Added some tests for book imports.
[wolnelektury.git] / apps / catalogue / views.py
index fd0174f..e79e5d4 100644 (file)
@@ -7,6 +7,7 @@ import zipfile
 import sys
 import pprint
 import traceback
+import re
 
 from django.conf import settings
 from django.template import RequestContext
@@ -22,7 +23,10 @@ from django.contrib.auth.forms import UserCreationForm, AuthenticationForm
 from django.utils import simplejson
 from django.utils.functional import Promise
 from django.utils.encoding import force_unicode
+from django.utils.http import urlquote_plus
 from django.views.decorators import cache
+from django.utils.translation import ugettext as _
+from django.views.generic.list_detail import object_list
 
 from catalogue import models
 from catalogue import forms
@@ -40,7 +44,7 @@ class LazyEncoder(simplejson.JSONEncoder):
         return obj
 
 
-def main_page(request):    
+def main_page(request):
     if request.user.is_authenticated():
         shelves = models.Tag.objects.filter(category='set', user=request.user)
         new_set_form = forms.NewSetForm()
@@ -49,7 +53,7 @@ def main_page(request):
     fragment_tags = models.Tag.objects.usage_for_model(models.Fragment, counts=True,
         extra={'where': ["catalogue_tag.category = 'theme'"] + [extra_where]})
     categories = split_tags(tags)
-    
+
     form = forms.SearchForm()
     return render_to_response('catalogue/main_page.html', locals(),
         context_instance=RequestContext(request))
@@ -58,11 +62,11 @@ def main_page(request):
 def book_list(request):
     books = models.Book.objects.all()
     form = forms.SearchForm()
-    
+
     books_by_first_letter = SortedDict()
     for book in books:
         books_by_first_letter.setdefault(book.title[0], []).append(book)
-    
+
     return render_to_response('catalogue/book_list.html', locals(),
         context_instance=RequestContext(request))
 
@@ -71,47 +75,84 @@ def tagged_object_list(request, tags=''):
     # Prevent DoS attacks on our database
     if len(tags.split('/')) > 6:
         raise Http404
-        
+
     try:
         tags = models.Tag.get_tag_list(tags)
     except models.Tag.DoesNotExist:
         raise Http404
-    
+
     if len([tag for tag in tags if tag.category == 'book']):
         raise Http404
-    
-    model = models.Book
-    shelf = [tag for tag in tags if tag.category == 'set']
-    shelf_is_set = (len(tags) == 1 and tags[0].category == 'set')
-    theme_is_set = len([tag for tag in tags if tag.category == 'theme']) > 0
-    if theme_is_set:
-        model = models.Fragment
-    only_author = len(tags) == 1 and tags[0].category == 'author'
-    pd_counter = only_author and tags[0].goes_to_pd()
 
-    user_is_owner = (len(shelf) and request.user.is_authenticated() and request.user == shelf[0].user)
-    
-    extra_where = "catalogue_tag.category NOT IN ('set', 'book')"
-    related_tags = models.Tag.objects.related_for_model(tags, model, counts=True, extra={'where': [extra_where]})
-    categories = split_tags(related_tags)
+    theme_is_set = [tag for tag in tags if tag.category == 'theme']
+    shelf_is_set = len(tags) == 1 and tags[0].category == 'set'
+    my_shelf_is_set = shelf_is_set and request.user.is_authenticated() and request.user == tags[0].user
 
