def setUp(self):
"""
set up DAS core module
"""
debug = 0
self.db = 'test_analytics.db'
config = deepcopy(das_readconfig())
dburi = config['mongodb']['dburi']
logger = PrintManager('TestDASAnalytics', verbose=debug)
config['logger'] = logger
config['verbose'] = debug
config['analyticsdb'] = dict(dburi=dburi, history=5184000,
dbname='test_analytics', collname='db')
self.mgr = DASAnalytics(config)
def setUp(self):
"""
set up data used in the tests.
setUp is called before each test function execution.
"""
self.i1 = "find dataset, run, bfield where site = T2 and admin=VK and storage=castor"
self.i2 = " find dataset, run where (run=1 or run=2) and storage=castor or site = T2"
debug = 0
config = das_readconfig()
logger = DASLogger(verbose=debug, stdout=debug)
config['logger'] = logger
config['verbose'] = debug
config['mapping_dbhost'] = 'localhost'
config['mapping_dbport'] = 27017
config['mapping_dbname'] = 'mapping'
config['dasmapping'] = DASMapping(config)
config['dasanalytics'] = DASAnalytics(config)
self.parser = MongoParser(config)
self.operators = [o.strip() for o in DAS_OPERATORS]
def __init__(self, config=None):
if not config:
config = das_readconfig()
self.dasmapping = DASMapping(config)
if not self.dasmapping.check_maps():
msg = "No DAS maps found in MappingDB"
raise Exception(msg)
self.analytics = DASAnalytics(config)
self.dasservices = config['services']
self.daskeysmap = self.dasmapping.daskeys()
self.operators = list(das_operators())
self.daskeys = list(das_special_keys())
self.verbose = config['verbose']
self.logger = PrintManager('QLManger', self.verbose)
for val in self.daskeysmap.values():
for item in val:
self.daskeys.append(item)
parserdir = config['das']['parserdir']
self.parserdir = parserdir
self.enabledb = config['parserdb']['enable']
if self.enabledb:
self.parserdb = DASParserDB(config)
def __init__(self, config=None, debug=0,
nores=False, logger=None, engine=None, multitask=True):
if config:
dasconfig = config
else:
dasconfig = das_readconfig()
verbose = dasconfig['verbose']
self.stdout = debug
if isinstance(debug, int):
self.verbose = debug
dasconfig['verbose'] = debug
else:
self.verbose = verbose
das_timer('DASCore::init', self.verbose)
self.operators = das_operators()
# set noresults option
self.noresults = False
if nores:
dasconfig['write_cache'] = True
self.noresults = nores
self.multitask = dasconfig['das'].get('multitask', True)
if debug or self.verbose:
self.multitask = False # in verbose mode do not use multitask
dasconfig['das']['multitask'] = False
if not multitask: # explicitly call DASCore ctor, e.g. in analytics
self.multitask = False
dasconfig['das']['multitask'] = False
dasconfig['engine'] = engine
if self.multitask:
nworkers = dasconfig['das'].get('core_workers', 5)
if engine:
thr_name = 'DASCore:PluginTaskManager'
self.taskmgr = PluginTaskManager(\
engine, nworkers=nworkers, name=thr_name)
self.taskmgr.subscribe()
else:
thr_name = 'DASCore:TaskManager'
self.taskmgr = TaskManager(nworkers=nworkers, name=thr_name)
else:
self.taskmgr = None
if logger:
self.logger = logger
else:
self.logger = PrintManager('DASCore', self.verbose)
# define Mapping/Analytics/Parser in this order since Parser depends
# on first two
dasmapping = DASMapping(dasconfig)
dasconfig['dasmapping'] = dasmapping
self.mapping = dasmapping
self.analytics = DASAnalytics(dasconfig)
dasconfig['dasanalytics'] = self.analytics
self.keylearning = DASKeyLearning(dasconfig)
dasconfig['keylearning'] = self.keylearning
# init DAS cache
self.rawcache = DASMongocache(dasconfig)
