def mlt(self):
"""Returns a search with a morelikethis query for docs like this"""
# Short responses tend to not repeat any words, so then MLT
# returns nothing. This fixes that by setting min_term_freq to
# 1. Longer responses tend to repeat important words, so we can
# set min_term_freq to 2.
num_words = len(self.description.split(' '))
if num_words > 40:
min_term_freq = 2
else:
min_term_freq = 1
s = self.search()
if self.product:
s = s.filter('term', product=self.product)
if self.platform:
s = s.filter('term', platform=self.platform)
s = s.query(
'more_like_this',
fields=['description'],
docs=[
{
'_index': get_index_name(),
'_type': self._doc_type.name,
'_id': self.id
}
],
min_term_freq=min_term_freq,
stop_words=list(ANALYSIS_STOPWORDS)
)
return s
def mlt(self):
"""Returns a search with a morelikethis query for docs like this"""
# Short responses tend to not repeat any words, so then MLT
# returns nothing. This fixes that by setting min_term_freq to
# 1. Longer responses tend to repeat important words, so we can
# set min_term_freq to 2.
num_words = len(self.description.split(' '))
if num_words > 40:
min_term_freq = 2
else:
min_term_freq = 1
s = self.search()
if self.product:
s = s.filter('term', product=self.product)
if self.platform:
s = s.filter('term', platform=self.platform)
s = s.query('more_like_this',
fields=['description'],
docs=[{
'_index': get_index_name(),
'_type': self._doc_type.name,
'_id': self.id
}],
min_term_freq=min_term_freq,
stop_words=list(ANALYSIS_STOPWORDS))
return s
def teardown_indexes(self):
try:
get_es().indices.delete(get_index_name())
except NotFoundError:
# If we get this error, it means the index didn't exist
# so there's nothing to delete.
pass
def teardown_indexes(self):
try:
get_es().indices.delete(get_index_name())
except NotFoundError:
# If we get this error, it means the index didn't exist
# so there's nothing to delete.
pass
def test_index_chunk_task(self):
responses = ResponseFactory.create_batch(10)
# With live indexing, that'll create items in the index. Since
# we want to test index_chunk_test, we need a clean index to
# start with so we delete and recreate it.
self.setup_indexes(empty=True)
# Verify there's nothing in the index.
assert ResponseDocType.docs.search().count() == 0
# Create the record and the chunk and then run it through
# celery.
batch_id = 'ou812'
rec = RecordFactory(batch_id=batch_id)
chunk = (
to_class_path(ResponseDocType),
[item.id for item in responses]
)
index_chunk_task.delay(get_index_name(), batch_id, rec.id, chunk)
self.refresh()
# Verify everything is in the index now.
assert ResponseDocType.docs.search().count() == 10
# Verify the record was marked succeeded.
rec = Record.objects.get(pk=rec.id)
assert rec.status == Record.STATUS_SUCCESS
def test_index_chunk_task(self):
responses = ResponseFactory.create_batch(10)
# With live indexing, that'll create items in the index. Since
# we want to test index_chunk_test, we need a clean index to
# start with so we delete and recreate it.
self.setup_indexes(empty=True)
# Verify there's nothing in the index.
assert ResponseDocType.docs.search().count() == 0
# Create the record and the chunk and then run it through
# celery.
batch_id = 'ou812'
rec = RecordFactory(batch_id=batch_id)
chunk = (to_class_path(ResponseDocType),
[item.id for item in responses])
index_chunk_task.delay(get_index_name(), batch_id, rec.id, chunk)
self.refresh()
# Verify everything is in the index now.
assert ResponseDocType.docs.search().count() == 10
# Verify the record was marked succeeded.
rec = Record.objects.get(pk=rec.id)
assert rec.status == Record.STATUS_SUCCESS
def search_admin_view(request):
"""Render the admin view containing search tools"""
error_messages = []
stats = None
es_deets = None
indexes = []
try:
if 'reset' in request.POST:
return handle_reset(request)
if 'reindex' in request.POST:
return handle_reindex(request)
if 'recreate_reindex' in request.POST:
return handle_recreate_reindex(request)
except Exception as exc:
error_messages.append(u'Error: %s' % exc.message)
try:
# This gets index stats, but also tells us whether ES is in
# a bad state.
try:
stats = get_index_stats()
except NotFoundError:
stats = None
indexes = get_indexes()
indexes.sort(key=lambda m: m[0])
# TODO: Input has a single ES_URL and that's the ZLB and does
# the balancing. If that ever changes and we have multiple
# ES_URLs, then this should get fixed.
es_deets = requests.get(settings.ES_URLS[0]).json()
except ConnectionError:
error_messages.append('Error: Elastic Search is not set up on this '
'machine or timed out trying to respond. '
'(ConnectionError/Timeout)')
except NotFoundError:
error_messages.append('Error: Index is missing. Press the reindex '
'button below. (ElasticHttpNotFoundError)')
outstanding_records = Record.outstanding()
recent_records = Record.objects.order_by('-creation_time')[:100]
return render(
request, 'admin/search_admin_view.html', {
'title': 'Search',
'es_deets': es_deets,
'doc_type_stats': stats,
'indexes': indexes,
'index': get_index_name(),
'error_messages': error_messages,
'recent_records': recent_records,
'outstanding_records': outstanding_records,
'now': datetime.now(),
})
def search_admin_view(request):
"""Render the admin view containing search tools"""
error_messages = []
stats = None
es_deets = None
indexes = []
try:
if 'reset' in request.POST:
return handle_reset(request)
if 'reindex' in request.POST:
return handle_reindex(request)
if 'recreate_reindex' in request.POST:
return handle_recreate_reindex(request)
except Exception as exc:
error_messages.append(u'Error: %s' % exc.message)
try:
# This gets index stats, but also tells us whether ES is in
# a bad state.
try:
stats = get_index_stats()
except NotFoundError:
stats = None
indexes = get_indexes()
indexes.sort(key=lambda m: m[0])
# TODO: Input has a single ES_URL and that's the ZLB and does
# the balancing. If that ever changes and we have multiple
# ES_URLs, then this should get fixed.
es_deets = requests.get(settings.ES_URLS[0]).json()
except ConnectionError:
error_messages.append('Error: Elastic Search is not set up on this '
'machine or timed out trying to respond. '
'(ConnectionError/Timeout)')
except NotFoundError:
error_messages.append('Error: Index is missing. Press the reindex '
'button below. (ElasticHttpNotFoundError)')
outstanding_records = Record.outstanding()
recent_records = Record.objects.order_by('-creation_time')[:100]
return render(request, 'admin/search_admin_view.html', {
'title': 'Search',
'es_deets': es_deets,
'doc_type_stats': stats,
'indexes': indexes,
'index': get_index_name(),
'error_messages': error_messages,
'recent_records': recent_records,
'outstanding_records': outstanding_records,
'now': datetime.now(),
})
def timezone_view(request):
"""Admin view showing times and timezones in data."""
# Note: This is an admin page that gets used once in a blue moon.
# As such, I'm taking some liberties (hand-indexing the response,
# time.sleep, etc) that I would never take if it was used more
# often or was viewable by users. If these two assumptions ever
# change, then this should be rewritten.
from fjord.feedback.models import (
Response,
ResponseDocType,
ResponseDocTypeManager
)
from fjord.feedback.tests import ResponseFactory
from fjord.search.index import get_es, get_index_name
server_time = datetime.now()
# Create a new response.
resp = ResponseFactory()
resp_time = resp.created
# Index the response by hand so we know it gets to
# Elasticsearch. Otherwise it gets done by celery and we don't
# know how long that'll take.
doc = ResponseDocType.extract_doc(resp)
ResponseDocTypeManager.bulk_index(docs=[doc])
# Fetch the response from the db.
resp = Response.objects.get(id=resp.id)
resp2_time = resp.created
# Refresh and sleep 5 seconds as a hand-wavey way to make sure
# that Elasticsearch has had time to refresh the index.
get_es().indices.refresh(get_index_name())
time.sleep(5)
s = ResponseDocTypeManager.search().filter('term', id=resp.id).execute()
