def admin_collection_template(request, collection_id):
hue_collection = Collection.objects.get(id=collection_id)
solr_collection = SolrApi(SOLR_URL.get(), request.user).collection_or_core(hue_collection)
sample_data = {}
if request.method == 'POST':
hue_collection.result.update_from_post(request.POST)
hue_collection.result.save()
return HttpResponse(json.dumps({}), mimetype="application/json")
solr_query = {}
solr_query['collection'] = hue_collection.name
solr_query['q'] = ''
solr_query['fq'] = ''
solr_query['rows'] = 5
solr_query['start'] = 0
solr_query['facets'] = 0
try:
response = SolrApi(SOLR_URL.get(), request.user).query(solr_query, hue_collection)
sample_data = json.dumps(response["response"]["docs"])
except PopupException, e:
message = e
try:
message = json.loads(e.message.message)['error']['msg'] # Try to get the core error
except:
pass
request.error(_('No preview available, some facets are invalid: %s') % message)
LOG.exception(e)
def create_collection(self, name, fields, unique_key_field='id', df='text'):
"""
Create solr collection or core and instance dir.
Create schema.xml file so that we can set UniqueKey field.
"""
if self.is_solr_cloud_mode():
# solrcloud mode
# Need to remove path afterwards
tmp_path, solr_config_path = utils.copy_configs(fields, unique_key_field, df, True)
# Create instance directory.
solrctl_path = get_solrctl_path()
process = subprocess.Popen([solrctl_path, "instancedir", "--create", name, solr_config_path],
stdout=subprocess.PIPE,
stderr=subprocess.PIPE,
env={
'SOLR_ZK_ENSEMBLE': conf.SOLR_ZK_ENSEMBLE.get()
})
status = process.wait()
# Don't want directories laying around
shutil.rmtree(tmp_path)
if status != 0:
LOG.error("Could not create instance directory.\nOutput: %s\nError: %s" % process.communicate())
raise PopupException(_('Could not create instance directory. '
'Check if solr_zk_ensemble and solrctl_path are correct in Hue config [indexer].'))
api = SolrApi(SOLR_URL.get(), self.user, SECURITY_ENABLED.get())
if not api.create_collection(name):
# Delete instance directory if we couldn't create a collection.
process = subprocess.Popen([solrctl_path, "instancedir", "--delete", name],
stdout=subprocess.PIPE,
stderr=subprocess.PIPE,
env={
'SOLR_ZK_ENSEMBLE': conf.SOLR_ZK_ENSEMBLE.get()
})
if process.wait() != 0:
LOG.error("Cloud not delete collection.\nOutput: %s\nError: %s" % process.communicate())
raise PopupException(_('Could not create collection. Check error logs for more info.'))
else:
# Non-solrcloud mode
# Create instance directory locally.
instancedir = os.path.join(conf.CORE_INSTANCE_DIR.get(), name)
if os.path.exists(instancedir):
raise PopupException(_("Instance directory %s already exists! Please remove it from the file system.") % instancedir)
tmp_path, solr_config_path = utils.copy_configs(fields, unique_key_field, df, False)
shutil.move(solr_config_path, instancedir)
shutil.rmtree(tmp_path)
api = SolrApi(SOLR_URL.get(), self.user, SECURITY_ENABLED.get())
if not api.create_core(name, instancedir):
# Delete instance directory if we couldn't create a collection.
shutil.rmtree(instancedir)
raise PopupException(_('Could not create collection. Check error logs for more info.'))
def fields_data(self, user):
schema_fields = SolrApi(SOLR_URL.get(), user).fields(self.name)
schema_fields = schema_fields['schema']['fields']
dynamic_fields = SolrApi(SOLR_URL.get(), user).fields(self.name, dynamic=True)
dynamic_fields = dynamic_fields['fields']
schema_fields.update(dynamic_fields)
return sorted([{'name': str(field), 'type': str(attributes.get('type', ''))}
for field, attributes in schema_fields.iteritems()])
def get_fields(self, collection_or_core_name):
try:
field_data = SolrApi(SOLR_URL.get(), self.user, SECURITY_ENABLED.get()).fields(collection_or_core_name)
fields = self._format_flags(field_data['schema']['fields'])
except:
LOG.exception(_('Could not fetch fields for collection %s.') % collection_or_core_name)
raise PopupException(_('Could not fetch fields for collection %s. See logs for more info.') % collection_or_core_name)
try:
uniquekey = SolrApi(SOLR_URL.get(), self.user, SECURITY_ENABLED.get()).uniquekey(collection_or_core_name)
except:
LOG.exception(_('Could not fetch unique key for collection %s.') % collection_or_core_name)
raise PopupException(_('Could not fetch unique key for collection %s. See logs for more info.') % collection_or_core_name)
return uniquekey, fields
def create_collection(self, name, fields, unique_key_field='id', df='text'):
"""
Create solr collection or core and instance dir.
Create schema.xml file so that we can set UniqueKey field.
"""
if self.is_solr_cloud_mode():
# solrcloud mode
# Need to remove path afterwards
tmp_path, solr_config_path = copy_configs(fields, unique_key_field, df, True)
zc = ZookeeperClient(hosts=get_solr_ensemble(), read_only=False)
root_node = '%s/%s' % (ZK_SOLR_CONFIG_NAMESPACE, name)
config_root_path = '%s/%s' % (solr_config_path, 'conf')
try:
zc.copy_path(root_node, config_root_path)
except Exception, e:
zc.delete_path(root_node)
raise PopupException(_('Error in copying Solr configurations.'), detail=e)
# Don't want directories laying around
shutil.rmtree(tmp_path)
api = SolrApi(SOLR_URL.get(), self.user, SECURITY_ENABLED.get())
if not api.create_collection(name):