dasconfig['rawcache'] = self.rawcache
# plug-in architecture: loop over registered data-services in
# dasconfig; load appropriate module/class; register data
# service with DASCore.
self.systems = dasmapping.list_systems()
# pointer to the DAS top level directory
dasroot = '/'.join(__file__.split('/')[:-3])
for name in self.systems:
try:
klass = 'DAS/services/%s/%s_service.py' \
% (name, name)
srvfile = os.path.join(dasroot, klass)
with file(srvfile) as srvclass:
for line in srvclass:
if line.find('(DASAbstractService)') != -1:
klass = line.split('(DASAbstractService)')[0]
klass = klass.split('class ')[-1]
break
mname = 'DAS.services.%s.%s_service' % (name, name)
module = __import__(mname, fromlist=[klass])
obj = getattr(module, klass)(dasconfig)
setattr(self, name, obj)
SERVICES[name] = obj
except IOError as err:
if debug > 1:
# we have virtual services, so IOError can be correct
print_exc(err)
try:
mname = 'DAS.services.generic_service'
module = __import__(mname, fromlist=['GenericService'])
obj = module.GenericService(name, dasconfig)
setattr(self, name, obj)
except Exception as exc:
print_exc(exc)
msg = "Unable to load %s data-service plugin" % name
raise Exception(msg)
except Exception as exc:
print_exc(exc)
msg = "Unable to load %s data-service plugin" % name
raise Exception(msg)
# loop over systems and get system keys, add mapping keys to final list
self.service_keys = {}
self.service_parameters = {}
for name in self.systems:
skeys = getattr(self, name).keys()
self.service_keys[getattr(self, name).name] = skeys
sparams = getattr(self, name).parameters()
self.service_parameters[getattr(self, name).name] = sparams
self.service_keys['special'] = das_special_keys()
self.dasconfig = dasconfig
das_timer('DASCore::init', self.verbose)
class DASCore(object):
"""
DAS core class.
"""
def __init__(self, config=None, debug=0,
nores=False, logger=None, engine=None, multitask=True):
if config:
dasconfig = config
else:
dasconfig = das_readconfig()
verbose = dasconfig['verbose']
self.stdout = debug
if isinstance(debug, int):
self.verbose = debug
dasconfig['verbose'] = debug
else:
self.verbose = verbose
das_timer('DASCore::init', self.verbose)
self.operators = das_operators()
# set noresults option
self.noresults = False
if nores:
dasconfig['write_cache'] = True
self.noresults = nores
self.multitask = dasconfig['das'].get('multitask', True)
if debug or self.verbose:
self.multitask = False # in verbose mode do not use multitask
dasconfig['das']['multitask'] = False
if not multitask: # explicitly call DASCore ctor, e.g. in analytics
self.multitask = False
dasconfig['das']['multitask'] = False
dasconfig['engine'] = engine
if self.multitask:
nworkers = dasconfig['das'].get('core_workers', 5)
if engine:
thr_name = 'DASCore:PluginTaskManager'
self.taskmgr = PluginTaskManager(\
engine, nworkers=nworkers, name=thr_name)
self.taskmgr.subscribe()
else:
thr_name = 'DASCore:TaskManager'
self.taskmgr = TaskManager(nworkers=nworkers, name=thr_name)
else:
self.taskmgr = None
if logger:
self.logger = logger
else:
self.logger = PrintManager('DASCore', self.verbose)
# define Mapping/Analytics/Parser in this order since Parser depends
# on first two
dasmapping = DASMapping(dasconfig)
dasconfig['dasmapping'] = dasmapping
self.mapping = dasmapping
self.analytics = DASAnalytics(dasconfig)
dasconfig['dasanalytics'] = self.analytics
self.keylearning = DASKeyLearning(dasconfig)
dasconfig['keylearning'] = self.keylearning
# init DAS cache
self.rawcache = DASMongocache(dasconfig)