es_time = s[0].created
# Delete the test response which also deletes it in the index.
resp.delete()
return render(request, 'admin/timezone_view.html', {
'server_time': server_time,
'resp_time': resp_time,
'resp2_time': resp2_time,
'es_time': es_time
})
def reindex():
"""Calculates and creates indexing chunks"""
index = get_index_name()
batch_id = create_batch_id()
# Break up all the things we want to index into chunks. This
# chunkifies by class then by chunk size.
chunks = []
for cls, indexable in get_indexable():
chunks.extend((cls, chunk) for chunk in chunked(indexable, CHUNK_SIZE))
for cls, id_list in chunks:
chunk_name = '%s %d -> %d' % (cls._doc_type.name, id_list[0],
id_list[-1])
rec = Record(batch_id=batch_id, name=chunk_name)
rec.save()
index_chunk_task.delay(index, batch_id, rec.id,
(to_class_path(cls), id_list))
def reindex():
"""Calculates and creates indexing chunks"""
index = get_index_name()
batch_id = create_batch_id()
# Break up all the things we want to index into chunks. This
# chunkifies by class then by chunk size.
chunks = []
for cls, indexable in get_indexable():
chunks.extend(
(cls, chunk) for chunk in chunked(indexable, CHUNK_SIZE))
for cls, id_list in chunks:
chunk_name = '%s %d -> %d' % (cls._doc_type.name,
id_list[0], id_list[-1])
rec = Record(batch_id=batch_id, name=chunk_name)
rec.save()
index_chunk_task.delay(index, batch_id, rec.id,
(to_class_path(cls), id_list))
def monitor_view(request):
"""View for services monitor."""
# Dict of infrastructure name -> list of output tuples of (INFO,
# msg) or (ERROR, msg)
status = {}
# Note: To add a new component, do your testing and then add a
# name -> list of output tuples map to status.
# Check memcached.
memcache_results = []
try:
for cache_name, cache_props in settings.CACHES.items():
result = True
backend = cache_props['BACKEND']
location = cache_props['LOCATION']
# LOCATION can be a string or a list of strings
if isinstance(location, basestring):
location = location.split(';')
if 'memcache' in backend:
for loc in location:
# TODO: this doesn't handle unix: variant
ip, port = loc.split(':')
result = test_memcached(ip, int(port))
memcache_results.append(
(INFO, '%s:%s %s' % (ip, port, result)))
if not memcache_results:
memcache_results.append((ERROR, 'memcache is not configured.'))
elif len(memcache_results) < 2:
memcache_results.append(
(ERROR, ('You should have at least 2 memcache servers. '
'You have %s.' % len(memcache_results))))
else:
memcache_results.append((INFO, 'memcached servers look good.'))
except Exception as exc:
memcache_results.append(
(ERROR, 'Exception while looking at memcached: %s' % str(exc)))
status['memcached'] = memcache_results
# Check ES.
es_results = []
try:
get_index_stats()
es_results.append(
(INFO, ('Successfully connected to ElasticSearch and index '
'exists.')))
except ConnectionError as exc:
es_results.append(
(ERROR, 'Cannot connect to ElasticSearch: %s' % str(exc)))
except NotFoundError:
es_results.append(
(ERROR, 'Index "%s" missing.' % get_index_name()))
except Exception as exc:
es_results.append(
(ERROR, 'Exception while looking at ElasticSearch: %s' % str(exc)))
status['ElasticSearch'] = es_results
# Check RabbitMQ.
rabbitmq_results = []
try:
rabbit_conn = establish_connection(connect_timeout=2)
rabbit_conn.connect()
rabbitmq_results.append(
(INFO, 'Successfully connected to RabbitMQ.'))
except (socket.error, IOError) as exc:
rabbitmq_results.append(
(ERROR, 'Error connecting to RabbitMQ: %s' % str(exc)))
except Exception as exc:
rabbitmq_results.append(
(ERROR, 'Exception while looking at RabbitMQ: %s' % str(exc)))
status['RabbitMQ'] = rabbitmq_results
status_code = 200
status_summary = {}
for component, output in status.items():
if ERROR in [item[0] for item in output]:
status_code = 500
status_summary[component] = False
else:
status_summary[component] = True
return render(request, 'services/monitor.html',
{'component_status': status,
'status_summary': status_summary},
status=status_code)