# Delete instance directory if we couldn't create a collection.
try:
zc.delete_path(root_node)
except Exception, e:
raise PopupException(_('Error in deleting Solr configurations.'), detail=e)
def get_terms(request):
result = {"status": -1, "message": "Error"}
try:
collection = json.loads(request.POST.get("collection", "{}"))
analysis = json.loads(request.POST.get("analysis", "{}"))
field = analysis["name"]
properties = {
"terms.limit": 25,
"terms.prefix": analysis["terms"]["prefix"]
# lower
# limit
# mincount
# maxcount
}
result["terms"] = SolrApi(SOLR_URL.get(), request.user).terms(collection["name"], field, properties)
result["terms"] = pairwise2(field, [], result["terms"]["terms"][field])
result["status"] = 0
result["message"] = ""
except Exception, e:
result["message"] = force_unicode(e)
if "not currently supported" in result["message"]:
result["status"] = 1
result["message"] = _("This field does not support stats")
def _fetch_collections(request):
from libsolr.api import SolrApi
from search.conf import SOLR_URL
path = request.GET['path']
item = None
name = None
if path:
item = path
if '/' in path:
item, name = path.split('/')
api = SolrApi(SOLR_URL.get(), request.user)
if not item:
return {"databases": ["collections", "configs", "admin"]}
elif item and name:
return {"authorizable_link": "/indexer/#edit/%s" % name, "extended_columns": [], "columns": [], "partition_keys": []}
elif item == 'collections':
return {"tables_meta": [{"comment": None, "type": "Table", "name": col} for col in api.collections2()]}
elif item == 'configs':
return {"tables_meta": [{"comment": None, "type": "Table", "name": conf} for conf in api.configs()]}
elif item == 'admin':
return {"tables_meta": [{"comment": None, "type": "Table", "name": 'collections'}, {"comment": None, "type": "Table", "name": "cores"}]}
else:
raise PopupException(_('Authorizable %s could not be retrieved') % path)
def test_query(self):
collection = Collection2(user=self.user, name='log_analytics_demo')
collection = json.loads(collection.get_json(self.user))
query = {'qs': [{'q': ''}], 'fqs': [], 'start': 0}
SolrApi(SOLR_URL.get(), self.user).query(collection['collection'], query)
def update_document(request):
result = {'status': -1, 'message': 'Error'}
if not can_edit_index(request.user):
result['message'] = _('Permission to edit the document denied')
return JsonResponse(result)
try:
collection = json.loads(request.POST.get('collection', '{}'))
document = json.loads(request.POST.get('document', '{}'))
doc_id = request.POST.get('id')
if document['hasChanged']:
edits = {
"id": doc_id,
}
version = None # If there is a version, use it to avoid potential concurrent update conflicts
for field in document['details']:
if field['hasChanged']:
edits[field['key']] = {"set": field['value']}
if field['key'] == '_version_':
version = field['value']
if SolrApi(SOLR_URL.get(), request.user).update(collection['name'], json.dumps([edits]), content_type='json', version=version):
result['status'] = 0
result['message'] = _('Document successfully updated.')
else:
result['status'] = 0
result['message'] = _('Document has no modifications to change.')
except Exception, e:
result['message'] = force_unicode(e)
def get_all_indexes(self, show_all=False):
indexes = []
try:
indexes = self.get_solr_collection().keys()
except:
pass
try:
indexes += SolrApi(SOLR_URL.get(), self.user).aliases().keys()
except:
pass
if show_all or not indexes:
return indexes + SolrApi(SOLR_URL.get(), self.user).cores().keys()
else:
return indexes
def get_all_indexes(self, show_all=False):
indexes = []
try:
indexes = self.get_solr_collections().keys()
except:
LOG.exception('failed to get indexes')
try:
indexes += SolrApi(SOLR_URL.get(), self.user).aliases().keys()
except:
LOG.exception('failed to get index aliases')
if show_all or not indexes:
return indexes + SolrApi(SOLR_URL.get(), self.user).cores().keys()
else:
return indexes
def update_data_from_hive(self, db, collection_or_core_name, database, table, columns, indexing_strategy='upload'):
"""
Add hdfs path contents to index
"""
# Run a custom hive query and post data to collection
from beeswax.server import dbms
import tablib
api = SolrApi(SOLR_URL.get(), self.user, SECURITY_ENABLED.get())
if indexing_strategy == 'upload':
table = db.get_table(database, table)
hql = "SELECT %s FROM `%s.%s` %s" % (','.join(columns), database, table.name, db._get_browse_limit_clause(table))
query = dbms.hql_query(hql)
try:
handle = db.execute_and_wait(query)
if handle:
result = db.fetch(handle, rows=100)
db.close(handle)
dataset = tablib.Dataset()
dataset.append(columns)
for row in result.rows():
dataset.append(row)
if not api.update(collection_or_core_name, dataset.csv, content_type='csv'):
raise PopupException(_('Could not update index. Check error logs for more info.'))
else:
raise PopupException(_('Could not update index. Could not fetch any data from Hive.'))
except Exception, e:
raise PopupException(_('Could not update index.'), detail=e)
def update_data_from_hdfs(self, fs, collection_or_core_name, fields, path, data_type='separated', indexing_strategy='upload', **kwargs):
"""
Add hdfs path contents to index
"""
api = SolrApi(SOLR_URL.get(), self.user, SECURITY_ENABLED.get())
if indexing_strategy == 'upload':
stats = fs.stats(path)
if stats.size > MAX_UPLOAD_SIZE:
raise PopupException(_('File size is too large to handle!'))
else:
# Get fields for filtering
unique_key, fields = self.get_fields(collection_or_core_name)
fields = [{'name': field, 'type': fields[field]['type']} for field in fields]
fh = fs.open(path)
if data_type == 'log':
# Transform to JSON then update
data = json.dumps([value for value in field_values_from_log(fh, fields)])
content_type = 'json'
elif data_type == 'separated':
data = json.dumps([value for value in field_values_from_separated_file(fh, kwargs.get('separator', ','), kwargs.get('quote_character', '"'), fields)], indent=2)
content_type = 'json'
else:
raise PopupException(_('Could not update index. Unknown type %s') % data_type)
fh.close()
if not api.update(collection_or_core_name, data, content_type=content_type):
raise PopupException(_('Could not update index. Check error logs for more info.'))
