Add Haystack search plugin (NB! Whoosh backend is broken upstream)

.gitignore

@@ -4,6 +4,7 @@ dist
 *.egg-info
 README.rst
 testproject/testproject/whoosh_index/
+testproject/testproject/xapian_index/
 
 #Docs
 docs/_build

testproject/testproject/settings/haystack.py

@@ -13,9 +13,22 @@ HAYSTACK_CONNECTIONS = {
         'ENGINE': 'haystack.backends.simple_backend.SimpleEngine',
     },
 }
+
+
 HAYSTACK_CONNECTIONS = {
     'default': {
-        'ENGINE': 'haystack.backends.whoosh_backend.WhooshEngine',
-        'PATH': os.path.join(PROJECT_PATH, 'whoosh_index'),
+        'ENGINE': 'xapian_backend.XapianEngine',
+        'PATH': os.path.join(PROJECT_PATH, 'xapian_index'),
     },
 }
+
+# Whoosh backend is completely broken
+# https://github.com/toastdriven/django-haystack/issues/522
+# https://github.com/toastdriven/django-haystack/issues/382
+# https://github.com/toastdriven/django-haystack/issues/447
+#HAYSTACK_CONNECTIONS = {
+#    'default': {
+#        'ENGINE': 'haystack.backends.whoosh_backend.WhooshEngine',
+#        'PATH': os.path.join(PROJECT_PATH, 'whoosh_index'),
+#    },
+#}

wiki/plugins/haystack/README.md

+Haystack plugin
+===============
+
+***July 16 2013***
+Tested backend, xapian with Haystack 2.0. NB! Get the xapian from Git
+
+https://github.com/notanumber/xapian-haystack
+
+Not working
+-----------
+
+Do not use Whoosh, it has broken SearchQuerySet support and therefore will leak
+articles that are set non-public.

wiki/plugins/haystack/__init__.py

+from wiki.core import settings
+from wiki.core import permissions
+from wiki import models
+
+from haystack import views as haystack_views
+from django.db.models import Q
+from django.utils.decorators import classonlymethod
+from django.shortcuts import redirect
+
+class SearchViewHaystack(haystack_views.SearchView):
+    results_per_page = 25
+    template = "search/search.html"
+
+    def __name__(self): #@ReservedAssignment
+        return "SearchViewHaystack"
+
+    @classonlymethod
+    def as_view(cls,*args,**kwargs):
+        return haystack_views.search_view_factory(view_class=cls,*args,**kwargs)
+
+    def dispatch(self, request, *args, **kwargs):
+        # Do not allow anonymous users to search if they cannot read content
+        if request.user.is_anonymous() and not settings.ANONYMOUS:
+            return redirect(settings.LOGIN_URL)
+        return super(SearchViewHaystack, self).dispatch(request, *args, **kwargs)
+
+    def __filter_can_read(self, user):
+        """Filter objects so only the ones with a user's reading access
+         are included"""
+        if user.has_perm('wiki.moderator'):
+            return self.results
+        if user.is_anonymous():
+            q = self.results.filter(other_read='True')
+            return q 
+        else:
+            q = self.results.filter(
+                Q(other_read=True) |
+                Q(owner=user) |
+                (Q(group__user=user) & Q(group_read=True))
+            )
+        return q
+
+
+    def __call__(self, request):
+        self.request = request
+        if self.request.user.is_anonymous and not settings.ANONYMOUS:
+            return redirect(settings.LOGIN_URL)
+
+        self.form = self.build_form()
+        self.query = self.get_query()
+        self.results = self.get_results()
+        if not permissions.can_moderate(models.URLPath.root().article, self.request.user):
+            self.results = self.__filter_can_read(self.request.user)         
+            #self.results = self.results.filter(current_revision__deleted=False)
+          
+        return self.create_response()    
+
+    def extra_context(self):
+        extra = super(SearchViewHaystack, self).extra_context()
+        extra['search_query'] = self.query
+        return extra
+
+

