fix
[wolnelektury.git] / src / search / forms.py
old mode 100755 (executable)
new mode 100644 (file)
index d62ab64..176c73e
@@ -1,10 +1,17 @@
 # This file is part of Wolnelektury, licensed under GNU Affero GPLv3 or later.
 # Copyright © Fundacja Nowoczesna Polska. See NOTICE for more information.
 #
+from django.apps import apps
+from django.conf import settings
+from django.contrib.postgres.search import SearchHeadline, SearchQuery
 from django import forms
-from django.utils.translation import ugettext_lazy as _
-
-from search.fields import JQueryAutoCompleteSearchField
+from django.utils.translation import gettext_lazy as _
+from catalogue.constants import LANGUAGES_3TO2
+import catalogue.models
+import pdcounter.models
+import picture.models
+from .fields import JQueryAutoCompleteSearchField, InlineRadioWidget
+from .utils import UnaccentSearchQuery, UnaccentSearchVector
 
 
 class SearchForm(forms.Form):
@@ -18,4 +25,180 @@ class SearchForm(forms.Form):
         self.fields['q'].widget.attrs['autocomplete'] = 'off'
         self.fields['q'].widget.attrs['data-source'] = source
         if 'q' not in self.data:
-            self.fields['q'].widget.attrs['placeholder'] = _('title, author, theme/topic, epoch, kind, genre, phrase')
+            self.fields['q'].widget.attrs['placeholder'] = _('title, author, epoch, kind, genre, phrase')
+
+
+class SearchFilters(forms.Form):
+    q = forms.CharField(
+        required=False, widget=forms.HiddenInput(),
+        min_length=2, max_length=256,
+    )
+    format = forms.ChoiceField(required=False, choices=[
+        ('', 'wszystkie'),
+        ('text', 'tekst'),
+        ('audio', 'audiobook'),
+        ('daisy', 'Daisy'),
+        ('art', 'obraz'),
+        #('theme', 'motywy'),
+    ], widget=InlineRadioWidget())
+    lang = forms.ChoiceField(required=False)
+    epoch = forms.ChoiceField(required=False)
+    genre = forms.ChoiceField(required=False)
+    category = forms.ChoiceField(required=False, choices=[
+        ('', 'wszystkie'),
+        ('author', 'autor'),
+        #('translator', 'tłumacz'),
+        ('theme', 'motyw'),
+        ('genre', 'gatunek'),
+        ('book', 'tytuł'),
+        ('art', 'obraz'),
+        ('collection', 'kolekcja'),
+        ('quote', 'cytat'),
+    ], widget=InlineRadioWidget())
+
+    def __init__(self, *args, **kwargs):
+        super().__init__(*args, **kwargs)
+
+        langs = dict(settings.LANGUAGES)
+        self.fields['lang'].choices = [('', 'wszystkie')] + [
+            (
+                b,
+                langs.get(LANGUAGES_3TO2.get(b, b), b)
+            )
+            for b in catalogue.models.Book.objects.values_list(
+                    'language', flat=True
+            ).distinct().order_by()
+        ]
+        self.fields['epoch'].choices = [('', 'wszystkie')] + [
+            (b.slug, b.name)
+            for b in catalogue.models.Tag.objects.filter(category='epoch')
+        ]
+        self.fields['genre'].choices = [('', 'wszystkie')] + [
+            (b.slug, b.name)
+            for b in catalogue.models.Tag.objects.filter(category='genre')
+        ]
+
+    def get_querysets(self):
+        qs = {
+            'author': catalogue.models.Tag.objects.filter(category='author'),
+            'pdauthor': pdcounter.models.Author.objects.all(),
+            'theme': catalogue.models.Tag.objects.filter(category='theme'),
+            'genre': catalogue.models.Tag.objects.filter(category='genre'),
+            'collection': catalogue.models.Collection.objects.all(),
+            'book': catalogue.models.Book.objects.filter(findable=True),
+            'pdbook': pdcounter.models.BookStub.objects.all(),
+            'snippet': catalogue.models.Snippet.objects.filter(book__findable=True),
+            'art': picture.models.Picture.objects.all(),
+            # art pieces
+        }
+        if self.cleaned_data['category']:
+            c = self.cleaned_data['category']
+            if c != 'author':
+                qs['author'] = qs['author'].none()
+                qs['pdauthor'] = qs['pdauthor'].none()
+            if c != 'theme': qs['theme'] = qs['theme'].none()
+            if c != 'genre': qs['genre'] = qs['genre'].none()
+            if c != 'collection': qs['collection'] = qs['collection'].none()
+            if c != 'book':
+                qs['book'] = qs['book'].none()
+                qs['pdbook'] = qs['pdbook'].none()
+            if c != 'quote': qs['snippet'] = qs['snippet'].none()
+            if c != 'art': qs['art'] = qs['art'].none()