else:
raise PopupException(_('Could not update index. Indexing strategy %s not supported.') % indexing_strategy)
def get_all_indexes(self):
indexes = []
try:
indexes = self.get_solr_collection().keys()
except:
pass
return indexes + SolrApi(SOLR_URL.get(), self.user).cores().keys()
def _create_facet(collection, user, facet_id, facet_label, facet_field, widget_type):
properties = {
'sort': 'desc',
'canRange': False,
'stacked': False,
'limit': 10,
'mincount': 0,
'isDate': False,
'andUp': False, # Not used yet
}
solr_api = SolrApi(SOLR_URL.get(), user)
range_properties = _new_range_facet(solr_api, collection, facet_field, widget_type)
if range_properties:
facet_type = 'range'
properties.update(range_properties)
elif widget_type == 'hit-widget':
facet_type = 'query'
else:
facet_type = 'field'
if widget_type == 'map-widget':
properties['scope'] = 'world'
properties['mincount'] = 1
properties['limit'] = 100
return {
'id': facet_id,
'label': facet_label,
'field': facet_field,
'type': facet_type,
'widgetType': widget_type,
'properties': properties
}
def get_terms(request):
result = {'status': -1, 'message': 'Error'}
try:
collection = json.loads(request.POST.get('collection', '{}'))
analysis = json.loads(request.POST.get('analysis', '{}'))
field = analysis['name']
properties = {
'terms.limit': 25,
'terms.prefix': analysis['terms']['prefix']
# lower
# limit
# mincount
# maxcount
}
result['terms'] = SolrApi(SOLR_URL.get(), request.user).terms(collection['name'], field, properties)
result['terms'] = pairwise2(field, [], result['terms']['terms'][field])
result['status'] = 0
result['message'] = ''
except Exception, e:
result['message'] = force_unicode(e)
if 'not currently supported' in result['message']:
result['status'] = 1
result['message'] = _('This field does not support stats')
def zkensemble():
"""
Try to guess the value if no values are specified.
"""
from django.conf import settings
if 'zookeeper' in settings.INSTALLED_APPS:
try:
# Backward compatibility until Hue 4
from zookeeper.conf import CLUSTERS
clusters = CLUSTERS.get()
if clusters['default'].HOST_PORTS.get() != 'localhost:2181':
return '%s' % clusters['default'].HOST_PORTS.get()
except:
LOG.warn('Could not get zookeeper ensemble from the zookeeper app')
if 'search' in settings.INSTALLED_APPS:
try:
from search.conf import SOLR_URL
parsed = urlparse(SOLR_URL.get())
return "%s:2181" % (parsed.hostname or 'localhost')
except:
LOG.warn('Could not get zookeeper ensemble from the search app')
return "localhost:2181"
def index(request):
hue_collections = Collection.objects.filter(enabled=True)
if not hue_collections:
if request.user.is_superuser:
return admin_collections(request, True)
else:
return no_collections(request)
initial_collection = request.COOKIES.get('hueSearchLastCollection', 0)
search_form = QueryForm(request.GET, initial_collection=initial_collection)
response = {}
error = {}
solr_query = {}
hue_collection = None
if search_form.is_valid():
collection_id = search_form.cleaned_data['collection']
solr_query['q'] = search_form.cleaned_data['query'].encode('utf8')
solr_query['fq'] = search_form.cleaned_data['fq']
if search_form.cleaned_data['sort']:
solr_query['sort'] = search_form.cleaned_data['sort']
solr_query['rows'] = search_form.cleaned_data['rows'] or 15
solr_query['start'] = search_form.cleaned_data['start'] or 0
solr_query['facets'] = search_form.cleaned_data['facets'] or 1
try:
hue_collection = Collection.objects.get(id=collection_id)
solr_query['collection'] = hue_collection.name
response = SolrApi(SOLR_URL.get(), request.user).query(solr_query, hue_collection)
except Exception, e:
error['message'] = unicode(str(e), "utf8")
def index(request):
hue_collections = Collection.objects.filter(enabled=True)
if not hue_collections:
if request.user.is_superuser:
return admin_collections(request, True)
else:
return no_collections(request)
init_collection = initial_collection(request, hue_collections)
search_form = QueryForm(request.GET, initial_collection=init_collection)
response = {}
error = {}
solr_query = {}
if search_form.is_valid():
try:
collection_id = search_form.cleaned_data['collection']
hue_collection = Collection.objects.get(id=collection_id)
solr_query = search_form.solr_query_dict
response = SolrApi(SOLR_URL.get(), request.user).query(solr_query, hue_collection)
solr_query['total_pages'] = int(math.ceil((float(response['response']['numFound']) / float(solr_query['rows']))))
solr_query['search_time'] = response['responseHeader']['QTime']
except Exception, e:
error['title'] = force_unicode(e.title) if hasattr(e, 'title') else ''
error['message'] = force_unicode(str(e))
def _create_solr_cloud_collection(self, name, fields, unique_key_field, df):
with ZookeeperClient(hosts=get_solr_ensemble(), read_only=False) as zc:
root_node = '%s/%s' % (ZK_SOLR_CONFIG_NAMESPACE, name)
tmp_path, solr_config_path = copy_configs(fields, unique_key_field, df, True)
try:
config_root_path = '%s/%s' % (solr_config_path, 'conf')
try:
zc.copy_path(root_node, config_root_path)
except Exception, e:
zc.delete_path(root_node)
raise PopupException(_('Error in copying Solr configurations.'), detail=e)
finally:
# Don't want directories laying around
shutil.rmtree(tmp_path)
api = SolrApi(SOLR_URL.get(), self.user, SECURITY_ENABLED.get())
if not api.create_collection(name):
# Delete instance directory if we couldn't create a collection.
try:
zc.delete_path(root_node)
except Exception, e:
raise PopupException(_('Error in deleting Solr configurations.'), detail=e)
raise PopupException(_('Could not create collection. Check error logs for more info.'))