dasconfig['rawcache'] = self.rawcache
# plug-in architecture: loop over registered data-services in
# dasconfig; load appropriate module/class; register data
# service with DASCore.
self.systems = dasmapping.list_systems()
# pointer to the DAS top level directory
dasroot = '/'.join(__file__.split('/')[:-3])
for name in self.systems:
try:
klass = 'DAS/services/%s/%s_service.py' \
% (name, name)
srvfile = os.path.join(dasroot, klass)
with file(srvfile) as srvclass:
for line in srvclass:
if line.find('(DASAbstractService)') != -1:
klass = line.split('(DASAbstractService)')[0]
klass = klass.split('class ')[-1]
break
mname = 'DAS.services.%s.%s_service' % (name, name)
module = __import__(mname, fromlist=[klass])
obj = getattr(module, klass)(dasconfig)
setattr(self, name, obj)
SERVICES[name] = obj
except IOError as err:
if debug > 1:
# we have virtual services, so IOError can be correct
print_exc(err)
try:
mname = 'DAS.services.generic_service'
module = __import__(mname, fromlist=['GenericService'])
obj = module.GenericService(name, dasconfig)
setattr(self, name, obj)
except Exception as exc:
print_exc(exc)
msg = "Unable to load %s data-service plugin" % name
raise Exception(msg)
except Exception as exc:
print_exc(exc)
msg = "Unable to load %s data-service plugin" % name
raise Exception(msg)
# loop over systems and get system keys, add mapping keys to final list
self.service_keys = {}
self.service_parameters = {}
for name in self.systems:
skeys = getattr(self, name).keys()
self.service_keys[getattr(self, name).name] = skeys
sparams = getattr(self, name).parameters()
self.service_parameters[getattr(self, name).name] = sparams
self.service_keys['special'] = das_special_keys()
self.dasconfig = dasconfig
das_timer('DASCore::init', self.verbose)
def keys(self):
"""
Return map of data service keys
"""
return self.service_keys
def das_keys(self):
"""
Return map of data service keys
"""
_keys = ['records']
for values in self.service_keys.values():
for key in values:
if key not in _keys:
_keys.append(key)
return _keys
def result(self, query, idx=0, limit=None):
"""
Get results either from cache or from explicit call
"""
self.logger.info('input query=%s' % query)
results = []
dasquery = DASQuery(query)
dasquery.add_to_analytics()
query = dasquery.mongo_query
# check if we have any service which cover the query
# otherwise decompose it into list of queries
service_map = dasquery.service_apis_map()
if not service_map:
msg = 'no APIs found to answer input query, will decompose it'
self.logger.info(msg)
skeys = query['fields']
if not skeys:
skeys = []
for key in skeys:
newquery = DASQuery(dict(fields=[key], spec=query['spec']))
self.call(newquery) # process query
else:
self.call(dasquery) # process query
# lookup provided query in a cache
if not self.noresults:
results = self.get_from_cache(dasquery, idx, limit)
return results
def remove_from_cache(self, dasquery):
"""
Delete in cache entries about input query
"""
self.rawcache.remove_from_cache(dasquery)
def get_status(self, dasquery):
"""
Look-up status of provided query in a cache.
Return status of the query request and its hash.
"""
status = None
error = None
reason = None
for col in ['merge', 'cache']:
self.rawcache.remove_expired(dasquery, col)
if dasquery and 'fields' in dasquery.mongo_query:
fields = dasquery.mongo_query['fields']
if fields and isinstance(fields, list) and 'queries' in fields:
return 'ok', error, reason
record = self.rawcache.find(dasquery)
error, reason = self.rawcache.is_error_in_records(dasquery)
try:
if record and 'das' in record and 'status' in record['das']:
status = record['das']['status']
if not error:
error = record['das'].get('error', error)
if not reason:
reason = record['das'].get('reason', reason)
return status, error, reason
except Exception as exc:
print_exc(exc)
status = error = reason = None
self.rawcache.remove_from_cache(dasquery)
return status, error, reason
def worker(self, srv, dasquery):
"""Main worker function which calls data-srv call function"""
self.logger.info('##### %s ######\n' % srv)
das_timer(srv, self.verbose)
getattr(getattr(self, srv), 'call')(dasquery)
das_timer(srv, self.verbose)
def insert_query_records(self, dasquery):
"""
Insert DAS query records into DAS cache and return list of services
which will answer this query
"""
services = dasquery.services
self.logger.info('Potential services = %s' % services)
if not services:
msg = 'No data-services for query %s' % dasquery
msg += 'mongo_query: %s' % dasquery.mongo_query
msg += 'params: %s' % dasquery.params()
print dastimestamp('DAS WARNING '), msg
# get list of URI which can answer this query
ack_services = []
for srv in services:
gen = [t for t in getattr(getattr(self, srv), 'apimap')(dasquery)]
for url, api, args, iformat, expire in gen:
header = dasheader(srv, dasquery, expire, api, url, ctime=0)
self.rawcache.insert_query_record(dasquery, header)
if srv not in ack_services:
ack_services.append(srv)
if not ack_services:
ack_services = services
if dasquery.query.find('records ') != -1:
srv_status = True # skip DAS queries w/ records request
expire = 2*60 # 2 minutes, it should be overwriten by data-srv
header = dasheader("das", dasquery, expire, api='das_core',
services=dict(das=ack_services))
header['lookup_keys'] = []
self.rawcache.insert_query_record(dasquery, header)
das_timer('das_record', self.verbose)
return ack_services
def call(self, query, add_to_analytics=True, **kwds):
"""
Top level DAS api which execute a given query using underlying
data-services. It follows the following steps:
- parse input query
- identify data-sercices based on selection keys
and where clause conditions
- construct DAS workflow and execute data-service
API calls. At this step individual
data-services store results into DAS cache.
Return status 0/1 depending on success of the calls, can be
used by workers on cache server.
kwds is provided for compatibility with web layer, e.g. it
may invoke this method with additional pid parameter.
"""
def update_das_query(dasquery, status, reason=None):
"Update DAS query record with given status and reason"
self.rawcache.update_query_record(dasquery, status, reason=reason)
self.rawcache.add_to_record(\
dasquery, {'das.timer': get_das_timer()}, system='das')
# make sure that das record is updated, we use 7 iteration which
# sum up into 1 minute to cover default syncdelay value of mongo
# server (in a future it would be better to find programatically
# this syncdelay value, but it seems pymongo driver does not
# provide any API for it.
for idx in xrange(1, 7):
spec = {'qhash':dasquery.qhash, 'das.system':['das']}
res = self.rawcache.col.find_one(spec)
if res:
dbstatus = res.get('das', {}).get('status', None)
if dbstatus == status:
break
msg = 'qhash %s, das.status=%s, status=%s, wait for update' \
% (dasquery.qhash, dbstatus, status)
print dastimestamp('DAS WARNING'), msg
time.sleep(idx*idx)
self.rawcache.update_query_record(dasquery, status, reason=reason)
self.logger.info('input query=%s' % query)
das_timer('DASCore::call', self.verbose)
if isinstance(query, object) and hasattr(query, '__class__')\
and query.__class__.__name__ == 'DASQuery':
dasquery = query
else:
dasquery = DASQuery(query)
for col in ['merge', 'cache']:
self.rawcache.remove_expired(dasquery, col)
if add_to_analytics:
dasquery.add_to_analytics()
query = dasquery.mongo_query
spec = query.get('spec')
fields = query.get('fields')
if fields == ['records']:
msg = 'look-up all records in cache'
self.logger.info(msg)
return 'in cache'
if spec == dict(records='*'):
self.logger.info("look-up everything in cache")
return 'in cache'
for record in self.rawcache.find_specs(dasquery):
status = record['das']['status']
msg = 'found query %s in cache, status=%s\n' \
% (record['query'], status)
self.logger.info(msg)
print dastimestamp('DAS INFO'), msg
return status
self.logger.info(dasquery)
das_timer('das_record', self.verbose)
services = self.insert_query_records(dasquery)
if not services:
msg = 'unable to locate data-services to fulfill this request'
msg += ', will iterate over all registered services'
print dastimestamp('DAS WARNING '), dasquery, msg
services = dasquery.services if dasquery.services else self.systems
try:
if self.multitask:
jobs = []
for srv in sorted(services):
jobs.append(self.taskmgr.spawn(self.worker, srv, dasquery))
self.taskmgr.joinall(jobs)
else:
for srv in services:
self.worker(srv, dasquery)
except Exception as exc:
print_exc(exc)
return 'fail'
self.logger.info('\n##### merging ######\n')
update_das_query(dasquery, 'merging')
das_timer('merge', self.verbose)
self.rawcache.merge_records(dasquery)
das_timer('merge', self.verbose)
# check if we have service records and properly setup status
self.logger.info('\n##### check services ######\n')
das_services = self.rawcache.check_services(dasquery)
reason = ''
status = 'ok'
if not das_services:
if 'records' in dasquery.query:
status = 'ok' # keep status ok for 'records' queries
else:
reason = 'no data records found in DAS cache'
status = 'fail'
print dastimestamp('DAS ERROR '), dasquery, reason
update_das_query(dasquery, status, reason)
das_timer('DASCore::call', self.verbose)
return status
def processing_time(self, dasquery):
"Look-up and return DAS query processing time"
query_record = self.rawcache.find(dasquery)
if query_record:
das = query_record.get('das', None)
if isinstance(das, dict):
ctime = das.get('ctime', [])
if ctime:
return ctime[-1]-ctime[0]
return None
def nresults(self, dasquery, coll='merge'):
"""
Return total number of results (count) for provided query
Code should match body of get_from_cache method.
