_to_response('catalogue/search_too_short.html', {'prefix': query}, context_instance=RequestContext(request)) + + query = remove_query_syntax_chars(query) + search = Search() + theme_terms = search.index.analyze(text=query, field="themes_pl") \ + + search.index.analyze(text=query, field="themes") + # change hints tags = search.hint_tags(query, pdcounter=True, prefix=False) tags = split_tags(tags) author_results = search.search_phrase(query, 'authors', book=True) + translator_results = search.search_phrase(query, 'translators', book=True) + title_results = search.search_phrase(query, 'title', book=True) # 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 = search.search_some(query, ['authors', 'title', 'tags']) + author_title_mixed = search.search_some(query, ['authors', 'translators', 'title', 'tags'], query_terms=theme_terms) author_title_rest = [] for b in author_title_mixed: - also_in_mixed = filter(lambda ba: ba.book_id == b.book_id, author_results + title_results) + also_in_mixed = filter(lambda ba: ba.book_id == b.book_id, author_results + translator_results + title_results) for b2 in also_in_mixed: b2.boost *= 1.1 if also_in_mixed is []: @@ -139,9 +155,9 @@ def main(request): # Because the query is using only one field. text_phrase = SearchResult.aggregate( search.search_phrase(query, 'text', snippets=True, book=False), - search.search_some(query, ['text'], snippets=True, book=False)) + search.search_some(query, ['text'], snippets=True, book=False, query_terms=theme_terms)) - everywhere = search.search_everywhere(query) + everywhere = search.search_everywhere(query, query_terms=theme_terms) def already_found(results): def f(e): @@ -152,30 +168,24 @@ def main(request): return True return False return f - f = already_found(author_results + title_results + text_phrase) + f = already_found(author_results + translator_results + title_results + text_phrase) everywhere = filter(lambda x: not f(x), everywhere) author_results = SearchResult.aggregate(author_results) + translator_results = SearchResult.aggregate(translator_results) title_results = SearchResult.aggregate(title_results) everywhere = SearchResult.aggregate(everywhere, author_title_rest) for field, res in [('authors', author_results), + ('translators', translator_results), ('title', title_results), ('text', text_phrase), ('text', everywhere)]: res.sort(reverse=True) - print "get snips %s, res size %d" % (field, len(res)) for r in res: - print "Get snippets for %s" % r search.get_snippets(r, query, field, 3) - # for r in res: - # for h in r.hits: - # h['snippets'] = map(lambda s: - # re.subn(r"(^[ \t\n]+|[ \t\n]+$)", u"", - # re.subn(r"[ \t\n]*\n[ \t\n]*", u"\n", s)[0])[0], h['snippets']) - # suggestion = did_you_mean(query, search.get_tokens(toks, field="SIMPLE")) suggestion = u'' def ensure_exists(r): @@ -185,22 +195,25 @@ def main(request): return False author_results = filter(ensure_exists, author_results) + translator_results = filter(ensure_exists, translator_results) title_results = filter(ensure_exists, title_results) text_phrase = filter(ensure_exists, text_phrase) everywhere = filter(ensure_exists, everywhere) - results = author_results + title_results + text_phrase + everywhere + results = author_results + translator_results + title_results + text_phrase + everywhere # ensure books do exists & sort them - results.sort(reverse=True) - - if len(results) == 1: - 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(fragment_hits[0]['fragment'].get_absolute_url()) - return HttpResponseRedirect(results[0].book.get_absolute_url()) - elif len(results) == 0: + for res in (author_results, translator_results, title_results, text_phrase, everywhere): + res.sort(reverse=True) + + # We don't want to redirect to book text, but rather display result page even with one result. + # if len(results) == 1: + # 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(fragment_hits[0]['fragment'].get_absolute_url()) + # return HttpResponseRedirect(results[0].book.get_absolute_url()) + if len(results) == 0: form = PublishingSuggestForm(initial={"books": query + ", "}) return render_to_response('catalogue/search_no_hits.html', {'tags': tags, @@ -213,6 +226,7 @@ def main(request): {'tags': tags, 'prefix': query, 'results': {'author': author_results, + 'translator': translator_results, 'title': title_results, 'content': text_phrase, 'other': everywhere},