wiki/plugins/haystack/search_indexes.py

+from haystack import indexes
+from wiki import models 
+
+
+class ArticleIndex(indexes.SearchIndex, indexes.Indexable):
+    text = indexes.CharField(document=True, use_template=True)
+    created = indexes.DateTimeField(model_attr='created')
+    modified = indexes.DateTimeField(model_attr='modified')
+
+    #default because indexing fails with whoosh. see. 
+    #http://stackoverflow.com/questions/11995367/how-do-i-use-a-boolean-field-in-django-haystack-search-query
+    #https://github.com/toastdriven/django-haystack/issues/382
+    other_read = indexes.BooleanField(model_attr='other_read',default=False)
+    group_read = indexes.BooleanField(model_attr='group_read',default=False)
+
+    owner_id = indexes.IntegerField(model_attr='owner__id')
+    group_id = indexes.IntegerField(model_attr='group__id')
+    
+
+    def get_model(self):
+        return models.Article
+
+    def index_queryset(self, using=None):
+        """Used when the entire index for model is updated."""
+        return self.get_model().objects.all()

wiki/plugins/haystack/templates/search/indexes/wiki/article_text.txt

+{{ object.current_revision.title }}
+{{ object.current_revision.content }}

wiki/plugins/haystack/templates/wiki/plugins/haystack/search.html

+{% extends "wiki/search.html" %}
+
+{% block wiki_search_loop %}
+{% with article.object as article %}
+{% include "wiki/includes/searchresult.html" %}
+{% endwith %}
+{% endblock %}
+

wiki/plugins/haystack/views.py

+from wiki.views.article import SearchView
+from haystack.query import SearchQuerySet
+from haystack.inputs import AutoQuery
+from wiki.core import permissions
+from wiki import models
+
+class HaystackSearchView(SearchView):
+    
+    template_name = 'wiki/plugins/haystack/search.html'
+    
+    def get_queryset(self):
+        qs = SearchQuerySet().all()
+        if self.request.user.is_authenticated():
+            # TODO: This has a leak! It should say:
+            # group=self.request.group.id AND group_read=True
+            if not permissions.can_moderate(models.URLPath.root().article, 
+                self.request.user):
+                qs = qs.filter_or(
+                    owner=self.request.user.id, 
+                    group=self.request.group.id,
+                    other_read=True
+                )
+        else:
+            qs = qs.exclude(other_read=False)
+        
+        qs = qs.filter(content=AutoQuery(self.query))
+        qs = qs.exclude(other_read=False)
+        qs = qs.load_all()
+        return qs
+    
\ No newline at end of file

wiki/templates/wiki/includes/searchresult.html

+{% load wiki_tags i18n humanize %}
+{% load url from future %}
+
+<tr>
+  <td>
+    {% for urlpath in article.urlpath_set.all %}
+      <a href="{% url 'wiki:get' path=urlpath.path %}">{{ article.current_revision.title }}<br /><small class="muted">/{{ urlpath.path }}</small></a>
+    {% empty %}
+      <a href="{% url 'wiki:get' article_id=article.id %}">{{ article.current_revision.title }}</a>
+    {% endfor %}
+    {% if article.current_revision.deleted %}
+      <span class="icon-trash"></span>
+    {% endif %}
+    {% if article.current_revision.locked %}
+      <span class="icon-lock"></span>
+    {% endif %}
+    <p class="muted"><small>{{ article.current_revision.content|get_content_snippet:search_query }}</small></p>
+  </td>
+  <td style="white-space: nowrap">
+    {{ article.current_revision.created|naturaltime }}
+  </td>
+</tr>

