def get(self, request, format=None):
query = request.GET.get('q', '').strip()
sqs = SearchQuerySet().models(self.model_class)
if query:
sqs = sqs.autocomplete(name=query)
data = [{'id': int(result.pk), 'name': result.name} for result in sqs]
return Response(json.dumps(data))
def test_load_all_read_queryset(self):
# Stow.
old_ui = connections["default"]._index
ui = UnifiedIndex()
gafmmsi = GhettoAFifthMockModelSearchIndex()
ui.build(indexes=[gafmmsi])
connections["default"]._index = ui
gafmmsi.update()
sqs = SearchQuerySet()
results = sqs.load_all().all()
results.query.backend = ReadQuerySetMockSearchBackend("default")
results._fill_cache(0, 2)
# The deleted result isn't returned
self.assertEqual(len([result for result in results._result_cache if result is not None]), 1)
# Register a SearchIndex with a read_queryset that returns deleted items
rqstsi = TextReadQuerySetTestSearchIndex()
ui.build(indexes=[rqstsi])
rqstsi.update()
sqs = SearchQuerySet()
results = sqs.load_all().all()
results.query.backend = ReadQuerySetMockSearchBackend("default")
results._fill_cache(0, 2)
# Both the deleted and not deleted items are returned
self.assertEqual(len([result for result in results._result_cache if result is not None]), 2)
# Restore.
connections["default"]._index = old_ui
def search(self, query=None, *args, **kwargs):
"""
Uses haystack to query news.
Returns a SearchQuerySet
"""
sqs = SearchQuerySet()
user = kwargs.get('user', None)
# check to see if there is impersonation
if hasattr(user,'impersonated_user'):
if isinstance(user.impersonated_user, User):
user = user.impersonated_user
is_an_admin = user.profile.is_superuser
if query:
sqs = sqs.auto_query(sqs.query.clean(query))
if user:
if not is_an_admin:
return []
else:
sqs = sqs.all()
if user:
if not is_an_admin:
return []
return sqs.models(self.model).order_by('-update_dt')
def get(self, request, relation=None, name=None, year=None):
filter_criteria = request.session.get("filter_criteria", {})
filter_criteria.update(date_published__gte=year + "-01-01T00:00:00Z",
date_published__lte=year + "-12-31T00:00:00Z")
if relation == RelationClassifier.JOURNAL.value:
filter_criteria.update(container__name=name)
elif relation == RelationClassifier.SPONSOR.value:
filter_criteria.update(sponsors__name=name)
elif relation == RelationClassifier.PLATFORM.value:
filter_criteria.update(platforms__name=name)
elif relation == RelationClassifier.MODELDOCUMENDTATION.value:
filter_criteria.update(model_documentation__name=name)
elif relation == RelationClassifier.AUTHOR.value:
filter_criteria.update(authors__name__exact=name.replace("/", " "))
sqs = SearchQuerySet()
sqs = sqs.filter(**filter_criteria).models(Publication)
pubs_pk = queryset_gen(sqs)
pubs = Publication.api.primary(pk__in=pubs_pk)
paginator = CatalogPagination()
result_page = paginator.paginate_queryset(pubs, request)
serializer = PublicationListSerializer(result_page, many=True)
response = paginator.get_paginated_response(serializer.data)
return Response({'json': json.dumps(response)}, template_name="publication/list.html")
def test_models(self):
# Stow.
old_unified_index = connections["default"]._index
ui = UnifiedIndex()
bmmsi = BasicMockModelSearchIndex()
bammsi = BasicAnotherMockModelSearchIndex()
ui.build(indexes=[bmmsi, bammsi])
connections["default"]._index = ui
msqs = SearchQuerySet()
sqs = msqs.all()
self.assertTrue(isinstance(sqs, SearchQuerySet))
self.assertEqual(len(sqs.query.models), 0)
sqs = msqs.models(MockModel)
self.assertTrue(isinstance(sqs, SearchQuerySet))
self.assertEqual(len(sqs.query.models), 1)
sqs = msqs.models(MockModel, AnotherMockModel)
self.assertTrue(isinstance(sqs, SearchQuerySet))
self.assertEqual(len(sqs.query.models), 2)
# This will produce a warning.
ui.build(indexes=[bmmsi])
sqs = msqs.models(AnotherMockModel)
self.assertTrue(isinstance(sqs, SearchQuerySet))
self.assertEqual(len(sqs.query.models), 1)
def get_global_context(self, context):
# Find all the models to search over...
models = set(
self.search_sections[section]['model']
for section in self.search_sections
)
show_top_hits = (self.page == '1' or not self.page)
top_hits_ids = []
if show_top_hits:
context['top_hits'] = []
for section, max_for_top_hits in SearchBaseView.top_hits_under.items():
data = self.get_section_data(section)
if data['results_count'] <= max_for_top_hits:
context['top_hits'] += data['results']
top_hits_ids = set(r.id for r in context['top_hits'])
sqs = SearchQuerySet().models(*list(models))
# Exclude anything that will already have been shown in the top hits:
for top_hit_id in top_hits_ids:
sqs = sqs.exclude(id=top_hit_id)
sqs = sqs. \
exclude(hidden=True). \
filter(content=AutoQuery(self.query)). \
highlight()
context['results'] = self.get_paginated_results(sqs)
return context
def autocomplete(request):
"""Return autocomple JSON results"""
term = request.GET.get("term", "").strip()
response_data = []
if len(term):
# Does not work - probably because the FLAG_PARTIAL is not set on Xapian
# (trying to set it in settings.py as documented appears to have no effect)
# sqs = SearchQuerySet().autocomplete(name_auto=term)
# Split the search term up into little bits
terms = re.split(r"\s+", term)
# Build up a query based on the bits
sqs = SearchQuerySet()
for bit in terms:
# print "Adding '%s' to the '%s' query" % (bit,term)
sqs = sqs.filter_and(name_auto__startswith=sqs.query.clean(bit))
# collate the results into json for the autocomplete js
for result in sqs.all()[0:10]:
response_data.append({"url": result.object.get_absolute_url(), "label": result.object.name})
# send back the results as JSON
return HttpResponse(simplejson.dumps(response_data), mimetype="application/json")
def search_filter(self, filters=None, *args, **kwargs):
sqs = SearchQuerySet()
user = kwargs.get('user', None)
groups = []
if user and user.is_authenticated():
groups = [g.pk for g in user.group_set.all()]
admin = user.profile.is_superuser
# permission filters
if user:
if not user.profile.is_superuser:
if not user.is_anonymous():
# (status+status_detail+(anon OR user)) OR (who_can_view__exact)
anon_query = Q(allow_anonymous_view=True)
user_query = Q(allow_user_view=True)
sec1_query = Q(status=True, status_detail='active')
user_perm_q = Q(users_can_view__in=[user.pk])
group_perm_q = Q(groups_can_view__in=groups)
query = reduce(operator.or_, [anon_query, user_query])
query = reduce(operator.and_, [sec1_query, query])
query = reduce(operator.or_, [query, user_perm_q, group_perm_q])
else:
sqs = sqs.filter(allow_anonymous_view=True)
else:
sqs = sqs.filter(allow_anonymous_view=True)
# custom filters
for filter in filters:
sqs = sqs.filter(content='"%s"' % filter)
return sqs.models(self.model)
def personal_search(request):
"""
The user can search any item within his own collections and can search **only shared items** of other users
TODO: Build a hash table to store item_id in the result of user_item to reduce time from O(n^2) to O(n)
Reference: http://docs.haystacksearch.org/dev/searchqueryset_api.html#field-lookups
"""
#Two parameters to tune
RESULTS_PER_PAGE = 10
load_all = False
query = request.GET['q'].strip() #Heystack only accepts key name as 'q'
user_id = int(request.GET['pid'])
if query == '':
sqs = EmptySearchQuerySet()
else:
searchqueryset = SearchQuerySet()
if user_id == request.user.pk:
pronoun = '我'
own_items = User_Item.objects.filter(user__pk=request.user.pk)
else:
pronoun = Profile.objects.get(pk=user_id).name
own_items = User_Item.objects.filter(user__pk=user_id).exclude(status=1)
own_items_ids = []
for oi in own_items:
own_items_ids.append(int(oi.item_id))
sqs = searchqueryset.auto_query(query).filter(primary_key__in=own_items_ids)
if load_all:
sqs = sqs.load_all()
paginator = Paginator(sqs, RESULTS_PER_PAGE)
try:
page = paginator.page(request.GET.get('page', 1))
feeds_id = ''
for result in page.object_list:
feeds_id += str(result.object.id) + ','
feeds_id = feeds_id[:-1]
topics_of_item_dict = get_topics_of_item(feeds_id, request.user.pk)
friends_of_item_dict = get_friends_of_item(feeds_id, request.user.pk)
user_item_status_dict = get_user_items(feeds_id, request.user.pk)
except InvalidPage:
raise Http404
context = {
'query': query,
'page': page,
'page_type':'search',
'topics_of_item_dict':topics_of_item_dict,
'friends_of_item_dict':friends_of_item_dict,
'user_item_status_dict':user_item_status_dict,
'paginator': paginator,
'suggestion': None,
'pronoun': pronoun,
'num_results': len(sqs),
'user_id': user_id
}
from django.template import add_to_builtins
add_to_builtins('haystack.templatetags.highlight')
return render_to_response('main/search/personal_search_results.html', context, context_instance=RequestContext(request))
def current_bills_sorted(self):
from haystack.query import SearchQuerySet
qs = SearchQuerySet().using("bill").filter(indexed_model_name__in=["Bill"], congress=CURRENT_CONGRESS, committees=self.id).order_by('-proscore')
return {
"count": qs.count(),
"bills": [ b.object for b in qs[0:100] ],
}
def _specific_view(user, obj):
"""
determines if a user has specific permissions to view the object.