def get_collections(self):
solr_collections = {}
solr_aliases = {}
solr_cores = {}
try:
api = SolrApi(SOLR_URL.get(), self.user, SECURITY_ENABLED.get())
if self.is_solr_cloud_mode():
solr_collections = api.collections()
for name in solr_collections:
solr_collections[name]['isCoreOnly'] = False
solr_aliases = api.aliases()
for name in solr_aliases:
solr_aliases[name] = {
'isCoreOnly': False,
'isAlias': True,
'collections': solr_aliases[name]
}
solr_cores = api.cores()
for name in solr_cores:
solr_cores[name]['isCoreOnly'] = True
except Exception, e:
LOG.warn('No Zookeeper servlet running on Solr server: %s' % e)
def zkensemble(): """ ZooKeeper Ensemble """ from search.conf import SOLR_URL parsed = urlparse(SOLR_URL.get()) return "%s:2181/solr" % (parsed.hostname or 'localhost')
def admin_collection_schema(request, collection_id):
hue_collection = Collection.objects.get(id=collection_id)
solr_schema = SolrApi(SOLR_URL.get(), request.user).schema(hue_collection.name)
content = {
'solr_schema': solr_schema.decode('utf-8')
}
return HttpResponse(json.dumps(content), mimetype="application/json")
def get_new_collections(self):
try:
solr_collections = SolrApi(SOLR_URL.get()).collections()
for name in Collection.objects.values_list('name', flat=True):
solr_collections.pop(name, None)
except Exception, e:
LOG.warn('No Zookeeper servlet running on Solr server: %s' % e)
solr_collections = []
def get_new_cores(self):
try:
solr_cores = SolrApi(SOLR_URL.get()).cores()
for name in Collection.objects.values_list('name', flat=True):
solr_cores.pop(name, None)
except Exception, e:
solr_cores = []
LOG.warn('No Single core setup on Solr server: %s' % e)
def is_solr_cloud_mode(self):
api = SolrApi(SOLR_URL.get(), self.user, SECURITY_ENABLED.get())
if not hasattr(self, '_solr_cloud_mode'):
try:
api.collections()
setattr(self, '_solr_cloud_mode', True)
except:
setattr(self, '_solr_cloud_mode', False)
return getattr(self, '_solr_cloud_mode')
def get_autocomplete(self):
autocomplete = {}
try:
api = SolrApi(SOLR_URL.get(), self.user, SECURITY_ENABLED.get())
autocomplete['collections'] = api.collections2()
autocomplete['configs'] = api.configs()
except Exception, e:
LOG.warn('No Zookeeper servlet running on Solr server: %s' % e)
def fields_data(self, user, name):
api = SolrApi(SOLR_URL.get(), user)
try:
schema_fields = api.fields(name)
schema_fields = schema_fields['schema']['fields']
except Exception, e:
LOG.warn('/luke call did not succeed: %s' % e)
fields = api.schema_fields(name)
schema_fields = Collection2._make_luke_from_schema_fields(fields)
def is_solr_cloud_mode(self):
api = SolrApi(SOLR_URL.get(), self.user, SECURITY_ENABLED.get())
if not hasattr(self, '_solr_cloud_mode'):
try:
api.collections()
setattr(self, '_solr_cloud_mode', True)
except Exception, e:
LOG.info('Non SolrCloud server: %s' % e)
setattr(self, '_solr_cloud_mode', False)
def admin_collection_solr_properties(request, collection_id):
hue_collection = Collection.objects.get(id=collection_id)
solr_collection = SolrApi(SOLR_URL.get(), request.user).collection_or_core(hue_collection)
content = render('admin_collection_properties_solr_properties.mako', request, {
'solr_collection': solr_collection,
'hue_collection': hue_collection,
}, force_template=True).content
return HttpResponse(json.dumps({'content': content}), mimetype="application/json")
def update_document(request):
result = {'status': -1, 'message': 'Error'}
if not can_edit_index(request.user):
result['message'] = _('Permission to edit the document denied')
return JsonResponse(result)
try:
collection = json.loads(request.POST.get('collection', '{}'))
document = json.loads(request.POST.get('document', '{}'))
doc_id = request.POST.get('id')
if document['hasChanged']:
edits = {
"id": doc_id,
}
version = None # If there is a version, use it to avoid potential concurrent update conflicts
for field in document['details']:
if field['hasChanged'] and field['key'] != '_version_':
edits[field['key']] = {"set": field['value']}
if field['key'] == '_version_':
version = field['value']
result['update'] = SolrApi(SOLR_URL.get(), request.user).update(
collection['name'],
json.dumps([edits]),
content_type='json',
version=version)
result['message'] = _('Document successfully updated.')
result['status'] = 0
else:
result['status'] = 0
result['message'] = _('Document has no modifications to change.')