wiki/templates/wiki/search.html

@@ -11,7 +11,7 @@
 <form class="form-search directory-toolbar">
 <p class="lead">
   <div class="pull-right">
-    {{ search_form.query }}
+    {{ search_form.q }}
     <button class="btn"><span class="icon-search"></span></button>
   </div>
   <p>{% blocktrans with paginator.object_list.count as cnt %}Your search returned <strong>{{ cnt }}</strong> results.{% endblocktrans %}</p>
@@ -25,25 +25,9 @@
     <th>{% trans "Last modified" %}</th>
   </tr>
   {% for article in articles %}
-    <tr>
-      <td>
-        {% for urlpath in article.urlpath_set.all %}
-          <a href="{% url 'wiki:get' path=urlpath.path %}">{{ article.current_revision.title }}<br /><small class="muted">/{{ urlpath.path }}</small></a>
-        {% empty %}
-          <a href="{% url 'wiki:get' article_id=article.id %}">{{ article.current_revision.title }}</a>
-        {% endfor %}
-        {% if article.current_revision.deleted %}
-          <span class="icon-trash"></span>
-        {% endif %}
-        {% if article.current_revision.locked %}
-          <span class="icon-lock"></span>
-        {% endif %}
-        <p class="muted"><small>{{ article.current_revision.content|get_content_snippet:search_query }}</small></p>
-      </td>
-      <td style="white-space: nowrap">
-        {{ article.current_revision.created|naturaltime }}
-      </td>
-    </tr>
+    {% block wiki_search_loop %}
+    {% include "wiki/includes/searchresult.html" %}
+    {% endblock %}
     {% empty%}
     <tr>
       <td colspan="100">

wiki/views/article.py

@@ -11,7 +11,6 @@ from django.utils.translation import ugettext as _
 from django.views.generic.base import TemplateView, View
 from django.views.generic.edit import FormView
 from django.views.generic.list import ListView
-from django.utils.decorators import classonlymethod
 
 from wiki.views.mixins import ArticleMixin
 from wiki import editors, forms, models
@@ -24,7 +23,7 @@ from django.db import transaction
 from wiki.core.exceptions import NoRootURL
 from wiki.core import permissions
 from django.http import Http404
-from haystack import views as haystack_views
+
 
 class ArticleView(ArticleMixin, TemplateView):
 
@@ -465,58 +464,6 @@ class Dir(ListView, ArticleMixin):
 
         return kwargs
 
-class SearchViewHaystack(haystack_views.SearchView):
-    results_per_page = 25
-    template = "search/search.html"
-
-    def __name__(self):
-        return "SearchViewHaystack"
-
-    def dispatch(self, request, *args, **kwargs):
-        # Do not allow anonymous users to search if they cannot read content
-        if request.user.is_anonymous() and not settings.ANONYMOUS:
-            return redirect(settings.LOGIN_URL)
-        return super(SearchViewHaystack, self).dispatch(request, *args, **kwargs)
-
-    @classonlymethod
-    def as_view(cls,*args,**kwargs):
-        return haystack_views.search_view_factory(view_class=cls,*args,**kwargs)
-
-    def __filter_can_read(self, user):
-        """Filter objects so only the ones with a user's reading access
-         are included"""
-        if user.has_perm('wiki.moderator'):
-            return self.results
-        if user.is_anonymous():
-            q = self.results.filter(other_read='True')
-            return q 
-        else:
-           q = self.results.filter(Q(other_read=True) |
-                        Q(owner=user) |
-                        (Q(group__user=user) & Q(group_read=True))
-                        )
-        return q
-
-
-    def __call__(self, request):
-        self.request = request
-        if self.request.user.is_anonymous and not settings.ANONYMOUS:
-            return redirect(settings.LOGIN_URL)
-
-        self.form = self.build_form()
-        self.query = self.get_query()
-        self.results = self.get_results()
-        if not permissions.can_moderate(models.URLPath.root().article, self.request.user):
-            self.results = self.__filter_can_read(self.request.user)         
-            #self.results = self.results.filter(current_revision__deleted=False)
-          
-        return self.create_response()    
-
-    def extra_context(self):
-        extra = super(SearchViewHaystack, self).extra_context()
-        extra['search_query'] = self.query
-        return extra
-
 
 class SearchView(ListView):