note this is based only on:
(users_can_view contains user)
+
(groups_can_view contains one of user's groups)
"""
sqs = SearchQuerySet()
sqs = sqs.models(obj.__class__)
groups = [g.pk for g in user.group_set.all()]
q_primary_key = SQ(primary_key=obj.pk)
q_groups = SQ(groups_can_view__in=groups)
q_users = SQ(users_can_view__in=[user.pk])
if groups:
sqs = sqs.filter(q_primary_key & (q_groups | q_users))
else:
sqs = sqs.filter(q_primary_key & q_users)
if sqs:
return True
return False
def get(self, request):
#takes a latitude and longitude and returns artifacts within 2 miles
try:
lon = float(request.GET.get('lon', None))
lat = float(request.GET.get('lat', None))
except:
return HttpResponse('You Must Specify a latitude and longitude', content_type='application/json', status=status.HTTP_400_BAD_REQUEST)
username = request.user.username
pnt = Point(lon, lat)
# Within a two miles.
max_dist = D(mi=2)
# 'location' is the fieldname from our ``SearchIndex``...
# Do the radius query.
sqs = SearchQuerySet().distance('location',pnt).order_by('distance')
within = sqs.dwithin('location', pnt, max_dist)
if len(within) != len(sqs) and len(within) !=0 and len(sqs) !=0:
closest_not_within = pnt.distance(sqs[len(within)])*100*0.6214
else: closest_not_within='null'
the_data = []
for result in sqs:
# get number of pickups as well as whether or not the user has already picked up
already_picked_up = (username in result.pickup_count)
_pickup_count = count_m2m(result.pickup_count)
# get number of upvotes as well as whether or not the user has already upvoted
already_upvoted = (username in result.upvoted)
upvote_count = count_m2m(result.upvoted)
# make a point
artifactpnt = Point(float(result.longitude), float(result.latitude))
loopy_data = {'author': result.author, 'upvote_count': upvote_count, 'already_upvoted': already_upvoted, 'already_picked_up': already_picked_up, 'longitude': result.longitude, 'radius': result.radius, 'latitude': result.latitude, 'pub_date': str(result.pub_date), 'pickup_count': _pickup_count, 'distance': (artifactpnt.distance(pnt) *100)* 0.6214, 'pk': result.pk}
the_data += [loopy_data]
the_data = [json.dumps(the_data + [{'closest_not_within':closest_not_within}])]
return HttpResponse(the_data, content_type='application/json', status=status.HTTP_200_OK)
def __init__(self, *args, user, report_ids=(), **kwargs):
sqs = SearchQuerySet().models(Report)
# Ensure anonymous/public users cannot see non-public reports in all
# cases.
if not user.is_active:
if report_ids:
sqs = sqs.filter(SQ(id__in=report_ids) | SQ(is_public=True))
else:
sqs = sqs.filter(is_public=True)
super().__init__(*args, searchqueryset=sqs, **kwargs)
self.user = user
self.report_ids = report_ids
# Only certain fields on this form can be used by members of the
# public.
if not user.is_active:
for name in self.fields.keys():
if name not in self.public_fields:
self.fields.pop(name)
if user.is_anonymous():
if report_ids:
source_field = self.fields['source']
source_choices = source_field.choices
source_field.choices = [
(value, label) for (value, label) in source_choices if value != 'invited']
else:
self.fields.pop('source')
def build_form(self, form_kwargs=None):
log = logging.getLogger(".".join((__name__, "PostSearch")))
request = self.request
log.debug("which author has domain %r?", request.META["HTTP_HOST"])
try:
self.author = User.objects.get(authorsite__site__domain=request.META["HTTP_HOST"])
except User.DoesNotExist:
log.debug(" no such author! no results at all!")
self.author = None
sqs = SearchQuerySet().none()
else:
sqs = SearchQuerySet().filter(author_pk=self.author.pk)
# What visibility of posts can the searcher see?
if request.user.is_anonymous():
log.debug(" viewer is anonymous, so only %s's public posts", self.author.username)
sqs = sqs.filter(private=0)
elif request.user.pk == self.author.pk:
log.debug(" viewer is %s, so all their posts", self.author.username)
else:
# TODO: honor trust groups instead of giving everyone else only public posts
log.debug(" viewer is logged in as somebody else, so only %s's public posts", self.author.username)
sqs = sqs.filter(private=0)
self.searchqueryset = sqs
return super(PostSearch, self).build_form(form_kwargs)
def search(request):
sqs = SearchQuerySet().filter(content=Fuzzy(request.GET['q']))
if request.GET['id'] != '':
sqs = sqs.exclude(django_id=int(request.GET['id']))
res_list = sorted([x for x in sqs], key=lambda x: x.score, reverse=True)
ser = serializers.serialize('json', [x.object for x in res_list])
return HttpResponse(ser, content_type="application/json")
class SearchForm(forms.Form):
q = forms.CharField(required=False, label=_('Search'))
def __init__(self, *args, **kwargs):
self.searchqueryset = kwargs.pop('searchqueryset', None)
self.load_all = kwargs.pop('load_all', False)
if self.searchqueryset is None:
self.searchqueryset = SearchQuerySet()
super(SearchForm, self).__init__(*args, **kwargs)
def search(self):
if self.is_valid():
sqs = self.searchqueryset.auto_query(self.cleaned_data['q'])
if self.load_all:
sqs = sqs.load_all()
return sqs
else:
return []
def get_suggestion(self):
if not self.is_valid():
return None
return self.searchqueryset.spelling_suggestion(self.cleaned_data['q'])
def all_results(self):
from inventory.models import Product
sqs = SearchQuerySet().load_all().models(Product)
sqs = sqs.filter(shop_id=self.shop.id)
return sqs
def get_search(self, request, **kwargs):
self.method_check(request, allowed=['get'])
self.is_authenticated(request)
self.throttle_check(request)
sqs = SearchQuerySet().models(Tracks).load_all()
sqs = sqs.auto_query(request.GET.get('q', ''))
paginator = Paginator(sqs, RESULTS_PER_PAGE)
try:
page = paginator.page(int(request.GET.get('page', 1)))
except InvalidPage:
raise Http404("No such page exists.")
except ValueError:
raise Http404("Invalid page number.")
objects = []
for result in page.object_list:
if result is None:
continue
bundle = self.build_bundle(obj=result.object, request=request)
bundle = self.full_dehydrate(bundle)
objects.append(bundle)
self.log_throttled_access(request)
return self.create_response(request, {'objects': objects})
def get_related(self):
query_string = u' '.join(self.tags.names())
if query_string:
query_set = SearchQuerySet().exclude(django_id=self.pk)
return query_set.filter(content=query_string, type='thread')
return tuple()
def get_evaluations(request):
qs = SearchQuerySet().filter(
evaluated_rubrics__in=[0] + list(Rubric.objects.values_list("id", flat=True))
).narrow("is_displayed:true")
size = int(request.REQUEST.get("size", 100))
start = int(request.REQUEST.get("start", 0))
total_items = qs.count()
items = []
if start < total_items and start >= 0:
for r in qs[start:start+size]:
fields = r.get_stored_fields()
items.append(dict(
title=fields["title"],
url=fields["url"],
rubric_1=fields["evaluation_score_rubric_0"],
rubric_2=fields["evaluation_score_rubric_1"],
rubric_3=fields["evaluation_score_rubric_2"],
rubric_4=fields["evaluation_score_rubric_3"],
rubric_5=fields["evaluation_score_rubric_4"],
rubric_6=fields["evaluation_score_rubric_5"],
rubric_7=fields["evaluation_score_rubric_6"],
))
return dict(items=items, total_items=total_items)
def search(self, query=None, *args, **kwargs):
"""
haystack to query corporate memberships.
Returns a SearchQuerySet
"""
from corporate_memberships.models import CorporateMembership
from perms.utils import is_admin
user = kwargs.get('user', None)
if user.is_anonymous():
return SearchQuerySet().models().none()
is_an_admin = is_admin(user)
sqs = SearchQuerySet().models(CorporateMembership)
if query:
sqs = sqs.filter(content=sqs.query.clean(query))
else:
sqs = sqs.all()
if not is_an_admin:
# reps__contain
sqs = sqs.filter(Q(content='rep\:%s' % user.username) |
Q(creator=user) |
Q(owner=user)).filter(status_detail='active')
return sqs
def _specific_view(user, obj):
"""
determines if a user has specific permissions to view the object.
note this is based only on:
(users_can_view contains user)
+
(groups_can_view contains one of user's groups)
"""
sqs = SearchQuerySet()
sqs = sqs.models(obj.__class__)
groups = [g.pk for g in user.group_set.all()]
q_primary_key = SQ(primary_key=obj.pk)
q_groups = SQ(groups_can_view__in=groups)
q_users = SQ(users_can_view__in=[user.pk])
if groups:
sqs = sqs.filter(q_primary_key & (q_groups | q_users))
else:
sqs = sqs.filter(q_primary_key & q_users)
if sqs:
# Make sure the index isn't doing something unexpected with the query,
# like when the Whoosh StopFilter caused the primary_key portion of the
# query to be ignored.
assert len(sqs) == 1, "Index returned an unexpected result set when searching for view permissions on an object"
return True
return False
def setUp(self):
super(LiveSolrMoreLikeThisTestCase, self).setUp()
self.sqs = SearchQuerySet()
# Wipe it clean.
self.sqs.query.backend.clear()
# With the models registered, you get the proper bits.
import haystack
from haystack.sites import SearchSite
# Stow.
self.old_site = haystack.site
test_site = SearchSite()
test_site.register(MockModel, SolrMockModelSearchIndex)
test_site.register(AnotherMockModel, SolrAnotherMockModelSearchIndex)
haystack.site = test_site
# Force indexing of the content.
for mock in MockModel.objects.all():
mock.save()
# Force indexing of the content.
for mock in AnotherMockModel.objects.all():
mock.save()
self.sqs = SearchQuerySet()
def get_urls(self):
# Build SQS
sqs = SearchQuerySet()
for facet in settings.OSCAR_SEARCH_FACETS["fields"].values():
sqs = sqs.facet(facet["field"])
for facet in settings.OSCAR_SEARCH_FACETS["queries"].values():
for query in facet["queries"]:
sqs = sqs.query_facet(facet["field"], query[1])
# The form class has to be passed to the __init__ method as that is how
# Haystack works. It's slightly different to normal CBVs.
urlpatterns = patterns(
"",
url(r"^$", self.search_view(form_class=forms.MultiFacetedSearchForm), name="search"),
url(r"^suggest/$", self.suggest_view.as_view(), name="suggest"),
url(
r"^default/$",
search_view_factory(
view_class=views.FacetedSearchView,
form_class=forms.PriceRangeSearchForm,
searchqueryset=sqs,
template="search/results.html",
),
name="search_default",
),
)
return self.post_process_urls(urlpatterns)
def get_searchqueryset(self, form):
"""Get the Haystack searchqueryset (which we treat as
a regular Django queryset."""