-    if not (theme_is_set or shelf_is_set):
-        model=models.Book.objects.filter(parent=None)
-    
-    return newtagging_views.tagged_object_list(
+    objects = only_author = pd_counter = categories = None
+
+    if theme_is_set:
+        shelf_tags = [tag for tag in tags if tag.category == 'set']
+        fragment_tags = [tag for tag in tags if tag.category != 'set']
+        fragments = models.Fragment.tagged.with_all(fragment_tags)
+
+        if shelf_tags:
+            books = models.Book.tagged.with_all(shelf_tags).order_by()
+            l_tags = [models.Tag.objects.get(slug='l-' + book.slug) for book in books]
+            fragments = models.Fragment.tagged.with_any(l_tags, fragments)
+
+        # newtagging goes crazy if we just try:
+        #related_tags = models.Tag.objects.usage_for_queryset(fragments, counts=True, 
+        #                    extra={'where': ["catalogue_tag.category != 'book'"]})
+        fragment_keys = [fragment.pk for fragment in fragments]
+        if fragment_keys:
+            related_tags = models.Fragment.tags.usage(counts=True,
+                                filters={'pk__in': fragment_keys},
+                                extra={'where': ["catalogue_tag.category != 'book'"]})
+            related_tags = (tag for tag in related_tags if tag not in fragment_tags)
+            categories = split_tags(related_tags)
+
+            objects = fragments
+    else:
+        books = models.Book.tagged.with_all(tags).order_by()
+        l_tags = [models.Tag.objects.get(slug='l-' + book.slug) for book in books]
+        book_keys = [book.pk for book in books]
+        # newtagging goes crazy if we just try:
+        #related_tags = models.Tag.objects.usage_for_queryset(books, counts=True, 
+        #                    extra={'where': ["catalogue_tag.category NOT IN ('set', 'book', 'theme')"]})
+        if book_keys:
+            related_tags = models.Book.tags.usage(counts=True,
+                                filters={'pk__in': book_keys},
+                                extra={'where': ["catalogue_tag.category NOT IN ('set', 'book', 'theme')"]})
+            categories = split_tags(related_tags)
+
+            fragment_keys = [fragment.pk for fragment in models.Fragment.tagged.with_any(l_tags)]
+            if fragment_keys:
+                categories['theme'] = models.Fragment.tags.usage(counts=True,
+                                    filters={'pk__in': fragment_keys},
+                                    extra={'where': ["catalogue_tag.category = 'theme'"]})
+
+            books = books.exclude(parent__in=book_keys)
+            objects = books
+
+    if not objects:
+        only_author = len(tags) == 1 and tags[0].category == 'author'
+        pd_counter = only_author and tags[0].goes_to_pd()
+        objects = models.Book.objects.none()
+
+    return object_list(
         request,
-        tag_model=models.Tag,
-        queryset_or_model=model,
-        tags=tags,
+        objects,
         template_name='catalogue/tagged_object_list.html',
-        extra_context = {
+        extra_context={
             'categories': categories,
             'shelf_is_set': shelf_is_set,
             'only_author': only_author,
             'pd_counter': pd_counter,
-            'user_is_owner': user_is_owner,
+            'user_is_owner': my_shelf_is_set,
             'formats_form': forms.DownloadFormatsForm(),
-        },
+
+            'tags': tags,
+        }
     )
 
 
@@ -120,7 +161,7 @@ def book_fragments(request, book_slug, theme_slug):
     book_tag = get_object_or_404(models.Tag, slug='l-' + book_slug)
     theme = get_object_or_404(models.Tag, slug=theme_slug)
     fragments = models.Fragment.tagged.with_all([book_tag, theme])
-    
+
     form = forms.SearchForm()
     return render_to_response('catalogue/book_fragments.html', locals(),
         context_instance=RequestContext(request))
@@ -132,14 +173,14 @@ def book_detail(request, slug):
     except models.Book.DoesNotExist:
         return book_stub_detail(request, slug)
 
-    book_tag = get_object_or_404(models.Tag, slug = 'l-' + slug)
+    book_tag = get_object_or_404(models.Tag, slug='l-' + slug)
     tags = list(book.tags.filter(~Q(category='set')))
     categories = split_tags(tags)
     book_children = book.children.all().order_by('parent_number')
     extra_where = "catalogue_tag.category = 'theme'"
     book_themes = models.Tag.objects.related_for_model(book_tag, models.Fragment, counts=True, extra={'where': [extra_where]})
     extra_info = book.get_extra_info_value()
-    
+
     form = forms.SearchForm()
     return render_to_response('catalogue/book_detail.html', locals(),
         context_instance=RequestContext(request))
@@ -149,10 +190,10 @@ def book_stub_detail(request, slug):
     book = get_object_or_404(models.BookStub, slug=slug)
     pd_counter = book.pd
     form = forms.SearchForm()
-    
+
     return render_to_response('catalogue/book_stub_detail.html', locals(),
         context_instance=RequestContext(request))
-    
+
 