"""
fields = dasquery.mongo_query.get('fields', None)
if dasquery.mapreduce:
result = self.rawcache.map_reduce(dasquery.mapreduce, dasquery)
return len([1 for _ in result])
elif dasquery.aggregators:
return len(dasquery.aggregators)
elif isinstance(fields, list) and 'queries' in fields:
return len([1 for _ in self.get_queries(dasquery)])
return self.rawcache.nresults(dasquery, coll)
def apilist(self, dasquery):
"Return list of APIs answer given das query"
return self.rawcache.apilist(dasquery)
def incache(self, dasquery, coll='merge'):
"""
Answer the question if given query in DAS cache or not
"""
return self.rawcache.incache(dasquery, collection=coll)
def get_from_cache(self, dasquery, idx=0, limit=0, collection='merge'):
"""
Look-up results from the merge cache and yield them for
further processing.
"""
das_timer('DASCore::get_from_cache', self.verbose)
msg = 'col=%s, query=%s, idx=%s, limit=%s'\
% (collection, dasquery, idx, limit)
self.logger.info(msg)
fields = dasquery.mongo_query.get('fields', None)
if dasquery.mapreduce:
res = self.rawcache.map_reduce(dasquery.mapreduce, dasquery)
elif dasquery.aggregators:
# extract das information from rawcache
rows = self.rawcache.get_from_cache(\
dasquery, collection=collection)
first = rows.next()
sinfo = das_sinfo(first)
# to perform aggregation we need:
# - loop over all aggregator functions
# - loop over all data-services
# - loop over all APIs within a data-services
# the code below does that, it applies aggregator
# to selected (based on key/srv/api) records
res = []
_id = 0
time0 = time.time()
expire = 300 # min expire
for func, key in dasquery.aggregators:
afunc = getattr(das_aggregator, 'das_%s' % func)
found = False
for srv, apis, in sinfo.items():
for api in apis:
rows = self.rawcache.get_from_cache(\
dasquery, collection=collection)
gen = api_rows(rows, api)
data = afunc(key, gen)
ctime = time.time() - time0
das = dasheader(srv, dasquery, expire, api=api,
ctime=ctime)
if isinstance(data, dict) and data['value'] != 'N/A':
aggr = {'_id':_id, 'function': func,
'key': key, 'result': data}
aggr.update(das)
res.append(aggr)
_id += 1
found = True
if not found: # when we got nothing add empty result record
empty = {'value':'N/A'}
ctime = time.time() - time0
das = dasheader('das', dasquery, expire, api='das_core',
ctime=ctime)
rec = {'_id':0, 'function':func, 'key':key, 'result':empty}
rec.update(das)
res.append(rec)
elif isinstance(fields, list) and 'queries' in fields:
res = itertools.islice(self.get_queries(dasquery), idx, idx+limit)
else:
res = self.rawcache.get_from_cache(dasquery, idx, limit, \
collection=collection)
for row in res:
fix_times(row)
yield row
das_timer('DASCore::get_from_cache', self.verbose)
def get_queries(self, dasquery):
"""
Look-up (popular) queries in DAS analytics/logging db
"""
das_timer('DASCore::get_queries', self.verbose)
fields = dasquery.mongo_query.get('fields')
spec = dasquery.mongo_query.get('spec')
if 'popular' in fields:
res = self.analytics.get_popular_queries(spec)
else:
datestamp = spec.get('date')
if isinstance(datestamp, dict):