sqs = SearchQuerySet()
if self.model:
sqs = sqs.models(self.model)
# FIXME: Move somewhere more sensible
if settings.PORTAL_HIDE_DRAFTS and not self.request.user.is_staff:
sqs = sqs.narrow("publication_status:%d" % models.Resource.PUBLISHED)
for facet in self.facetclasses:
sqs = facet.apply(sqs)
# apply the query
if form.is_valid():
sqs = form.filter(sqs)
for facetclass in self.facetclasses:
sqs = facetclass.narrow(sqs, self.request.GET.getlist(
facetclass.paramname))
counts = sqs.facet_counts()
current = sqs.query.narrow_queries
for facetclass in self.facetclasses:
facetclass.parse(counts, current)
# FIXME: Find way around assigning the sqs to the instance,
# but it seems to be difficult to prevent it from running
# multiple times otherwise, e.g. when checking for a spelling
# suggestion.
self.searchqueryset = sqs
return sqs
class LiveSearchQuerySetTestCase(HaystackBackendTestCase, TestCase):
"""
SearchQuerySet specific tests
"""
fixtures = ['base_data.json']
def get_index(self):
return MockSearchIndex()
def setUp(self):
super(LiveSearchQuerySetTestCase, self).setUp()
self.backend.update(self.index, MockModel.objects.all())
self.sq = connections['default'].get_query()
self.sqs = SearchQuerySet()
def test_result_class(self):
# Assert that we're defaulting to ``SearchResult``.
sqs = self.sqs.all()
self.assertTrue(isinstance(sqs[0], SearchResult))
# Custom class.
sqs = self.sqs.result_class(MockSearchResult).all()
self.assertTrue(isinstance(sqs[0], MockSearchResult))
# Reset to default.
sqs = self.sqs.result_class(None).all()
self.assertTrue(isinstance(sqs[0], SearchResult))
def test_facet(self):
self.assertEqual(len(self.sqs.facet('name').facet_counts()['fields']['name']), 3)
def handle_save(self, sender, instance, **kwargs):
if sender not in self.index_models:
return
# If IgniteUser was saved and changed name, then update all its authored projects owner_name:
if sender == IgniteUser:
owner_projects_qs = Project.objects.filter(owner=instance)
base_owner_projects_sq = SearchQuerySet()
base_owner_projects_sq = base_owner_projects_sq.models(Project)
for owner_project_instance in owner_projects_qs:
# get owner project from haystack index:
owner_projects_sq = base_owner_projects_sq.filter(id=owner_project_instance.id)
if owner_projects_sq.count() == 0:
# continue till get to the first project found in haystack.
# Note: all projects should be found in haystack always, so this would never happen.
continue
old_owner_name = owner_projects_sq[0].owner_name
# if indexed owner_name is different from the current user name:
if old_owner_name != instance.name:
using_backends = self.connection_router.for_write(instance=owner_project_instance)
for using in using_backends:
try:
index = self.connections[using].get_unified_index().get_index(Project)
if hasattr(index, 'update_owner_projects'):
# update the owner_name of the owner projects:
index.update_owner_projects(owner=instance, using=using)
except NotHandled:
# TODO: Maybe log it or let the exception bubble?
pass
break
else:
super(IgniteSignalProcessor, self).handle_save(sender, instance, **kwargs)
def get_box(request):
long1 = float(request.GET.get('long1',0))
lat1 = float(request.GET.get('lat1',0))
long2 = float(request.GET.get('long2',0))
lat2 = float(request.GET.get('lat2',0))
bl = Point(long1, lat1)
tr = Point(long2, lat2)
max_dist = D(mi=20)
sqs = SearchQuerySet().within('location', bl, tr)
print (tr,bl)
data = {'counter':0 ,'places':[]}
if sqs.count()==0:
return HttpResponse(content=json.dumps(data))
for elem in sqs.all():
if elem.object.long:
data['counter'] += 1
place = {'long':0, 'lat':0, 'title':''}
place['long'] = float(elem.object.long)
place['lat'] = float(elem.object.lat)
place['title'] = elem.object.name
place['level'] = elem.object.level
data['places'].append(place)
return HttpResponse(content=json.dumps(data))
def search(self):
if not hasattr(self, "cleaned_data"):
return self.no_query_found()
search_fields = [key for key, value in self.cleaned_data.iteritems() if value == True]
if 'title' not in search_fields:
sqs = SearchQuerySet()
else:
sqs = super(SearchOptionsForm, self).search()
# title is a document field and has been used for filtering in super method search()
search_fields = [key for key in search_fields if key != 'title']
query = sqs.query.clean(self.cleaned_data.pop('q'))
galleries = [g.id for g in self.cleaned_data.get('galleries', [])]
search_galleries = self.cleaned_data.get('search_galleries_choice', "ALL")
query_words = query.split()
for key in search_fields:
if key == "tags":
sqs = sqs.filter_or(tags__in=[query.lower() for query in query_words])
else:
sqs = self._filter_or_query_words(sqs, key, query_words)
if search_galleries == 'SELECTED':
sqs = sqs.filter_and(galleries_ids__in=galleries)
return sqs
def get_response(project_uri, query_string, include_n3=True):
d = {
'results': list(),
}
project_graph = projects.get_project_graph(project_uri)
graph = Graph()
query_set = SearchQuerySet().models(Text).filter(
content=AutoQuery(query_string), project__exact=project_uri
)
highlighter = Highlighter(query_string, html_tag='span', css_class=CSS_RESULT_MATCH_CLASS)
title_highlighter = TitleHighlighter(query_string, html_tag='span', css_class=CSS_RESULT_MATCH_CLASS)
d['spelling_suggestion'] = query_set.spelling_suggestion()
for result in query_set:
text_uri = URIRef(result.get_stored_fields()['identifier'])
if annotations.has_annotation_link(project_graph, text_uri) or projects.is_top_level_project_resource(project_uri, text_uri):
d['results'].append(search_result_to_dict(result, project_uri, highlighter, title_highlighter))
if include_n3:
graph += utils.metadata_triples(project_graph, text_uri)
if include_n3:
d['n3'] = graph.serialize(format='n3')
return d
from aristotle_mdr.views.search import PermissionSearchView
from haystack.views import search_view_factory
from haystack.query import SearchQuerySet
urlpatterns = [
url(r'^', include('aristotle_mdr.urls')),
url(
r'^extension_test/',
include('extension_test.extension_urls',
app_name="extension_test",
namespace="extension_test")),
url(r'^fail_search/?',
search_view_factory(view_class=PermissionSearchView,
template='search/search.html',
searchqueryset=SearchQuerySet(),
form_class=PermissionSearchForm),
name='fail_search'),
url(
r'^',
include('aristotle_mdr.contrib.links.urls',
app_name="aristotle_mdr_links",
namespace="aristotle_mdr_links")),
url(
r'^publish/',
include('aristotle_mdr.contrib.self_publish.urls',
app_name="aristotle_self_publish",
namespace="aristotle_self_publish")),
url(
r'^',
include('aristotle_mdr.contrib.slots.urls',
def get_estimates_common(request, data):
""" Append common page content to the argument data
:param request: web request
:param data: A dict that contains page render data.
:return: the results of the search query
"""
post = None
if request.method == "POST":
post = request.POST
# Ensure timeout for session manually.
# We could have used set_expiry but then we do not get sessions
# to automatically expire when browser closes.
last_access = request.session.get('estimates_last_access_time', 0.0)
import time
current_time = time.time()
if last_access < (current_time - (20 * 60.0)):
request.session.pop("estimates_post_data", None)
request.session['estimates_last_access_time'] = current_time
# Get post from session if this is not a POST request.
if post is None:
# If the user had made queries before during this session, recover the state here.
post = request.session.get("estimates_post_data")
else:
# Store the POST data for possible use later in this session.
request.session["estimates_post_data"] = post
query = {}
def is_checked(prefix, element, reset):
"""
Helper function that checks post data (if any) to see if
the checkbox corresponding to the given element is checked.
:param prefix: the prefix used by the view to identify the
model of the element.
:param element: the model object whose check state is needed.
:param reset: the POST parameter name that may flag a full reset of the
checkboxes. If this parameter is set to a "Reset to default" value,
then the result of this function will be 1.
:return: 1 if the checkbox is checked, 0 otherwise
"""
if post is not None and prefix + str(element.pk) not in post and \
post.get(reset) != "Reset to default":
return 0
else:
return 1
EstimateManager.cache()
# Check which Flags (nations) are selected and include the selection in the query.
nations = [[
x.name, x.pk,
is_checked("checkbox_nation_", x, "submit_nation")
] for x in EstimateManager.nations.values()]
data['nations'] = nations
query["nation__in"] = [nation[0] for nation in nations if nation[2] == 1]
data['all_nations_selected'] = len(nations) == len(query["nation__in"])
export_regions = {}
for area, regions in EstimateManager.export_hierarchy.iteritems():
children = [[[x.name, x.pk],
is_checked("eregion-button-", x, "submit_regions")]
for x in regions]
checked = is_checked("earea-button-", area, "submit_regions")
if len(regions) == 1:
children[0][1] = checked
export_regions[(area, checked)] = children
import_regions = {}
for area, regions in EstimateManager.import_hierarchy.iteritems():
children = [[[x.name, x.pk],
is_checked("dregion-button-", x, "submit_regions")]
for x in regions]
checked = is_checked("darea-button-", area, "submit_regions")
if len(regions) == 1:
children[0][1] = checked
import_regions[(area, checked)] = children
data['export_regions'] = collections.OrderedDict(
sorted(export_regions.items(), key=lambda x: x[0][0].name))
data['import_regions'] = collections.OrderedDict(
sorted(import_regions.items(), key=lambda x: x[0][0].name))
def query_region(query_key, regions_dict, all_selected_key):
"""
Obtain a list of the names of selected regions in the regions_dict
:param query_key: The key used when inserting this list on the query
dict
:param regions_dict: A dictionary with keys given by Area and whose
values are lists of the regions in that area in the format
[[name, pk], checked]
:param all_selected_key: The key to set a boolean value which indicates
whether all regions are selected.