 def book_text(request, slug):
     book = get_object_or_404(models.Book, slug=slug)
@@ -160,7 +201,7 @@ def book_text(request, slug):
     for fragment in book.fragments.all():
         for theme in fragment.tags.filter(category='theme'):
             book_themes.setdefault(theme, []).append(fragment)
-    
+
     book_themes = book_themes.items()
     book_themes.sort(key=lambda s: s[0].sort_key)
     return render_to_response('catalogue/book_text.html', locals(),
@@ -170,64 +211,135 @@ def book_text(request, slug):
 # ==========
 # = Search =
 # ==========
+
+def _no_diacritics_regexp(query):
+    """ returns a regexp for searching for a query without diacritics
+    
+    should be locale-aware """
+    names = {
+        u'a':u'aąĄ', u'c':u'cćĆ', u'e':u'eęĘ', u'l': u'lłŁ', u'n':u'nńŃ', u'o':u'oóÓ', u's':u'sśŚ', u'z':u'zźżŹŻ',
+        u'ą':u'ąĄ', u'ć':u'ćĆ', u'ę':u'ęĘ', u'ł': u'łŁ', u'ń':u'ńŃ', u'ó':u'óÓ', u'ś':u'śŚ', u'ź':u'źŹ', u'ż':u'żŻ'
+        }
+    def repl(m):
+        l = m.group()
+        return u"(%s)" % '|'.join(names[l])
+    return re.sub(u'[%s]' % (u''.join(names.keys())), repl, query)
+
+def unicode_re_escape(query):
+    """ Unicode-friendly version of re.escape """
+    return re.sub('(?u)(\W)', r'\\\1', query)
+
 def _word_starts_with(name, prefix):
-    """returns a Q object gettings models having `name` contain a word
+    """returns a Q object getting models having `name` contain a word
     starting with `prefix`
+    
+    We define word characters as alphanumeric and underscore, like in JS.
+    
+    Works for MySQL, PostgreSQL, Oracle.
+    For SQLite, _sqlite* version is substituted for this.
     """
     kwargs = {}
-    if settings.DATABASE_ENGINE in ('mysql', 'postgresql_psycopg2', 'postgresql'):
-        # we must escape `prefix` so that it only matches literally
-        for special in r'\^$.*+?|(){}[]':
-            prefix = prefix.replace(special, '\\' + special)
-        
-        # we could use a [[:<:]] (word start), 
-        # but we want both `xy` and `(xy` to catch `(xyz)`
-        kwargs['%s__iregex' % name] = u"(^|[^[:alpha:]])%s" % prefix
-    else:
-        # don't know how to do a generic regex
-        # checking for simple icontain instead
-        kwargs['%s__icontains' % name] = prefix
+
+    prefix = _no_diacritics_regexp(unicode_re_escape(prefix))
+    # can't use [[:<:]] (word start), 
+    # but we want both `xy` and `(xy` to catch `(xyz)`
+    kwargs['%s__iregex' % name] = u"(^|[^[:alnum:]_])%s" % prefix
+
+    return Q(**kwargs)
+
+
+def _sqlite_word_starts_with(name, prefix):
+    """ version of _word_starts_with for SQLite 
+    
+    SQLite in Django uses Python re module
+    """
+    kwargs = {}
+    prefix = _no_diacritics_regexp(unicode_re_escape(prefix))
+    kwargs['%s__iregex' % name] = ur"(^|(?<=[^\wąćęłńóśźżĄĆĘŁŃÓŚŹŻ]))%s" % prefix
     return Q(**kwargs)
 
 
-def _tags_starting_with(prefix, user):
+if settings.DATABASE_ENGINE == 'sqlite3':
+    _word_starts_with = _sqlite_word_starts_with
+
+
+def _tags_starting_with(prefix, user=None):
+    prefix = prefix.lower()
     book_stubs = models.BookStub.objects.filter(_word_starts_with('title', prefix))
     books = models.Book.objects.filter(_word_starts_with('title', prefix))
     book_stubs = filter(lambda x: x not in books, book_stubs)
     tags = models.Tag.objects.filter(_word_starts_with('name', prefix))
-    if user.is_authenticated():
+    if user and user.is_authenticated():
         tags = tags.filter(~Q(category='book') & (~Q(category='set') | Q(user=user)))
     else:
         tags = tags.filter(~Q(category='book') & ~Q(category='set'))
 