except RestException, e:
try:
result['message'] = json.loads(e.message)['error']['msg']
except:
LOG.exception('Failed to parse json response')
result['message'] = force_unicode(e)
def search(request):
response = {}
collection = json.loads(request.POST.get('collection', '{}'))
query = json.loads(request.POST.get('query', '{}'))
query['download'] = 'download' in request.POST
if collection:
try:
response = SolrApi(SOLR_URL.get(),
request.user).query(collection, query)
response = augment_solr_response(response, collection, query)
except RestException, e:
try:
response['error'] = json.loads(e.message)['error']['msg']
except:
LOG.exception('failed to parse json response')
response['error'] = force_unicode(e)
except Exception, e:
raise PopupException(e, title=_('Error while accessing Solr'))
response['error'] = force_unicode(e)
def index_fields_dynamic(request):
result = {'status': -1, 'message': 'Error'}
try:
name = request.POST['name']
dynamic_fields = SolrApi(SOLR_URL.get(), request.user).luke(name)
result['message'] = ''
result['fields'] = [
Collection2._make_field(name, properties)
for name, properties in dynamic_fields['fields'].iteritems()
if 'dynamicBase' in properties
]
result['gridlayout_header_fields'] = [
Collection2._make_gridlayout_header_field({'name': name}, True)
for name, properties in dynamic_fields['fields'].iteritems()
if 'dynamicBase' in properties
]
result['status'] = 0
except Exception, e:
result['message'] = force_unicode(e)
def get_range_facet(request):
result = {'status': -1, 'message': ''}
try:
collection = json.loads(request.POST.get('collection', '{}')) # Perms
facet = json.loads(request.POST.get('facet', '{}'))
action = request.POST.get('action', 'select')
solr_api = SolrApi(SOLR_URL.get(), request.user)
if action == 'select':
properties = _guess_gap(solr_api, collection, facet,
facet['properties']['start'],
facet['properties']['end'])
else:
properties = _zoom_range_facet(solr_api, collection, facet)
result['properties'] = properties
result['status'] = 0
except Exception, e:
result['message'] = unicode(str(e), "utf8")
def update_collection(self, name, fields):
"""
Only create new fields
"""
api = SolrApi(SOLR_URL.get(), self.user, SECURITY_ENABLED.get())
# Create only new fields
# Fields that already exist, do not overwrite since there is no way to do that, currently.
old_field_names = list(api.fields(name)['schema']['fields'].keys())
new_fields = [
field for field in fields if field['name'] not in old_field_names
]
new_fields_filtered = []
for field in new_fields:
new_field = {}
for attribute in [
attribute for attribute in ALLOWED_FIELD_ATTRIBUTES
if attribute in field
]:
new_field[attribute] = field[attribute]
new_fields_filtered.append(new_field)
api.add_fields(name, new_fields_filtered)
def _create_solr_cloud_collection(self, name, fields, unique_key_field,
df):
client = SolrClient(self.user)
with ZookeeperClient(hosts=client.get_zookeeper_host(),
read_only=False) as zc:
root_node = '%s/%s' % (ZK_SOLR_CONFIG_NAMESPACE, name)
tmp_path, solr_config_path = copy_configs(
fields=fields,
unique_key_field=unique_key_field,
df=df,
solr_cloud_mode=client.is_solr_cloud_mode(),
is_solr_six_or_more=client.is_solr_six_or_more(),
is_solr_hdfs_mode=client.is_solr_with_hdfs())
try:
config_root_path = '%s/%s' % (solr_config_path, 'conf')
try:
zc.copy_path(root_node, config_root_path)
except Exception, e:
zc.delete_path(root_node)
raise PopupException(
_('Error in copying Solr configurations: %s') % e)
finally:
# Don't want directories laying around
shutil.rmtree(tmp_path)
api = SolrApi(SOLR_URL.get(), self.user, SECURITY_ENABLED.get())
if not api.create_collection(name):
# Delete instance directory if we couldn't create a collection.
try:
zc.delete_path(root_node)
except Exception, e:
raise PopupException(
_('Error in deleting Solr configurations.'), detail=e)
raise PopupException(
_('Could not create collection. Check error logs for more info.'
))
def get_range_facet(request):
result = {'status': -1, 'message': ''}
try:
collection = json.loads(request.POST.get('collection', '{}'))
facet = json.loads(request.POST.get('facet', '{}'))
action = request.POST.get('action', 'select')
solr_api = SolrApi(SOLR_URL.get(), request.user)
if action == 'select':
properties = _guess_gap(solr_api, collection, facet, facet['properties']['start'], facet['properties']['end'])
else:
properties = _zoom_range_facet(solr_api, collection, facet) # Zoom out
result['properties'] = properties
result['status'] = 0
except Exception as e:
result['message'] = force_unicode(e)
return JsonResponse(result)
def get_stats(request):
result = {'status': -1, 'message': 'Error'}
try:
collection = json.loads(request.POST.get('collection', '{}'))
query = json.loads(request.POST.get('query', '{}'))
analysis = json.loads(request.POST.get('analysis', '{}'))
field = analysis['name']
facet = analysis['stats']['facet']
result['stats'] = SolrApi(SOLR_URL.get(),
request.user).stats(collection['name'],
[field], query, facet)
result['status'] = 0
result['message'] = ''
except Exception, e:
result['message'] = unicode(str(e), "utf8")
if 'not currently supported' in result['message']:
result['status'] = 1
result['message'] = _('This field does not support stats')
def admin_collection_properties(request, collection_id):
hue_collection = Collection.objects.get(id=collection_id)
solr_collection = SolrApi(SOLR_URL.get(), request.user).collection_or_core(hue_collection)
if request.method == 'POST':
collection_form = CollectionForm(request.POST, instance=hue_collection)
if collection_form.is_valid():
searcher = SearchController(request.user)
hue_collection = collection_form.save(commit=False)
hue_collection.is_core_only = not searcher.is_collection(hue_collection.name)
hue_collection.save()
return redirect(reverse('search:admin_collection_properties', kwargs={'collection_id': hue_collection.id}))
else:
request.error(_('Errors on the form: %s.') % collection_form.errors)
else:
collection_form = CollectionForm(instance=hue_collection)
return render('admin_collection_properties.mako', request, {
'solr_collection': solr_collection,
'hue_collection': hue_collection,
'collection_form': collection_form,
})
def delete_collection(self, name, core):
"""
Delete solr collection/core and instance dir
"""
api = SolrApi(SOLR_URL.get(), self.user, SECURITY_ENABLED.get())
client = SolrClient(self.user)
if core:
raise PopupException(_('Cannot remove Solr cores.'))