:return:
"""
from itertools import chain
# Flatten the regions so that we may generate the corresponding query term.
flat = list(chain.from_iterable(regions_dict.values()))
query[query_key] = [region[0][0] for region in flat if region[1] == 1]
data[all_selected_key] = len(flat) == len(query[query_key])
query_region("embarkation_region__in", export_regions,
"all_embarkations_selected")
query_region("disembarkation_region__in", import_regions,
"all_disembarkations_selected")
year_form = None
# Ensure that GET requests or Reset POST requests yield a fresh copy of the form with default values.
if post is not None and not post.get("submit_year") == "Reset to default":
year_form = EstimateYearForm(post)
if year_form is not None and year_form.is_valid():
query["year__gte"] = year_form.cleaned_data["frame_from_year"]
query["year__lte"] = year_form.cleaned_data["frame_to_year"]
else:
if year_form is not None:
import logging
logging.getLogger('voyages').error(year_form.errors)
year_form = EstimateYearForm(
initial={
'frame_from_year': globals.default_first_year,
'frame_to_year': globals.default_last_year
})
query["year__gte"] = globals.default_first_year
query["year__lte"] = globals.default_last_year
data['year_form'] = year_form
data['query'] = query
data['post'] = post
return SearchQuerySet().models(Estimate).filter(**query).load_all()
print
def _test_solr(self):
print '=== SOLR ==='
sb = backend.SearchBackend()
try:
video = Video.objects.all()[:1].get()
update_search_index(Video, video.pk)
sb.conn.commit()
except (IOError, SolrError), e:
raise Exception('Solr is unavailable')
except Video.DoesNotExist:
raise Exception('Database is empty to test Solr')
# failing on nf, will check later
sqs_count = SearchQuerySet().count()#(content=video.title)
#sqs = SearchQuerySet().filter(content=video.title)
assert sqs_count, 'Solr is unavailable. Can\'t find video'
print 'OK'
print
def _test_memcached(self):
print '=== CACHE ==='
print 'backend: ', settings.CACHE_BACKEND
val = random.random()
key = 'test-cache-%s' % base64.b64encode(str(random.random()))
cache.set(key, val)
assert val == cache.get(key), u'Cache is unavailable. Can\'t get value'
def perform_search(search, lang):
items = search['items']
search_terms = {}
custom_terms = []
sqs = SearchQuerySet()
for item in items:
term = item['searchTerm']
operator = _operators_dict[item['op']]
is_list = isinstance(term, list)
if is_list and not operator.list_type:
term = term[0]
skip = False
if operator.front_end_op_str == _op_contains.front_end_op_str:
m = re.match(u'^\s*["\u201c](\*?)([^\*]*)(\*?)["\u201d]\s*$', term)
if m:
# Change to exact match and remove quotes.
# Make sure we sanitize the input.
term = sqs.query.clean(m.group(2))
operator = _op_eq
# Here we are using Solr's format, which is not very portable,
# but at this stage this project is very dependent on Solr anyway.
# If the search is really for a full exact match, then we search
# on the plaintext_exact variant of the field. If it is a "contains"
# the exact search terms, then we use the plaintext variant instead.
custom_terms.append(u'var_' + unicode(item['varName']) +
'_plaintext' +
('_exact' if len(m.group(1)) +
len(m.group(3)) == 0 else '') + ':("' +
term + '")')
skip = True
if not skip:
search_terms[u'var_' + unicode(item['varName']) + u'__' +
unicode(operator.back_end_op_str)] = term
search_terms[u'var_intra_american_voyage__exact'] = json.loads(
search_terms.get(u'var_intra_american_voyage__exact', 'false'))
result = sqs.models(Voyage).filter(**search_terms)
for ct in custom_terms:
result = result.filter(content=Raw(ct, clean=True))
order_fields = search.get('orderBy')
if order_fields:
remaped_fields = []
for field in order_fields:
# Remap field names if they are plain text or language dependent.
order_by_field = u'var_' + unicode(field['name'])
if order_by_field.endswith('_partial'):
# Partial dates are encoded in a way that is terrible for sorting MM,DD,YYYY.
# Therefore we use the original Date value (which defaults month, day to 1).
order_by_field = order_by_field[0:-8]
if order_by_field.endswith('_lang'):
order_by_field += '_' + lang + '_exact'
elif order_by_field in translated_field_list:
order_by_field += '_lang_' + lang + '_exact'
elif order_by_field in plain_text_suffix_list:
order_by_field += '_plaintext_exact'
if field['direction'] == 'desc':
order_by_field = '-' + order_by_field
elif order_by_field.endswith('_exact'):
remaped_fields.append('eq(' + order_by_field + ', \' \')')
remaped_fields.append(order_by_field)
result = result.order_by(*remaped_fields)
return result
def get(self, request):
frm = ItemSearchForm(request.GET)
self.object_list = Item.objects.none()
if frm.is_valid():
cdata = frm.clean()
sqs = SearchQuerySet().filter(published=True)
psqs = sqs
# No blank values, please
for key in cdata.iterkeys():
if isinstance(cdata[key], basestring):
cdata[key] = cdata[key].strip()
if cdata['categories']:
sqs = sqs.filter(categories__in=[
x.strip() for x in cdata['categories'].split(' ')
])
if cdata['title']:
sqs = sqs.filter(title=cdata['title'])
if cdata['artist']:
sqs = sqs.filter(artist=cdata['artist'])
if cdata['date_from']:
sqs = sqs.filter(date_from__gte=cdata['date_from'])
if cdata['date_to']:
sqs = sqs.filter(date_to__lte=cdata['date_to'])
if cdata['origin_city']:
sqs = sqs.filter(origin_city=cdata['origin_city'])
if cdata['origin_country']:
sqs = sqs.filter(origin_country=cdata['origin_country'])
if cdata['materials']:
sqs = sqs.filter(materials__in=[
x.strip() for x in cdata['materials'].split(' ')
])
if cdata['video_only']:
sqs = sqs.filter(video_only=True)
# fulltext search
if cdata['q']:
sqs = sqs.filter(content=cdata['q'])
# No search data entered
if psqs == sqs:
return redirect(reverse('item_search'))
# Assigning a list to self.object_list won't work, it needs a QuerySet.
# We're basically loading the items twice :(
try:
results = list(
sqs.order_by('score')[:1000]
) # slicing the array prevents multiple queries to Solr.
except KeyError:
# In case we do not have score, will happen with Whoosh
results = list(sqs[:1000])
ids = [
x.object.id for x in results
] # sqs.values_list('django_id', flat=True) won't work with Haystack.
self.object_list = Item.objects.filter(id__in=ids)
self.parent_category = None
self.current_category = None
self.child_categories = None
context = self.get_context_data(object_list=self.object_list)
return (self.render_to_response(context))
def build_haystack_filters(self, parameters):
from haystack.inputs import Raw
from haystack.query import SearchQuerySet, SQ # noqa
sqs = None
# Retrieve Query Params
# Text search
query = parameters.get('q', None)
# Types and subtypes to filter (map, layer, vector, etc)
type_facets = parameters.getlist("type__in", [])
# If coming from explore page, add type filter from resource_name
resource_filter = self._meta.resource_name.rstrip("s")
if resource_filter != "base" and resource_filter not in type_facets:
type_facets.append(resource_filter)
# Publication date range (start,end)
date_end = parameters.get("date__lte", None)
date_start = parameters.get("date__gte", None)
# Topic category filter
category = parameters.getlist("category__identifier__in")
# Keyword filter
keywords = parameters.getlist("keywords__slug__in")
# Region filter
regions = parameters.getlist("regions__name__in")
# Owner filters
owner = parameters.getlist("owner__username__in")
# Sort order
sort = parameters.get("order_by", "relevance")
# Geospatial Elements
bbox = parameters.get("extent", None)
# Filter by Type and subtype
if type_facets is not None:
types = []
subtypes = []
for type in type_facets:
if type in ["map", "layer", "document", "user"]:
# Type is one of our Major Types (not a sub type)
types.append(type)
elif type in LAYER_SUBTYPES.keys():
subtypes.append(type)
if len(subtypes) > 0:
types.append("layer")
sqs = SearchQuerySet().narrow("subtype:%s" %
','.join(map(str, subtypes)))
if len(types) > 0:
sqs = (SearchQuerySet() if sqs is None else sqs).narrow(
"type:%s" % ','.join(map(str, types)))
# Filter by Query Params
# haystack bug? if boosted fields aren't included in the
# query, then the score won't be affected by the boost
if query:
if query.startswith('"') or query.startswith('\''):
# Match exact phrase
phrase = query.replace('"', '')
sqs = (SearchQuerySet() if sqs is None else sqs).filter(
SQ(title__exact=phrase) | SQ(description__exact=phrase)
| SQ(content__exact=phrase))
else:
words = [
w for w in re.split('\W', query, flags=re.UNICODE) if w
]
for i, search_word in enumerate(words):
if i == 0:
sqs = (SearchQuerySet() if sqs is None else sqs) \
.filter(
SQ(title=Raw(search_word)) |
SQ(description=Raw(search_word)) |
SQ(content=Raw(search_word))
)
elif search_word in ["AND", "OR"]:
pass
elif words[i - 1] == "OR": # previous word OR this word
sqs = sqs.filter_or(
SQ(title=Raw(search_word))
| SQ(description=Raw(search_word))
| SQ(content=Raw(search_word)))
else: # previous word AND this word
sqs = sqs.filter(
SQ(title=Raw(search_word))
| SQ(description=Raw(search_word))
| SQ(content=Raw(search_word)))
# filter by category
if category:
sqs = (SearchQuerySet() if sqs is None else sqs).narrow(
'category:%s' % ','.join(map(str, category)))
# filter by keyword: use filter_or with keywords_exact
# not using exact leads to fuzzy matching and too many results
# using narrow with exact leads to zero results if multiple keywords
# selected
if keywords:
for keyword in keywords:
sqs = (SearchQuerySet() if sqs is None else sqs).filter_or(
keywords_exact=keyword)
# filter by regions: use filter_or with regions_exact
# not using exact leads to fuzzy matching and too many results
# using narrow with exact leads to zero results if multiple keywords
# selected
if regions:
for region in regions:
sqs = (SearchQuerySet() if sqs is None else sqs).filter_or(
regions_exact__exact=region)
# filter by owner
if owner:
sqs = (SearchQuerySet() if sqs is None else sqs).narrow(
"owner__username:%s" % ','.join(map(str, owner)))
# filter by date
if date_start:
sqs = (SearchQuerySet() if sqs is None else sqs).filter(
SQ(date__gte=date_start))
if date_end:
sqs = (SearchQuerySet() if sqs is None else sqs).filter(
SQ(date__lte=date_end))
# Filter by geographic bounding box
if bbox:
left, bottom, right, top = bbox.split(',')
sqs = (SearchQuerySet() if sqs is None else sqs).exclude(
SQ(bbox_top__lte=bottom) | SQ(bbox_bottom__gte=top)
| SQ(bbox_left__gte=right) | SQ(bbox_right__lte=left))
# Apply sort
if sort.lower() == "-date":
sqs = (SearchQuerySet() if sqs is None else sqs).order_by("-date")
elif sort.lower() == "date":
sqs = (SearchQuerySet() if sqs is None else sqs).order_by("date")
elif sort.lower() == "title":
sqs = (SearchQuerySet()
if sqs is None else sqs).order_by("title_sortable")
elif sort.lower() == "-title":
sqs = (SearchQuerySet()
if sqs is None else sqs).order_by("-title_sortable")
elif sort.lower() == "-popular_count":
sqs = (SearchQuerySet()
if sqs is None else sqs).order_by("-popular_count")
else:
sqs = (SearchQuerySet() if sqs is None else sqs).order_by("-date")
return sqs
def deploy_checks(request=None):
passed = []
failed = []
# cache something now to see if it's still there further down.
randval = random.randint(1, 1000000)
cache.set('check_things_cache_test', randval, 60)
# Django database
try:
n = Semester.objects.all().count()
if n > 0:
passed.append(('Main database connection', 'okay'))
else:
failed.append(('Main database connection',
"Can't find any coredata.Semester objects"))
except django.db.utils.OperationalError:
failed.append(
('Main database connection', "can't connect to database"))
except django.db.utils.ProgrammingError:
failed.append(('Main database connection', "database tables missing"))
# non-BMP Unicode in database
try:
l = LogEntry.objects.create(userid='ggbaker',
description='Test Unicode \U0001F600',
related_object=Semester.objects.first())
except OperationalError:
failed.append(('Unicode handling in database',
'non-BMP character not supported by connection'))
else:
l = LogEntry.objects.get(id=l.id)
if '\U0001F600' in l.description:
passed.append(('Unicode handling in database', 'okay'))
else:
failed.append(('Unicode handling in database',
'non-BMP character not stored correctly'))
# Celery tasks
celery_okay = False
try:
if settings.USE_CELERY:
try:
from coredata.tasks import ping
except ImportError:
failed.append(
('Celery task',
"Couldn't import task: probably missing MySQLdb module"))
else:
try:
t = ping.apply_async()
except kombu.exceptions.OperationalError:
failed.append(
('Celery task',
'Kombu error. Probably RabbitMQ not running.'))