     return list(books) + list(tags) + list(book_stubs)
-        
+
+
+def _get_result_link(match, tag_list):
+    if isinstance(match, models.Book) or isinstance(match, models.BookStub):
+        return match.get_absolute_url()
+    else:
+        return reverse('catalogue.views.tagged_object_list',
+            kwargs={'tags': '/'.join(tag.slug for tag in tag_list + [match])}
+        )
+
+def _get_result_type(match):
+    if isinstance(match, models.Book) or isinstance(match, models.BookStub):
+        type = 'book'
+    else:
+        type = match.category
+    return dict(models.TAG_CATEGORIES)[type]
+
+
+
+def find_best_matches(query, user=None):
+    """ Finds a Book, Tag or Bookstub best matching a query.
+    
+    Returns a with:
+      - zero elements when nothing is found,
+      - one element when a best result is found,
+      - more then one element on multiple exact matches
+    
+    Raises a ValueError on too short a query.
+    """
+
+    query = query.lower()
+    if len(query) < 2:
+        raise ValueError("query must have at least two characters")
+
+    result = tuple(_tags_starting_with(query, user))
+    exact_matches = tuple(res for res in result if res.name.lower() == query)
+    if exact_matches:
+        return exact_matches
+    else:
+        return result[:1]
+
 
 def search(request):
     tags = request.GET.get('tags', '')
     prefix = request.GET.get('q', '')
-    # Prefix must have at least 2 characters
-    if len(prefix) < 2:
-        return HttpResponse('')
-    
+
     try:
         tag_list = models.Tag.get_tag_list(tags)
     except:
         tag_list = []
-    
-    result = _tags_starting_with(prefix, request.user)
-    if len(result) > 0:
-        tag = result[0]
-        if isinstance(tag, models.Book) or isinstance(tag, models.BookStub):
-            return HttpResponseRedirect(tag.get_absolute_url())
-        else:
-            tag_list.append(tag)
-            
-            return HttpResponseRedirect(reverse('catalogue.views.tagged_object_list', 
-                kwargs={'tags': '/'.join(tag.slug for tag in tag_list)}
-            ))
+
+    try:
+        result = find_best_matches(prefix, request.user)
+    except ValueError:
+        return render_to_response('catalogue/search_too_short.html', {'tags':tag_list, 'prefix':prefix},
+            context_instance=RequestContext(request))
+
+    if len(result) == 1:
+        return HttpResponseRedirect(_get_result_link(result[0], tag_list))
+    elif len(result) > 1:
+        return render_to_response('catalogue/search_multiple_hits.html',
+            {'tags':tag_list, 'prefix':prefix, 'results':((x, _get_result_link(x, tag_list), _get_result_type(x)) for x in result)},
+            context_instance=RequestContext(request))
     else:
-        return render_to_response('catalogue/search_no_hits.html', {'query':prefix, 'tags':tag_list},
+        return render_to_response('catalogue/search_no_hits.html', {'tags':tag_list, 'prefix':prefix},
             context_instance=RequestContext(request))
 
 
@@ -236,7 +348,7 @@ def tags_starting_with(request):
     # Prefix must have at least 2 characters
     if len(prefix) < 2:
         return HttpResponse('')
-    
+
     return HttpResponse('\n'.join(tag.name for tag in _tags_starting_with(prefix, request.user)))
 
 
@@ -256,24 +368,24 @@ def book_sets(request, slug):
     book = get_object_or_404(models.Book, slug=slug)
     user_sets = models.Tag.objects.filter(category='set', user=request.user)
     book_sets = book.tags.filter(category='set', user=request.user)
-    
+
     if not request.user.is_authenticated():
         return HttpResponse(_('<p>To maintain your shelves you need to be logged in.</p>'))
-    
+
     if request.method == 'POST':
         form = forms.ObjectSetsForm(book, request.user, request.POST)
         if form.is_valid():
             old_shelves = list(book.tags.filter(category='set'))
             new_shelves = [models.Tag.objects.get(pk=id) for id in form.cleaned_data['set_ids']]
-            
+
             for shelf in [shelf for shelf in old_shelves if shelf not in new_shelves]:
                 shelf.book_count -= 1
                 shelf.save()
-                
+
             for shelf in [shelf for shelf in new_shelves if shelf not in old_shelves]:
                 shelf.book_count += 1
                 shelf.save()
-            
+
             book.tags = new_shelves + list(book.tags.filter(~Q(category='set') | ~Q(user=request.user)))
             if request.is_ajax():
                 return HttpResponse(_('<p>Shelves were sucessfully saved.</p>'))
@@ -282,7 +394,7 @@ def book_sets(request, slug):
     else:
         form = forms.ObjectSetsForm(book, request.user)
         new_set_form = forms.NewSetForm()
-    
+
     return render_to_response('catalogue/book_sets.html', locals(),
         context_instance=RequestContext(request))
 