+            qs['art'] = picture.models.Picture.objects.none()
+
+        if self.cleaned_data['format']:
+            c = self.cleaned_data['format']
+            qs['author'] = qs['author'].none()
+            qs['pdauthor'] = qs['pdauthor'].none()
+            qs['theme'] = qs['theme'].none()
+            qs['genre'] = qs['genre'].none()
+            qs['collection'] = qs['collection'].none()
+            if c == 'art':
+                qs['book'] = qs['book'].none()
+                qs['pdbook'] = qs['pdbook'].none()
+                qs['snippet'] = qs['snippet'].none()
+            if c in ('text', 'audio', 'daisy'):
+                qs['art'] = qs['art'].none()
+                if c == 'audio':
+                    qs['book'] = qs['book'].filter(media__type='mp3')
+                    qs['pdbook'] = qs['book'].none()
+                    qs['snippet'] = qs['snippet'].filter(book__media__type='mp3')
+                elif c == 'daisy':
+                    qs['book'] = qs['book'].filter(media__type='daisy')
+                    qs['snippet'] = qs['snippet'].filter(book__media__type='daisy')
+
+        if self.cleaned_data['lang']:
+            qs['author'] = qs['author'].none()
+            qs['pdauthor'] = qs['pdauthor'].none()
+            qs['theme'] = qs['theme'].none()
+            qs['genre'] = qs['genre'].none()
+            qs['art'] = qs['art'].none()
+            qs['collection'] = qs['collection'].none()
+            qs['book'] = qs['book'].filter(language=self.cleaned_data['lang'])
+            qs['pdbook'] = qs['pdbook'].none()
+            qs['snippet'] = qs['snippet'].filter(book__language=self.cleaned_data['lang'])
+
+        for tag_cat in ('epoch', 'genre'):
+            c = self.cleaned_data[tag_cat]
+            if c:
+                # FIXME nonexistent
+                t = catalogue.models.Tag.objects.get(category=tag_cat, slug=c)
+                qs['author'] = qs['author'].none()
+                qs['pdauthor'] = qs['pdauthor'].none()
+                qs['theme'] = qs['theme'].none()
+                qs['genre'] = qs['genre'].none()
+                qs['collection'] = qs['collection'].none()
+                qs['book'] = qs['book'].filter(tag_relations__tag=t)
+                qs['pdbook'] = qs['pdbook'].none()
+                qs['snippet'] = qs['snippet'].filter(book__tag_relations__tag=t)
+                qs['art'] = qs['art'].filter(tag_relations__tag=t)
+            
+        return qs
+
+    def results(self):
+        qs = self.get_querysets()
+        query = self.cleaned_data['q']
+        squery = UnaccentSearchQuery(query, config='polish')
+        query = SearchQuery(query, config='polish')
+        books = qs['book'].annotate(
+            search_vector=UnaccentSearchVector('title')
+        ).filter(search_vector=squery)
+        books = books.exclude(ancestor__in=books).order_by('-popularity__count')
+
+        snippets = qs['snippet'].filter(search_vector=squery).annotate(
+                    headline=SearchHeadline(
+                        'text',
+                        query,
+                        config='polish',
+                        start_sel='<strong>',
+                        stop_sel='</strong>',
+                    )
+                ).order_by('-book__popularity__count', 'sec')[:100]
+        snippets_by_book = {}
+        for snippet in snippets:
+            snippet_list = snippets_by_book.setdefault(snippet.book, [])
+            if len(snippet_list) < 3:
+                snippet_list.append(snippet)
+
+        return {
+            'author': qs['author'].annotate(
+                search_vector=UnaccentSearchVector('name_pl')
+            ).filter(search_vector=squery),
+            'theme': qs['theme'].annotate(
+                search_vector=UnaccentSearchVector('name_pl')
+            ).filter(search_vector=squery),
+            'genre': qs['genre'].annotate(
+                search_vector=UnaccentSearchVector('name_pl')
+            ).filter(search_vector=squery),
+            'collection': qs['collection'].annotate(
+                search_vector=UnaccentSearchVector('title')
+            ).filter(search_vector=squery),
+            'book': books[:100],
+            'art': qs['art'].annotate(
+                search_vector=UnaccentSearchVector('title')
+            ).filter(search_vector=squery)[:100],
+            'snippet': snippets_by_book,
+            'pdauthor': pdcounter.models.Author.search(squery, qs=qs['pdauthor']),
+            'pdbook': pdcounter.models.BookStub.search(squery, qs=qs['pdbook']),
+        }
+