else:
res = t.get(timeout=5)
if res == True:
passed.append(('Celery task', 'okay'))
celery_okay = True
else:
failed.append(
('Celery task', 'got incorrect result from task'))
else:
failed.append(('Celery task', 'celery disabled in settings'))
except celery.exceptions.TimeoutError:
failed.append(
('Celery task',
"didn't get result before timeout: celeryd maybe not running"))
except socket.error:
failed.append(('Celery task', "can't communicate with broker"))
except NotImplementedError:
failed.append(('Celery task', 'celery disabled'))
except django.db.utils.ProgrammingError:
failed.append(('Celery task', 'celery DB tables missing'))
except django.db.utils.OperationalError:
failed.append(('Celery task', 'djkombu tables missing: try migrating'))
# celery beat
try:
from coredata.tasks import BEAT_TEST_FILE, BEAT_FILE_MAX_AGE
beatfile_age = time.time() - os.stat(BEAT_TEST_FILE).st_mtime
if beatfile_age < BEAT_FILE_MAX_AGE:
passed.append(('Celery beat', 'okay'))
else:
failed.append((
'Celery beat',
'marker file is old: celery beat likely not processing tasks'))
except OSError:
failed.append((
'Celery beat',
'marker file is missing: celery beat likely not processing tasks'))
# Django cache
# (has a subprocess do something to make sure we're in a persistent shared cache, not DummyCache)
subprocess.call(
['python3', 'manage.py', 'check_things', '--cache_subcall'])
cache_okay = False
res = cache.get('check_things_cache_test')
if res == randval:
failed.append((
'Django cache',
'other processes not sharing cache: dummy/local probably being used instead of memcached'
))
elif res is None:
failed.append(
('Django cache', 'unable to retrieve anything from cache'))
elif res != randval + 1:
failed.append(('Django cache', 'unknown result'))
else:
passed.append(('Django cache', 'okay'))
cache_okay = True
# Reporting DB connection
try:
db = SIMSConn()
db.execute("SELECT last_name FROM ps_names WHERE emplid=301355288", ())
result = list(db)
# whoever this is, they have non-ASCII in their name: let's hope they don't change it.
lname = result[0][0]
if not isinstance(lname, str):
failed.append(
('Reporting DB connection',
'string result not a string: check Unicode decoding'))
elif lname[1] != u'\u00e4':
failed.append(('Reporting DB connection',
'returned incorrectly-decoded Unicode'))
elif len(result) == 0:
failed.append(('Reporting DB connection',
'query inexplicably returned nothing'))
else:
passed.append(('Reporting DB connection', 'okay'))
except SIMSProblem as e:
failed.append(
('Reporting DB connection', 'SIMSProblem, %s' % (str(e))))
except ImportError:
failed.append(
('Reporting DB connection', "couldn't import DB2 module"))
# compression enabled?
if settings.COMPRESS_ENABLED:
passed.append(('Asset compression enabled', 'okay'))
else:
failed.append(('Asset compression enabled', 'disabled in settings'))
# Haystack searching
from haystack.query import SearchQuerySet
try:
res = SearchQuerySet().filter(text='cmpt')
if res:
passed.append(('Haystack search', 'okay'))
else:
failed.append((
'Haystack search',
'nothing found: maybe update_index, or wait for search server to fully start'
))
except IOError:
failed.append(('Haystack search', "can't read/write index"))
# photo fetching
if cache_okay and celery_okay:
try:
res = do_photo_fetch(['301222726'])
if '301222726' not in res: # I don't know who 301222726 is, but he/she is real.
failed.append(
('Photo fetching', "didn't find photo we expect to exist"))
else:
passed.append(('Photo fetching', 'okay'))
except (KeyError, Unit.DoesNotExist, django.db.utils.ProgrammingError):
failed.append(('Photo fetching', 'photo password not set'))
except urllib.error.HTTPError as e:
failed.append(
('Photo fetching',
'failed to fetch photo (%s). Maybe wrong password?' % (e)))
else:
failed.append(
('Photo fetching', 'not testing since memcached or celery failed'))
# emplid/userid API
emplid = userid_to_emplid('ggbaker')
if not emplid:
failed.append(('Emplid API', 'no emplid returned'))
elif isinstance(emplid, str) and not emplid.startswith('2000'):
failed.append(('Emplid API', 'incorrect emplid returned'))
else:
passed.append(('Emplid API', 'okay'))
# Piwik API
#if not request:
# failed.append(('Piwik API', "can only check in web frontend with valid request object"))
#elif not settings.PIWIK_URL or not settings.PIWIK_TOKEN:
# failed.append(('Piwik API', "not configured in secrets.py"))
#else:
# # try to re-log this request in piwik and see what happens
# from piwik_middleware.tracking import PiwikTrackerLogic, urllib_errors
# tracking_logic = PiwikTrackerLogic()
# kwargs = tracking_logic.get_track_kwargs(request)
# try:
# tracking_logic.do_track_page_view(fail_silently=False, **kwargs)
# except urllib_errors as e:
# failed.append(('Piwik API', "API call failed: %s" % (e)))
# else:
# passed.append(('Piwik API', 'okay'))
# Backup server
#if not settings.BACKUP_SERVER or not settings.BACKUP_USER or not settings.BACKUP_PATH or not settings.BACKUP_PASSPHRASE:
# failed.append(('Backup server', 'Backup server settings not all present'))
#else:
# from coredata.management.commands.backup_remote import do_check
# try:
# do_check()
# except RuntimeError as e:
# failed.append(('Backup server', unicode(e)))
# passed.append(('Backup server', 'okay'))
# certificates
bad_cert = 0
res = _check_cert('/etc/stunnel/stunnel.pem')
if res:
failed.append(('Stunnel cert', res))
bad_cert += 1
res = _check_cert('/etc/nginx/cert.pem')
if res:
failed.append(('SSL PEM', res))
bad_cert += 1
res = _check_cert('/etc/nginx/cert.key')
if res:
failed.append(('SSL KEY', res))
bad_cert += 1
if bad_cert == 0:
passed.append((
'Certificates',
'All okay, but maybe check http://www.digicert.com/help/ or https://www.ssllabs.com/ssltest/'
))
# SVN database
if settings.SVN_DB_CONNECT:
from courselib.svn import SVN_TABLE, _db_conn
import MySQLdb
try:
db = _db_conn()
db.execute('SELECT count(*) FROM ' + SVN_TABLE, ())
n = list(db)[0][0]
if n > 0:
passed.append(('SVN database', 'okay'))
else:
failed.append(('SVN database', "couldn't access records"))
except MySQLdb.OperationalError:
failed.append(('SVN database', "can't connect to database"))
else:
failed.append(('SVN database', 'SVN_DB_CONNECT not set in secrets.py'))
# file creation in the necessary places
dirs_to_check = [
(settings.DB_BACKUP_DIR, 'DB backup dir'),
(settings.SUBMISSION_PATH, 'submitted files path'),
(os.path.join(settings.COMPRESS_ROOT,
'CACHE'), 'compressed media root'),
]
for directory, label in dirs_to_check:
res = _check_file_create(directory)
if res is None:
passed.append(('File creation in ' + label, 'okay'))
else:
failed.append(('File creation in ' + label, res))
# are any services listening publicly that shouldn't?
hostname = socket.gethostname()
ports = [
25, # mail server
#4369, # epmd, erlang port mapper daemon is okay to listen externally and won't start with ERL_EPMD_ADDRESS set. http://serverfault.com/questions/283913/turn-off-epmd-listening-port-4369-in-ubuntu-rabbitmq
45130, # beam? rabbitmq something
4000, # main DB stunnel
50000, # reporting DB
8000, # gunicorn
11211, # memcached
9200,
9300, # elasticsearch
]
connected = []
for p in ports:
s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
try:
s.connect((hostname, p))
except socket.error:
# couldn't connect: good
pass
else:
connected.append(p)
finally:
s.close()
if connected:
failed.append(
('Ports listening externally',
'got connections to port ' + ','.join(str(p) for p in connected)))
else:
passed.append(('Ports listening externally', 'okay'))
# is the server time close to real-time?
import ntplib
c = ntplib.NTPClient()
response = c.request('0.ca.pool.ntp.org')
if abs(response.offset) > 0.1:
failed.append(
('Server time',
'Time is %g seconds off NTP pool.' % (response.offset, )))
else:
passed.append(('Server time', 'okay'))
# library sanity
err = bitfield_check()
if err:
failed.append(('Library sanity', 'django-bitfield: ' + err))
else:
err = cache_check()
if err:
failed.append(('Library sanity', 'django cache: ' + err))
else:
passed.append(('Library sanity', 'okay'))
# github-flavoured markdown subprocess
from courselib.markup import markdown_to_html
try:
# checks that script runs; does github-flavour correctly; does Unicode correctly.
html = markdown_to_html(
'test *markup*\n\n```python\nprint(1)\n```\n\u2605\U0001F600')
if html.strip(
) == '<p>test <em>markup</em></p>\n<pre lang="python"><code>print(1)\n</code></pre>\n<p>\u2605\U0001F600</p>':
passed.append(('Markdown subprocess', 'okay'))
else:
failed.append(('Markdown subprocess',
'markdown script returned incorrect markup'))
except OSError:
failed.append((
'Markdown subprocess',
'failed to start ruby command: ruby package probably not installed'
))
except RuntimeError:
failed.append(('Markdown subprocess', 'markdown script failed'))
return passed, failed
# -*- coding: utf-8 -*-
from __future__ import absolute_import
from django.conf import settings
from haystack.query import SearchQuerySet
from .models import Product
__all__ = ['product_search']
product_search = SearchQuerySet().models(Product) \
.filter(is_archived=False) \
.facet('sites').facet('categories').facet('owner').facet('price').facet('pro_owner') \
.narrow('sites:%s' % settings.SITE_ID)
def search(request):
articles = SearchQuerySet().autocomplete(
content_auto=request.POST.get('search_text', ''))
return render(request, 'search/ajax_search.html', {'articles': articles})
def get_queryset(self):
self.result = self.request.GET.get('query')
return SearchQuerySet().filter(text=self.result)
from django.conf.urls.defaults import *
from haystack.query import SearchQuerySet
from oscar.core.loading import import_module
import_module('search.views', ['Suggestions', 'MultiFacetedSearchView'],
locals())
import_module('search.forms', ['MultiFacetedSearchForm'], locals())
import_module('search.search_indexes', ['ProductIndex'], locals())
sqs = SearchQuerySet()
for field_name, field in ProductIndex.fields.items():
if field.faceted is True:
# Ensure we facet the results set by the defined facetable fields
sqs = sqs.facet(field_name)
urlpatterns = patterns(
'search.apps.views',
url(r'^suggest/$', Suggestions.as_view(), name='oscar-search-suggest'),
url(r'^$',
MultiFacetedSearchView(form_class=MultiFacetedSearchForm,
searchqueryset=sqs),
name='oscar-search'),
)
def test_update_index(self):
self._add_message()
self.assertEqual(SearchQuerySet().count(), 0)
# Update the index
update_index()
self.assertEqual(SearchQuerySet().count(), 1)
def handle(self, *args, **options):
ind = BaseResourceIndex()
if len(options['resource_ids']
) > 0: # an array of resource short_id to check.
for rid in options['resource_ids']:
print("updating resource {}".format(rid))
try:
r = BaseResource.objects.get(short_id=rid)
# if ind.should_update(r): # always True
ind.update_object(r)
except BaseResource.DoesNotExist:
print("resource {} does not exist in Django".format(rid))
except Exception as e:
print("resource {} generated exception {}".format(
rid, str(e)))
else:
sqs = SearchQuerySet().all()
print("SOLR count = {}".format(sqs.count()))
dqs = BaseResource.objects.filter(
Q(raccess__discoverable=True) | Q(raccess__public=True))
print("Django count = {}".format(dqs.count()))
# what is in Django that isn't in SOLR
found_in_solr = set()
for r in list(sqs):
found_in_solr.add(r.short_id) # enable fast matching
django_indexed = 0
django_replaced = 0
django_refreshed = 0
for r in dqs:
try:
resource = get_resource_by_shortkey(r.short_id,
or_404=False)
repl = False
if hasattr(resource, 'metadata') and \
resource.metadata is not None and \
resource.metadata.relations is not None:
repl = resource.metadata.relations.filter(
type='isReplacedBy').exists()
if not repl:
django_indexed += 1
else:
django_replaced += 1
except BaseResource.DoesNotExist:
# race condition in processing while in production
print("resource {} no longer found in Django.".format(
r.short_id))
continue
except Exception as e:
print("resource {} generated exception {}".format(
r.short_id, str(e)))
if r.short_id not in found_in_solr:
print("{} {} NOT FOUND in SOLR: adding to index".format(
r.short_id, resource.discovery_content_type))
try:
ind.update_object(r)
django_refreshed += 1
except Exception as e:
print("resource {} generated exception {}".format(
r.short_id, str(e)))
# # This always returns True whether or not SOLR needs updating
# # This is likely a Haystack bug.
# elif ind.should_update(r):
# update everything to be safe.
elif options['force']:
print("{} {}: refreshing index (forced)".format(
r.short_id, resource.discovery_content_type))
try:
ind.update_object(r)
django_refreshed += 1
except Exception as e:
print("resource {} generated exception {}".format(
r.short_id, str(e)))
print(
"Django contains {} discoverable resources and {} replaced resources"
.format(django_indexed, django_replaced))
print("{} resources in Django refreshed in SOLR".format(
django_refreshed))
# what is in SOLR that isn't in Django:
sqs = SearchQuerySet().all() # refresh for changes from above
solr_indexed = 0
solr_replaced = 0
solr_deleted = 0
for r in sqs:
try:
resource = get_resource_by_shortkey(r.short_id,
or_404=False)
repl = False
if hasattr(resource,
'metadata') and resource.metadata is not None:
repl = resource.metadata.relations.filter(
type='isReplacedBy').exists()
if not repl:
solr_indexed += 1
else:
solr_replaced += 1
except BaseResource.DoesNotExist:
print(
"SOLR resource {} ({}) NOT FOUND in Django; removing from SOLR"
.format(r.short_id, resource.discovery_content_type))
ind.remove_object(r)
solr_deleted += 1
continue
print(
"SOLR contains {} discoverable resources and {} replaced resources"
.format(solr_indexed, solr_replaced))
print("{} resources not in Django removed from SOLR".format(
solr_deleted))
def autocomplete(request):
sqs = SearchQuerySet().autocomplete(
content_auto=request.GET.get('query', ''))
template = loader.get_template('reviews/autocomplete_template.html')
return HttpResponse(template.render({'reviews': sqs}, request))
def test_query__in_empty_list(self):
"""Confirm that an empty list avoids a Elasticsearch exception"""
sqs = SearchQuerySet(using='elasticsearch').filter(id__in=[])
self.assertEqual(sqs.query.build_query(), u'id:(!*:*)')
def do_site_search(q, allow_redirect=False):
if q.strip() == "":
return []
results = []
from bill.models import Bill
from vote.models import Vote
if "pass" in q or "fail" in q or "vote" in q:
results.append({
"title": "Tracking Federal Legislation",
"href": "/start",
"noun": "feeds",
"results": [
{"href": f.link,
"label": f.title,
"obj": f,
"feed": f,
"secondary": False }
for f in (
Bill.EnactedBillsFeed(), Bill.ActiveBillsExceptIntroductionsFeed(), Bill.ComingUpFeed(), Vote.AllVotesFeed(),
)
]
})
from haystack.query import SearchQuerySet
from events.models import Feed
results.append({
"title": "Members of Congress, Presidents, and Vice Presidents",
"href": "/congress/members/all",
"qsarg": "name",
"noun": "Members of Congress, Presidents, or Vice Presidents",
"results": [
{"href": p.object.get_absolute_url(),
"label": p.object.name,
"obj": p.object,
"feed": p.object.get_feed(),
"secondary": p.object.get_current_role() == None }
for p in SearchQuerySet().using("person").filter(indexed_model_name__in=["Person"], content=q).order_by('-is_currently_serving', '-score')[0:9]]
})
# Skipping states for now because we might want to go to the district maps or to
# the state's main page for state legislative information.
#import us
#results.append(("States", "/congress/members", "most_recent_role_state", "states",
# sorted([{"href": "/congress/members/%s" % s, "label": us.statenames[s] }
# for s in us.statenames
# if us.statenames[s].lower().startswith(q.lower())
# ], key=lambda p : p["label"])))
from committee.models import Committee
results.append({
"title": "Congressional Committees",
"href": "/congress/committees",
"noun": "committees in Congress",
"results": sorted([
{"href": c.get_absolute_url(),
"label": c.fullname,
"feed": c.get_feed(),
"obj": c,
"secondary": c.committee != None}
for c in Committee.objects.filter(name__icontains=q, obsolete=False)
], key=lambda c : c["label"])
})
from settings import CURRENT_CONGRESS
from bill.search import parse_bill_citation
bill = parse_bill_citation(q)
if not bill or not allow_redirect:
from haystack.inputs import AutoQuery
bills = [\
{"href": b.object.get_absolute_url(),
"label": b.object.title,
"obj": b.object,
"feed": b.object.get_feed() if b.object.is_alive else None,
"secondary": b.object.congress != CURRENT_CONGRESS }
for b in SearchQuerySet().using("bill").filter(indexed_model_name__in=["Bill"], content=AutoQuery(q)).order_by('-current_status_date')[0:9]]
else:
#bills = [{"href": bill.get_absolute_url(), "label": bill.title, "obj": bill, "secondary": bill.congress != CURRENT_CONGRESS }]
return HttpResponseRedirect(bill.get_absolute_url())
results.append({
"title": "Bills and Resolutions (Federal)",
"href": "/congress/bills/browse",
"qsarg": "congress=__ALL__&text",
"noun": "federal bills or resolutions",
"results": bills})
results.append({
"title": "State Legislation",
"href": "/states/bills/browse",
"qsarg": "text",
"noun": "state legislation",
"results": [
{"href": p.object.get_absolute_url(),
"label": p.object.short_display_title,
"obj": p.object,
"feed": Feed(feedname="states_bill:%d" % p.object.id),
"secondary": True }
for p in SearchQuerySet().using('states').filter(indexed_model_name__in=["StateBill"], content=q)[0:9]]
})
# subject terms, but exclude subject terms that look like committee names because
# that is confusing to also see with committee results
from bill.models import BillTerm, TermType
results.append({
"title": "Subject Areas (Federal Legislation)",
"href": "/congress/bills",
"noun": "subject areas",
"results": [
{"href": p.get_absolute_url(),
"label": p.name,
"obj": p,
"feed": p.get_feed(),
"secondary": not p.is_top_term() }
for p in BillTerm.objects.filter(name__icontains=q, term_type=TermType.new).exclude(name__contains=" Committee on ")[0:9]]
})
# in each group, make sure the secondary results are placed last, but otherwise preserve order
for grp in results:
for i, obj in enumerate(grp["results"]):
obj["index"] = i
grp["results"].sort(key = lambda o : (o.get("secondary", False), o["index"]))
# sort categories first by whether all results are secondary results, then by number of matches (fewest first, if greater than zero)
results.sort(key = lambda c : (
len([d for d in c["results"] if d.get("secondary", False) == False]) == 0,
len(c["results"]) == 0,
len(c["results"])))
return results
def test_narrow_sq(self):
sqs = SearchQuerySet(using='elasticsearch').narrow(SQ(foo='moof'))
self.assertTrue(isinstance(sqs, SearchQuerySet))
self.assertEqual(len(sqs.query.narrow_queries), 1)
self.assertEqual(sqs.query.narrow_queries.pop(), 'foo:(moof)')
def search(request):
if request.method == 'GET':
q = request.GET.get('q')
models = request.GET.getlist('models')
fields = request.GET.getlist('fields')
tu_results = []
bs_results = []
bp_results = []
if 'question' in fields and 'answer' not in fields:
if 'tossup' in models:
tu_results = SearchQuerySet().filter(
tossup_text=q).models(Tossup)
if 'bonus' in models:
bs_results = SearchQuerySet().filter(
leadin_text=q).models(Bonus)
bp_results = SearchQuerySet().filter(
part_text=q).models(BonusPart)
elif 'answer' in fields and 'question' not in fields:
if 'tossup' in models:
print 'getting tossups for {} with only questions'.format(q)
tu_results = SearchQuerySet().filter(
tossup_answer=q).models(Tossup)
if 'bonus' in models:
bp_results = SearchQuerySet().filter(
bonus_answer=q).models(BonusPart)
elif 'answer' in fields and 'question' in fields:
if 'tossup' in models:
print 'getting tossups for {} with questions and answers'.format(
q)
tu_results = SearchQuerySet().filter(content=q).models(Tossup)