if api.remove_collection(name):
# Delete instance directory.
try:
root_node = '%s/%s' % (ZK_SOLR_CONFIG_NAMESPACE, name)
with ZookeeperClient(hosts=client.get_zookeeper_host(),
read_only=False) as zc:
zc.delete_path(root_node)
except Exception, e:
# Re-create collection so that we don't have an orphan config
api.add_collection(name)
raise PopupException(
_('Error in deleting Solr configurations.'), detail=e)
def get_terms(request):
result = {'status': -1, 'message': 'Error'}
try:
collection = json.loads(request.POST.get('collection', '{}'))
analysis = json.loads(request.POST.get('analysis', '{}'))
limit = json.loads(request.POST.get('limit', '25'))
support_distributed = [
engine for engine in get_engines(request.user)
if engine['type'] == 'solr'
][0]['analytics']
field = analysis['name']
properties = {
'terms.limit': limit,
'terms.distrib': str(support_distributed).lower(),
# lower
# mincount
# maxcount
}
if analysis['terms']['prefix']:
properties['terms.regex'] = '.*%(prefix)s.*' % analysis[
'terms'] # Use regexp instead of case sensitive 'terms.prefix'
properties['terms.regex.flag'] = 'case_insensitive'
result['terms'] = SolrApi(SOLR_URL.get(),
request.user).terms(collection['name'],
field, properties)
result['terms'] = pairwise2(field, [], result['terms']['terms'][field])
result['status'] = 0
result['message'] = ''
except Exception, e:
result['message'] = force_unicode(e)
if 'not currently supported' in result['message']:
result['status'] = 1
result['message'] = _('This field does not support stats')
def get_service_info(service):
service_info = {}
if service.lower() == 'solr':
service_info['url'] = SOLR_URL.get()
service_info['security_enabled'] = SOLR_SECURITY_ENABLED.get()
if service.lower() == 'oozie':
service_info['url'] = OOZIE_URL.get()
service_info['security_enabled'] = OOZIE_SECURITY_ENABLED.get()
if service.lower() == 'httpfs':
hdfs_config = hdfs_conf.HDFS_CLUSTERS['default']
service_info['url'] = hdfs_config.WEBHDFS_URL.get()
service_info['security_enabled'] = hdfs_config.SECURITY_ENABLED.get()
if service.lower() == 'rm':
yarn_cluster = cluster.get_cluster_conf_for_job_submission()
service_info['url'] = yarn_cluster.RESOURCE_MANAGER_API_URL.get()
service_info['security_enabled'] = yarn_cluster.SECURITY_ENABLED.get()
if service.lower() == 'jhs':
yarn_cluster = cluster.get_cluster_conf_for_job_submission()
service_info['url'] = yarn_cluster.HISTORY_SERVER_API_URL.get()
service_info['security_enabled'] = yarn_cluster.SECURITY_ENABLED.get()
if service.lower() == 'sparkhs':
yarn_cluster = cluster.get_cluster_conf_for_job_submission()
service_info['url'] = yarn_cluster.SPARK_HISTORY_SERVER_URL.get()
service_info[
'security_enabled'] = yarn_cluster.SPARK_HISTORY_SERVER_SECURITY_ENABLED.get(
)
if 'url' not in service_info:
logging.info("Hue does not have %s configured, cannot test %s" %
(service, service))
elif service_info['url'] is None:
logging.info("Hue does not have %s configured, cannot test %s" %
(service, service))
if service_info['url'].endswith('/'):
service_info['url'] = service_info['url'][:-1]
return service_info
def download(request, format):
hue_collections = SearchController(request.user).get_search_collections()
if not hue_collections:
raise PopupException(_("No collection to download."))
init_collection = initial_collection(request, hue_collections)
search_form = QueryForm(request.GET, initial_collection=init_collection)
if search_form.is_valid():
try:
collection_id = search_form.cleaned_data['collection']
hue_collection = Collection.objects.get(id=collection_id)
solr_query = search_form.solr_query_dict
response = SolrApi(SOLR_URL.get(), request.user).query(solr_query, hue_collection)
LOG.debug('Download results for query %s' % smart_str(solr_query))
return export_download(response, format)
except Exception, e:
raise PopupException(_("Could not download search results: %s") % e)
def delete_collection(self, name, core):
"""
Delete solr collection/core and instance dir
"""
api = SolrApi(SOLR_URL.get(), self.user, SECURITY_ENABLED.get())
if core:
raise PopupException(_('Cannot remove Solr cores.'))
if api.remove_collection(name):
# Delete instance directory.
solrctl_path = get_solrctl_path()
process = subprocess.Popen([solrctl_path, "instancedir", "--delete", name],
stdout=subprocess.PIPE,
stderr=subprocess.PIPE,
env={
'SOLR_ZK_ENSEMBLE': conf.SOLR_ZK_ENSEMBLE.get()
})
if process.wait() != 0:
LOG.error("Cloud not delete instance directory.\nOutput stream: %s\nError stream: %s" % process.communicate())
raise PopupException(_('Could not create instance directory. Check error logs for more info.'))
else:
raise PopupException(_('Could not remove collection. Check error logs for more info.'))
def get_document(request):
result = {'status': -1, 'message': 'Error'}
try:
collection = json.loads(request.POST.get('collection', '{}'))
doc_id = request.POST.get('id')
if doc_id:
result['doc'] = SolrApi(SOLR_URL.get(),
request.user).get(collection['name'],
doc_id)
if result['doc']['doc']:
result['status'] = 0
result['message'] = ''
else:
result['status'] = 1
result['message'] = _('No document was returned by Solr.')
else:
result['message'] = _('This document does not have any index id.')