@@ -293,7 +405,7 @@ def book_sets(request, slug):
 def remove_from_shelf(request, shelf, book):
     book = get_object_or_404(models.Book, slug=book)
     shelf = get_object_or_404(models.Tag, slug=shelf, category='set', user=request.user)
-    
+
     if shelf in book.tags:
         models.Tag.objects.remove_tag(book, shelf)
 
@@ -326,18 +438,18 @@ def download_shelf(request, slug):
     be used for large dynamic PDF files.                                        
     """
     shelf = get_object_or_404(models.Tag, slug=slug, category='set')
-    
+
     formats = []
     form = forms.DownloadFormatsForm(request.GET)
     if form.is_valid():
         formats = form.cleaned_data['formats']
     if len(formats) == 0:
         formats = ['pdf', 'odt', 'txt', 'mp3', 'ogg']
-    
+
     # Create a ZIP archive
     temp = tempfile.TemporaryFile()
     archive = zipfile.ZipFile(temp, 'w')
-    
+
     for book in collect_books(models.Book.tagged.with_all(shelf)):
         if 'pdf' in formats and book.pdf_file:
             filename = book.pdf_file.path
@@ -355,11 +467,11 @@ def download_shelf(request, slug):
             filename = book.ogg_file.path
             archive.write(filename, str('%s.ogg' % book.slug))
     archive.close()
-    
+
     response = HttpResponse(content_type='application/zip', mimetype='application/x-zip-compressed')
     response['Content-Disposition'] = 'attachment; filename=%s.zip' % shelf.sort_key
     response['Content-Length'] = temp.tell()
-    
+
     temp.seek(0)
     response.write(temp.read())
     return response
@@ -373,7 +485,7 @@ def shelf_book_formats(request, shelf):
     shelf = get_object_or_404(models.Tag, slug=shelf, category='set')
 
     formats = {'pdf': False, 'odt': False, 'txt': False, 'mp3': False, 'ogg': False}
-    
+
     for book in collect_books(models.Book.tagged.with_all(shelf)):
         if book.pdf_file:
             formats['pdf'] = True
@@ -440,7 +552,7 @@ def register(request):
     if registration_form.is_valid():
         user = registration_form.save()
         user = auth.authenticate(
-            username=registration_form.cleaned_data['username'], 
+            username=registration_form.cleaned_data['username'],
             password=registration_form.cleaned_data['password1']
         )
         auth.login(request, user)
@@ -453,7 +565,7 @@ def register(request):
 @cache.never_cache
 def logout_then_redirect(request):
     auth.logout(request)
-    return HttpResponseRedirect(request.GET.get('next', '/'))
+    return HttpResponseRedirect(urlquote_plus(request.GET.get('next', '/'), safe='/?='))
 
 
 
@@ -472,7 +584,8 @@ def import_book(request):
             info = sys.exc_info()
             exception = pprint.pformat(info[1])
             tb = '\n'.join(traceback.format_tb(info[2]))
-            return HttpResponse(_("An error occurred: %s\n\n%s") % (exception, tb), mimetype='text/plain')
+            _('Today is %(month)s, %(day)s.') % {'month': m, 'day': d}
+            return HttpResponse(_("An error occurred: %(exception)s\n\n%(tb)s") % {'exception':exception, 'tb':tb}, mimetype='text/plain')
         return HttpResponse(_("Book imported successfully"))
     else:
         return HttpResponse(_("Error importing file: %r") % book_import_form.errors)
@@ -484,4 +597,4 @@ def clock(request):
     in a format suitable for Date.parse()
     """
     from datetime import datetime
-    return HttpResponse(datetime.now().strftime('%Y/%m/%d %H:%M:%S'))
\ No newline at end of file
+    return HttpResponse(datetime.now().strftime('%Y/%m/%d %H:%M:%S'))