if 'bonus' in models:
print 'getting bonuses for {} with questions and answers'.format(
q)
bs_results = SearchQuerySet().filter(content=q).models(Bonus)
bp_results = SearchQuerySet().filter(
content=q).models(BonusPart)
tossups_json = json.loads(
serializers.serialize('json', [r.object for r in tu_results]))
for tossup in tossups_json:
tossup['fields']['tournament_name'] = \
Tournament.objects.get(id=tossup['fields']['tournament']).tournament_name
tossup['fields']['author'] = Packet.objects.get(
id=tossup['fields']['packet']).author
tossup['id'] = tossup.pop('pk')
tossup['fields']['tour_id'] = tossup['fields'].pop('tournament')
tossup['fields']['pack_id'] = tossup['fields'].pop('packet')
for key, value in tossup['fields'].items():
tossup[key] = value
del tossup['fields']
# print q, tossups_json
bonuses = [r.object for r in bs_results]
for r in bp_results:
if r.object.bonus not in bonuses:
bonuses.append(r.object.bonus)
bonuses_json = json.loads(serializers.serialize('json', bonuses))
for bonus in bonuses_json:
bonus_parts = BonusPart.objects.filter(bonus__id=bonus['pk'])
bpart_json = json.loads(serializers.serialize('json', bonus_parts))
for bonus_part in bpart_json:
for key, value in bonus_part['fields'].items():
bonus_part[key] = value
del bonus_part['fields']
bonus['fields']['bonus_parts'] = bpart_json
bonus['fields']['tournament_name'] = \
Tournament.objects.get(id=bonus['fields']['tournament']).tournament_name
bonus['fields']['author'] = Packet.objects.get(
id=bonus['fields']['packet']).author
bonus['id'] = bonus.pop('pk')
bonus['fields']['tour_id'] = bonus['fields'].pop('tournament')
bonus['fields']['pack_id'] = bonus['fields'].pop('packet')
for key, value in bonus['fields'].items():
bonus[key] = value
del bonus['fields']
return HttpResponse(json.dumps({
'tossups': tossups_json,
'bonuses': bonuses_json
}),
content_type='application/json')
def map_search_endpoint(request, filter_group_id=None):
""" Maps API search endpoint using haystack search results. For parameters see ``MAP_SEARCH_PARAMETERS``
returns JSON with the contents of type ``HaystackMapResult``
@param filter_group_id: Will filter all items by group relation, where applicable
(i.e. users are filtered by group memberships for that group, events as events in that group)
"""
implicit_ignore_location = not any([
loc_param in request.GET
for loc_param in ['sw_lon', 'sw_lat', 'ne_lon', 'ne_lat']
])
params = _collect_parameters(request.GET, MAP_SEARCH_PARAMETERS)
query = force_text(params['q'])
limit = params['limit']
page = params['page']
item_id = params['item']
if params.get('cloudfiles', False):
return map_cloudfiles_endpoint(request, query, limit, page)
# TODO: set to params['external'] after the external switch button is in frontend!
external = settings.COSINNUS_EXTERNAL_CONTENT_ENABLED
prefer_own_portal = getattr(settings, 'MAP_API_HACKS_PREFER_OWN_PORTAL',
False)
if not is_number(limit) or limit < 0:
return HttpResponseBadRequest(
'``limit`` param must be a positive number or 0!')
limit = min(limit, SERVER_SIDE_SEARCH_LIMIT)
if not is_number(page) or page < 0:
return HttpResponseBadRequest(
'``page`` param must be a positive number or 0!')
# filter for requested model types
model_list = [
klass for klass, param_name in list(SEARCH_MODEL_NAMES.items())
if params.get(param_name, False)
]
sqs = SearchQuerySet().models(*model_list)
# filter for map bounds (Points are constructed ith (lon, lat)!!!)
if not params['ignore_location'] and not implicit_ignore_location:
sqs = sqs.within('location', Point(params['sw_lon'], params['sw_lat']),
Point(params['ne_lon'], params['ne_lat']))
# filter for user's own content
if params['mine'] and request.user.is_authenticated:
user_id = request.user.id
sqs = sqs.filter_and(
Q(creator=user_id) | Q(user_id=user_id) | Q(group_members=user_id))
# filter for search terms
if query:
sqs = sqs.auto_query(query)
# group-filtered-map view for on-group pages
if filter_group_id:
group = get_object_or_None(get_cosinnus_group_model(),
id=filter_group_id)
if group:
filtered_groups = [filter_group_id]
# get child projects of this group
filtered_groups += [
subproject.id for subproject in group.get_children()
if subproject.is_active
]
sqs = sqs.filter_and(
Q(membership_groups__in=filtered_groups)
| Q(group__in=filtered_groups))
# filter topics
topics = ensure_list_of_ints(params.get('topics', ''))
if topics:
sqs = sqs.filter_and(mt_topics__in=topics)
if settings.COSINNUS_ENABLE_SDGS:
sdgs = ensure_list_of_ints(params.get('sdgs', ''))
if sdgs:
sqs = sqs.filter_and(sdgs__in=sdgs)
if settings.COSINNUS_MANAGED_TAGS_ENABLED:
managed_tags = ensure_list_of_ints(params.get('managed_tags', ''))
if managed_tags:
sqs = sqs.filter_and(managed_tags__in=managed_tags)
# filter for portal visibility
sqs = filter_searchqueryset_for_portal(
sqs,
restrict_multiportals_to_current=prefer_own_portal,
external=external)
# filter for read access by this user
sqs = filter_searchqueryset_for_read_access(sqs, request.user)
# filter events by upcoming status and exclude hidden proxies
if params['events'] and Event is not None:
sqs = filter_event_searchqueryset_by_upcoming(sqs).exclude(
is_hidden_group_proxy=True)
# filter all default user groups if the new dashboard is being used (they count as "on plattform" and aren't shown)
if getattr(settings, 'COSINNUS_USE_V2_DASHBOARD', False):
sqs = sqs.exclude(is_group_model=True,
slug__in=get_default_user_group_slugs())
# kip score sorting and only rely on natural ordering?
skip_score_sorting = False
# if we hae no query-boosted results, use *only* our custom sorting (haystack's is very random)
if not query:
sort_args = ['-local_boost']
# if we only look at conferences, order them by their from_date, future first!
if prefer_own_portal:
sort_args = ['-portal'] + sort_args
"""
# this would be the way to force-sort a content type by a natural ordering instead of score if its the only type being shown
if params.get('conferences', False) and sum([1 if params.get(content_key, False) else 0 for content_key in MAP_CONTENT_TYPE_SEARCH_PARAMETERS.keys()]) == 1:
sort_args = ['-from_date'] + sort_args
skip_score_sorting = True
sqs = sqs.order_by(*sort_args)
"""
# sort results into one list per model
total_count = sqs.count()
sqs = sqs[limit * page:limit * (page + 1)]
results = []
for i, result in enumerate(sqs):
if skip_score_sorting:
# if we skip score sorting and only rely on the natural ordering, we make up fake high scores
result.score = 100000 - (limit * page) - i
elif not query:
# if we hae no query-boosted results, use *only* our custom sorting (haystack's is very random)
result.score = result.local_boost
if prefer_own_portal and is_number(result.portal) and int(
result.portal) == CosinnusPortal.get_current().id:
result.score += 100.0
results.append(HaystackMapResult(result, user=request.user))
# if the requested item (direct select) is not in the queryset snippet
# (might happen because of an old URL), then mix it in as first item and drop the last
if item_id:
item_id = str(item_id)
if not any([res['id'] == item_id for res in results]):
item_result = get_searchresult_by_itemid(item_id, request.user)
if item_result:
results = [HaystackMapResult(item_result, user=request.user)
] + results[:-1]
page_obj = None
if results:
page_obj = {
'index': page,
'count': len(results),
'total_count': total_count,
'start': (limit * page) + 1,
'end': (limit * page) + len(results),
'has_next': total_count > (limit * (page + 1)),
'has_previous': page > 0,
}
data = {
'results': results,
'page': page_obj,
}
return JsonResponse(data)
def test_query__in(self):
sqs = SearchQuerySet(using='elasticsearch').filter(id__in=[1, 2, 3])
self.assertEqual(sqs.query.build_query(), u'id:("1" OR "2" OR "3")')
from core.views import CustomRegistrationView
from data_set_manager.api import (AssayResource, AttributeOrderResource,
AttributeResource, InvestigationResource,
ProtocolReferenceParameterResource,
ProtocolReferenceResource, ProtocolResource,
PublicationResource, StudyResource)
from data_set_manager.urls import data_set_manager_router
from file_store.urls import file_store_router
from tool_manager.urls import django_docker_engine_url, tool_manager_router
from user_files_manager.urls import (user_files_csv_url, user_files_router,
user_files_url)
logger = logging.getLogger(__name__)
# NG: facets for Haystack
sqs = (SearchQuerySet().using("core").models(DataSet).facet(
'measurement').facet('technology').highlight())
# NG: added for tastypie URL
v1_api = Api(api_name='v1')
v1_api.register(AnalysisResource())
v1_api.register(ProjectResource())
v1_api.register(StudyResource())
v1_api.register(AssayResource())
v1_api.register(DataSetResource())
v1_api.register(AttributeOrderResource())
v1_api.register(NodeResource())
v1_api.register(NodeSetResource())
v1_api.register(NodeSetListResource())
v1_api.register(NodePairResource())
v1_api.register(NodeRelationshipResource())
class MySearchView(SearchView):
context_object_name = "search_list"
paginate_by = getattr(settings, "BASE_PAGE_BY", None)
paginate_orphans = getattr(settings, "BASE_ORPHANS", 0)
queryset = SearchQuerySet().order_by("-views")
def vocabulary_search(request): # view for search in vocabularies - remembers selection (vocabulary - class - property)
# get query parameter
if 'q' in request.GET:
q_in = request.GET['q']
else:
q_in = ''
if 'page' in request.GET:
try:
page = int(request.GET['page'])
except ValueError:
page = 1
else:
page = 1
# translate non english terms
if 'translate' in request.GET:
translate = True
# create a unique translator object to be used
translator = Translator(MS_TRANSLATOR_UID, MS_TRANSLATOR_SECRET)
q = translator.translate(text=q_in, to_lang='en', from_lang=None)
if q.startswith("TranslateApiException:"):
q = q_in
else:
translate = False
q = q_in
# get requested type
if 'type' in request.GET:
tp = request.GET['type']
else:
tp = "vocabularies"
# load the query set
if tp == "vocabularies":
clsname = 'Vocabulary'
sqs = SearchQuerySet().models(Vocabulary).filter(content=q)
elif tp == "classes":
clsname = 'VocabularyClass'
sqs = SearchQuerySet().models(VocabularyClass).filter(content=q)
elif tp == "properties":
clsname = 'VocabularyProperty'
sqs = SearchQuerySet().models(VocabularyProperty).filter(content=q)
else:
raise Http404
# remove non existing objects (may have been deleted but are still indexed)
obj_set = []
for res in sqs:
if res.object:
obj_set.append(res)
# search only inside a vocabulary
if request.GET.get('definedBy'):
defined_by = int(request.GET.get('definedBy'))
obj_set_old = obj_set[:]
obj_set = []
for res in obj_set_old:
try:
if res.object.vocabulary.id == defined_by:
obj_set.append(res)
except AttributeError:
continue # only return classes or properties
else:
defined_by = None
# order the results
if tp == "vocabularies":
qs = sorted(obj_set, key=attrgetter('object.lodRanking'), reverse=True) # order objects manually
elif tp == "classes":
qs = sorted(obj_set, key=attrgetter('object.vocabulary.lodRanking'), reverse=True)
elif tp == "properties":
qs = sorted(obj_set, key=attrgetter('object.vocabulary.lodRanking'), reverse=True)
# paginate the results
paginator = Paginator(qs, 15)
page_object = paginator.page(page)
# pass parameters and render the search template
params = {'q': q, 'type': tp, 'query': True, 'translate': translate,
'page_obj': page_object, 'url': "/vocabularies/?q=" + q + '&type=' + tp}
if defined_by:
params['vocabulary_define'] = Vocabulary.objects.get(pk=defined_by)
return render(request, 'search/search.html', params)
def _query_results(query, person):
"""
Actually build the query results for this person.