result['status'] = 1
except Exception, e:
result['message'] = unicode(str(e), "utf8")
def search(request):
response = {}
collection = json.loads(request.POST.get('collection', '{}'))
query = json.loads(request.POST.get('query', '{}'))
query['download'] = 'download' in request.POST
if collection:
try:
response = SolrApi(SOLR_URL.get(),
request.user).query(collection, query)
response = augment_solr_response(response, collection, query)
except RestException, e:
try:
message = json.loads(e.message)
response['error'] = message['error'].get(
'msg', message['error']['trace'])
except Exception, e2:
LOG.exception('failed to extract json message: %s' %
force_unicode(e2))
LOG.exception('failed to parse json response: %s' %
force_unicode(e))
response['error'] = force_unicode(e)
def __init__(self, solr_url=None, user=None, security_enabled=False, ssl_cert_ca_verify=SSL_CERT_CA_VERIFY.get()):
if solr_url is None and hasattr(SOLR_URL, 'get'):
solr_url = SOLR_URL.get()
if solr_url:
self._url = solr_url
self._user = user
self._client = HttpClient(self._url, logger=LOG)
self.security_enabled = security_enabled or SECURITY_ENABLED.get()
if self.security_enabled:
self._client.set_kerberos_auth()
self._client.set_verify(ssl_cert_ca_verify)
self._root = resource.Resource(self._client)
# The Kerberos handshake requires two requests in order to authenticate,
# but if our first request is a PUT/POST, it might flat-out reject the
# first request if the body is too large. So, connect here in order to get
# a cookie so future PUT/POSTs will be pre-authenticated.
if self.security_enabled:
self._root.invoke('HEAD', '/')
def query_suggest(request):
if request.method != 'POST':
raise PopupException(_('POST request required.'))
collection = json.loads(request.POST.get('collection', '{}'))
query = request.POST.get('query', '')
result = {'status': -1, 'message': ''}
solr_query = {}
solr_query['q'] = query
solr_query['dictionary'] = collection['suggest']['dictionary']
try:
response = SolrApi(SOLR_URL.get(),
request.user).suggest(collection['name'],
solr_query)
result['response'] = response
result['status'] = 0
except Exception as e:
result['message'] = force_unicode(e)
return JsonResponse(result)
def _create_facet(collection, user, facet_id, facet_label, facet_field,
widget_type):
properties = {
'sort': 'desc',
'canRange': False,
'stacked': False,
'limit': 10,
'mincount': 0,
'isDate': False,
'andUp': False, # Not used yet
}
solr_api = SolrApi(SOLR_URL.get(), user)
range_properties = _new_range_facet(solr_api, collection, facet_field,
widget_type)
if range_properties:
facet_type = 'range'
properties.update(range_properties)
elif widget_type == 'hit-widget':
facet_type = 'query'
else:
facet_type = 'field'
if widget_type == 'map-widget':
properties['scope'] = 'world'
properties['mincount'] = 1
properties['limit'] = 100
return {
'id': facet_id,
'label': facet_label,
'field': facet_field,
'type': facet_type,
'widgetType': widget_type,
'properties': properties
}
def create_collection(self,
name,
fields,
unique_key_field='id',
df='text'):
"""
Create solr collection or core and instance dir.
Create schema.xml file so that we can set UniqueKey field.
"""
if self.is_solr_cloud_mode():
# solrcloud mode
# Need to remove path afterwards
tmp_path, solr_config_path = copy_configs(fields, unique_key_field,
df, True)
zc = ZookeeperClient(hosts=get_solr_ensemble(), read_only=False)
root_node = '%s/%s' % (ZK_SOLR_CONFIG_NAMESPACE, name)
config_root_path = '%s/%s' % (solr_config_path, 'conf')
try:
zc.copy_path(root_node, config_root_path)
except Exception, e:
zc.delete_path(root_node)
raise PopupException(
_('Error in copying Solr configurations.'), detail=e)
# Don't want directories laying around
shutil.rmtree(tmp_path)
api = SolrApi(SOLR_URL.get(), self.user, SECURITY_ENABLED.get())
if not api.create_collection(name):
# Delete instance directory if we couldn't create a collection.
try:
zc.delete_path(root_node)
except Exception, e:
raise PopupException(
_('Error in deleting Solr configurations.'), detail=e)
def search(request):
response = {}
collection = json.loads(request.POST.get('collection', '{}'))
query = json.loads(request.POST.get('query', '{}'))
query['download'] = 'download' in request.POST
# todo: remove the selected histo facet if multiq
if collection['id']:
hue_collection = Collection.objects.get(id=collection['id']) # TODO perms
if collection:
try:
response = SolrApi(SOLR_URL.get(), request.user).query(collection, query)
response = augment_solr_response(response, collection, query)
except RestException, e:
try:
response['error'] = json.loads(e.message)['error']['msg']
except:
response['error'] = force_unicode(str(e))
except Exception, e:
raise PopupException(e, title=_('Error while accessing Solr'))
response['error'] = force_unicode(str(e))
def _fetch_collections(request):
from search.conf import SOLR_URL
path = request.GET['path']
item = None
name = None
if path:
item = path
if '/' in path:
item, name = path.split('/')
api = SolrApi(SOLR_URL.get(), request.user)
if not item:
return {"databases": ["collections", "configs"]}
elif item and name:
return {"authorizable_link": "/indexer/#edit/%s" % name, "extended_columns": [], "columns": [], "partition_keys": []}
elif item == 'collections':
return {"tables_meta": [{"comment": None, "type": "Table", "name": col} for col in api.collections2()]}
elif item == 'configs':
return {"tables_meta": [{"comment": None, "type": "Table", "name": conf} for conf in api.configs()]}
else:
raise PopupException(_('Authorizable %s could not be retrieved') % path)
def execute(self, notebook, snippet):
from search.conf import SOLR_URL
api = NativeSolrApi(SOLR_URL.get(), self.user.username)
collection = self.options.get('collection') or snippet.get('database')
if collection == 'default':
collection = api.collections2()[0]
response = api.sql(collection, snippet['statement'])
info = response['result-set']['docs'].pop(-1) # EOF, RESPONSE_TIME, EXCEPTION
if info.get('EXCEPTION'):
raise QueryError(info['EXCEPTION'])
data = [[cell for cell in doc.values()] for doc in response['result-set']['docs']]
has_result_set = bool(data)
return {
'sync': True,
'has_result_set': has_result_set,
'modified_row_count': 0,
'result': {
'has_more': False,
'data': data if has_result_set else [],
'meta': [{
'name': col,
'type': '',
'comment': ''
} for col in response['result-set']['docs'][0].keys()] if has_result_set else [],
'type': 'table'
},
'statement_id': 0,
'has_more_statements': False,
'statements_count': 1
}
def _create_non_solr_cloud_collection(self, name, fields, unique_key_field,
df):
# Non-solrcloud mode
# Create instance directory locally.
instancedir = os.path.join(CORE_INSTANCE_DIR.get(), name)
if os.path.exists(instancedir):
raise PopupException(
_("Instance directory %s already exists! Please remove it from the file system."