Make sure any result.content_type values are reflected in RESULT_TYPE_DISPLAY for display to the user.
"""
if len(query) < 2:
return []
query = query.replace('@sfu.ca', '') # hack to make email addresses searchable as userids
query = Clean(query)
# offerings person was a member of (coredata.CourseOffering)
if person:
members = Member.objects.filter(person=person).exclude(role='DROP').select_related('offering')
offering_slugs = set(m.offering.slug for m in members)
offering_results = SearchQuerySet().models(CourseOffering).filter(text=query) # offerings that match the query
offering_results = offering_results.filter(slug__in=offering_slugs) # ... and this person was in
else:
members = []
offering_results = []
# pages this person can view (pages.Page)
page_acl = set(['ALL'])
for m in members:
# builds a set of offering_slug+"_"+acl_value strings, which will match the permission_key field in the index
member_acl = set("%s_%s" % (m.offering.slug, acl) for acl in ACL_ROLES[m.role] if acl != 'ALL')
page_acl |= member_acl
page_results = SearchQuerySet().models(Page).filter(text=query) # pages that match the query
page_results = page_results.filter(permission_key__in=page_acl) # ... and are visible to this user
# discussion this person can view (discussion.DiscussionTopic)
if person:
discuss_results = SearchQuerySet().models(DiscussionTopic).filter(text=query) # discussions that match the query
discuss_results = discuss_results.filter(slug__in=offering_slugs) # ... and this person was in
else:
discuss_results = []
# students taught by instructor (coredata.Member)
instr_members = Member.objects.filter(person=person, role__in=['INST','TA']).exclude(offering__component='CAN') \
.select_related('offering')
if person and instr_members:
offering_slugs = set(m.offering.slug for m in instr_members)
member_results = SearchQuerySet().models(Member).filter(text=query) # members that match the query
member_results = member_results.filter(offering_slug__in=offering_slugs) # ... and this person was the instructor for
member_results = member_results.load_all()
else:
member_results = []
# combine and limit to best results
results = itertools.chain(
offering_results[:MAX_RESULTS],
page_results[:MAX_RESULTS],
member_results[:MAX_RESULTS],
discuss_results[:MAX_RESULTS],
)
results = (r for r in results if r is not None)
results = list(results)
results.sort(key=lambda result: -result.score)
results = results[:MAX_RESULTS] # (list before this could be n*MAX_RESULTS long)
return results
def get_popular_videos(cls, sort='-week_views'):
return SearchQuerySet().result_class(VideoSearchResult) \
.models(Video).order_by(sort)
def __init__(self, haystack_result, obj, user, *args, **kwargs):
message_url = None
if not settings.COSINNUS_IS_INTEGRATED_PORTAL and not 'cosinnus_message' in settings.COSINNUS_DISABLED_COSINNUS_APPS:
if settings.COSINNUS_ROCKET_ENABLED:
message_url = reverse('cosinnus:message-write-group',
kwargs={'slug': obj.slug})
else:
group_admins = list(obj.actual_admins)
message_url = message_group_admins_url(obj, group_admins)
kwargs.update({
'is_member':
check_ug_membership(user, obj),
'is_pending':
check_ug_pending(user, obj),
'is_invited':
check_ug_invited_pending(user, obj),
'action_url_1':
_prepend_url(user, obj.portal) +
group_aware_reverse('cosinnus:group-microsite',
kwargs={'group': obj},
skip_domain=True) + '?join=1',
'action_url_2': (_prepend_url(user, obj.portal) +
message_url) if message_url else None,
'youtube_url':
obj.video,
'twitter_username':
obj.twitter_username,
'flickr_url':
obj.flickr_url,
'website_url':
obj.website,
'contact':
linebreaksbr(escape(obj.contact_info)),
'followed':
obj.is_user_following(user),
'starred':
obj.is_user_starring(user)
})
""" TODO: check all read permissions on related objects! """
# collect upcoming and visible project/group events
sqs = SearchQuerySet().models(SEARCH_MODEL_NAMES_REVERSE['events'])
sqs = sqs.filter_and(group=obj.id)
sqs = filter_searchqueryset_for_read_access(sqs, user)
sqs = filter_event_searchqueryset_by_upcoming(sqs)
sqs = sqs.order_by('from_date')
kwargs.update(
{'events': [HaystackEventMapCard(result) for result in sqs]})
# collect administrator users. these are *not* filtered by visibility, as project admins are always visible!
sqs = SearchQuerySet().models(SEARCH_MODEL_NAMES_REVERSE['people'])
sqs = sqs.filter_and(admin_groups=obj.id)
#sqs = filter_searchqueryset_for_read_access(sqs, user)
sqs = sqs.order_by('title')
# private users are not visible to anonymous users, BUT they are visible to logged in users!
# because if a user chose to make his group visible, he has to take authorship responsibilities
if not user.is_authenticated:
sqs = filter_searchqueryset_for_read_access(sqs, user)
kwargs.update(
{'admins': [HaystackUserMapCard(result) for result in sqs]})
if settings.COSINNUS_ORGANIZATIONS_ENABLED:
sqs = SearchQuerySet().models(
SEARCH_MODEL_NAMES_REVERSE['organizations'])
sqs = sqs.filter_and(groups=obj.id)
sqs = filter_searchqueryset_for_read_access(sqs, user)
sqs = sqs.order_by('title')
kwargs.update({
'organizations':
[HaystackOrganizationMapCard(result) for result in sqs]
})
return super(DetailedBaseGroupMapResult,
self).__init__(haystack_result, obj, user, *args,
**kwargs)
def obj_get(self, request=None, **kwargs):
'''
Retrieves a detailed search item
'''
return SearchQuerySet().filter(id=kwargs['id'])[0]
def get_latest_videos(cls):
return SearchQuerySet().result_class(VideoSearchResult) \
.models(Video).order_by('-created')
def search(request,
template='search/search.html',
load_all=True,
form_class=ModelSearchForm,
searchqueryset=None,
context_class=RequestContext,
extra_context=None,
results_per_page=None):
query = ''
results = EmptySearchQuerySet()
if request.GET.get('q'):
searchqueryset = SearchQuerySet() \
.filter(Q(public=True) | Q(author=request.user)) \
.order_by('-pub_date')
if request.user.is_authenticated() and \
'mine-only' in request.GET:
searchqueryset = SearchQuerySet().filter(author=request.user) \
.order_by('-pub_date')
elif request.user.is_authenticated() and \
('author' in request.GET and
request.GET.get('author')):
author = request.GET.get('author')
if author == request.user.username:
searchqueryset = SearchQuerySet().filter(author=request.user) \
.order_by('-pub_date')
else:
team = get_object_or_None(Team, slug=author)
if team and team.user_is_member(request.user):
searchqueryset = SearchQuerySet().filter(author=team) \
.order_by('-pub_date')
form = ModelSearchForm(request.GET,
searchqueryset=searchqueryset,
load_all=load_all)
if form.is_valid():
query = form.cleaned_data['q']
results = form.search()
else:
form = form_class(searchqueryset=searchqueryset, load_all=load_all)
paginator = Paginator(results, results_per_page or RESULTS_PER_PAGE)
try:
page = paginator.page(int(request.GET.get('page', 1)))
except InvalidPage:
raise Http404("No such page of results!")
context = {
'form': form,
'has_snipts': True,
'page': page,
'paginator': paginator,
'query': query,
'suggestion': None,
}
if results.query.backend.include_spelling:
context['suggestion'] = form.get_suggestion()
if extra_context:
context.update(extra_context)
return render(request, template, context)
def get_featured_videos(cls):
return SearchQuerySet().result_class(VideoSearchResult) \
.models(Video).filter(featured__gt=datetime.datetime(datetime.MINYEAR, 1, 1)) \
.order_by('-featured')