) % instancedir)
tmp_path, solr_config_path = copy_configs(fields, unique_key_field, df,
False)
try:
shutil.move(solr_config_path, instancedir)
finally:
shutil.rmtree(tmp_path)
api = SolrApi(SOLR_URL.get(), self.user, SECURITY_ENABLED.get())
if not api.create_core(name, instancedir):
# Delete instance directory if we couldn't create a collection.
shutil.rmtree(instancedir)
raise PopupException(
_('Could not create collection. Check error logs for more info.'
))
response['columns'] = [col['name'] for col in columns]
response['extended_columns'] = columns
response['status'] = 0
return response
@query_error_handler
def get_sample_data(self,
snippet,
database=None,
table=None,
column=None,
async=False,
operation=None):
from search.conf import SOLR_URL
db = NativeSolrApi(SOLR_URL.get(), self.user)
assist = Assist(self, self.user, db)
response = {'status': -1}
if snippet.get('source') == 'sql':
sample_data = assist.get_sample_data_sql(database, table, column)
else:
sample_data = assist.get_sample_data(database, table, column)
if sample_data:
response['status'] = 0
response['headers'] = sample_data['headers']
response['full_headers'] = sample_data.get('full_headers')
response['rows'] = sample_data['rows']
else:
def admin_collection_schema(request, collection_id):
hue_collection = Collection.objects.get(id=collection_id)
solr_schema = SolrApi(SOLR_URL.get()).schema(hue_collection.name)
content = {'solr_schema': solr_schema.decode('utf-8')}
return HttpResponse(json.dumps(content), mimetype="application/json")
def is_core(self, core_name):
solr_cores = SolrApi(SOLR_URL.get()).cores()
return core_name in solr_cores
def is_collection(self, collection_name):
solr_collections = SolrApi(SOLR_URL.get()).collections()
return collection_name in solr_collections
class SolrApi(Api):
def __init__(self, user, interpreter=None):
Api.__init__(self, user, interpreter=interpreter)
self.options = interpreter['options']
@query_error_handler
def execute(self, notebook, snippet):
from search.conf import SOLR_URL
api = NativeSolrApi(SOLR_URL.get(), self.user.username)
collection = self.options.get('collection') or snippet.get('database')
if not collection or collection == 'default':
collection = api.collections2()[0]
response = api.sql(collection, snippet['statement'])
info = response['result-set']['docs'].pop(
-1) # EOF, RESPONSE_TIME, EXCEPTION
if info.get('EXCEPTION'):
raise QueryError(info['EXCEPTION'])
headers = []
for row in response['result-set']['docs']:
for col in row.keys():
if col not in headers:
headers.append(col)
data = [[doc.get(col) for col in headers]
for doc in response['result-set']['docs']]
has_result_set = bool(data)
return {
'sync': True,
'has_result_set': has_result_set,
'modified_row_count': 0,
'result': {
'has_more':
False,
'data':
data if has_result_set else [],
'meta': [{
'name': col,
'type': '',
'comment': ''
} for col in headers] if has_result_set else [],
'type':
'table'
},
'statement_id': 0,
'has_more_statements': False,
'statements_count': 1
}
@query_error_handler
def check_status(self, notebook, snippet):
return {'status': 'available'}
@query_error_handler
def fetch_result(self, notebook, snippet, rows, start_over):
return {'has_more': False, 'data': [], 'meta': [], 'type': 'table'}
@query_error_handler
def fetch_result_metadata(self):
pass
@query_error_handler
def cancel(self, notebook, snippet):
return {'status': 0}
@query_error_handler
def get_log(self, notebook, snippet, startFrom=None, size=None):
return 'No logs'
def download(self, notebook, snippet, format):
raise PopupException('Downloading is not supported yet')
@query_error_handler
def close_statement(self, snippet):
return {'status': -1}
@query_error_handler
def autocomplete(self,
snippet,
database=None,
table=None,
column=None,
nested=None):
from search.conf import SOLR_URL
api = NativeSolrApi(SOLR_URL.get(), self.user.username)
assist = Assist(self, self.user, api)
response = {'status': -1}
if database is None:
response['databases'] = [
self.options.get('collection') or snippet.get('database')
or 'default'
]
elif table is None:
response['tables_meta'] = assist.get_tables(database)
else:
columns = assist.get_columns(database, table)
response['columns'] = [col['name'] for col in columns]
response['extended_columns'] = columns
response['status'] = 0
return response
@query_error_handler
def get_sample_data(self,
snippet,
database=None,
table=None,
column=None,
async=False):
from search.conf import SOLR_URL
db = NativeSolrApi(SOLR_URL.get(), self.user)
assist = Assist(self, self.user, db)
response = {'status': -1}
if snippet.get('source') == 'sql':
sample_data = assist.get_sample_data_sql(database, table, column)
else:
sample_data = assist.get_sample_data(database, table, column)
if sample_data:
response['status'] = 0
response['headers'] = sample_data['headers']
response['full_headers'] = sample_data.get('full_headers')
response['rows'] = sample_data['rows']
else:
response['message'] = _('Failed to get sample data.')
return response
def get_solr_collection(self): return SolrApi(SOLR_URL.get(), self.user).collections()