value = datestamp.get('$in')
res = \
self.analytics.list_queries(after=value[0], before=value[1])
elif isinstance(datestamp, int):
res = self.analytics.list_queries(after=datestamp)
elif not datestamp:
res = self.analytics.list_queries()
else:
msg = 'Unsupported date value: %s' % datestamp
raise Exception(msg)
for row in res:
rid = row.pop('_id')
yield dict(das_query=row, _id=rid)
das_timer('DASCore::get_queries', self.verbose)
class testDASAnalytics(unittest.TestCase):
"""
A test class for the DAS analyticsdb class
"""
def setUp(self):
"""
set up DAS core module
"""
debug = 0
self.db = 'test_analytics.db'
config = deepcopy(das_readconfig())
dburi = config['mongodb']['dburi']
logger = PrintManager('TestDASAnalytics', verbose=debug)
config['logger'] = logger
config['verbose'] = debug
config['analyticsdb'] = dict(dburi=dburi, history=5184000,
dbname='test_analytics', collname='db')
self.mgr = DASAnalytics(config)
def tearDown(self):
"""Invoke after each test"""
self.mgr.delete_db()
def test_api(self):
"""test methods for api table"""
self.mgr.delete_db()
self.mgr.create_db()
query = 'find block'
dbs_api = 'listBlocks'
dbs_params = {'apiversion':'DBS_2_0_8',
'block_name':'*', 'storage_element_name':'*',
'user_type':'NORMAL'}
self.mgr.add_api('dbs', query, dbs_api, dbs_params)
phedex_api = 'blockReplicas'
phedex_params = {'node': '*', 'se': '*', 'block': '*'}
self.mgr.add_api('phedex', query, phedex_api, phedex_params)
res = self.mgr.list_systems()
res.sort()
self.assertEqual(['dbs', 'phedex'], res)
res = self.mgr.list_apis('dbs')
self.assertEqual([dbs_api], res)
self.mgr.add_api('dbs', query, dbs_api, dbs_params)
res = self.mgr.api_counter(dbs_api)
self.assertEqual(1, res) # we invoke API twice, so should get 2
self.mgr.update('dbs', query)
res = self.mgr.api_counter(dbs_api)
self.assertEqual(2, res) # we invoke API twice, so should get 2
res = self.mgr.api_params(phedex_api)
self.assertEqual([phedex_params], res)
class DASCore(object):
"""
DAS core class.
"""
def __init__(self, config=None, debug=None,
nores=False, logger=None, engine=None, multitask=True):
if config:
dasconfig = config
else:
dasconfig = das_readconfig()
verbose = dasconfig['verbose']
self.stdout = debug
if isinstance(debug, int):
self.verbose = debug
dasconfig['verbose'] = debug
else:
self.verbose = verbose
das_timer('DASCore::init', self.verbose)
self.operators = das_operators()
# set noresults option
self.noresults = False
if nores:
dasconfig['write_cache'] = True
self.noresults = nores
self.multitask = dasconfig['das'].get('multitask', True)
if debug or self.verbose:
self.multitask = False # in verbose mode do not use multitask
dasconfig['das']['multitask'] = False
if not multitask: # explicitly call DASCore ctor, e.g. in analytics
self.multitask = False
dasconfig['das']['multitask'] = False
dasconfig['engine'] = engine
if self.multitask:
nworkers = dasconfig['das'].get('core_workers', 5)
if engine:
thr_name = 'DASCore:PluginTaskManager'
self.taskmgr = PluginTaskManager(\
engine, nworkers=nworkers, name=thr_name)
self.taskmgr.subscribe()
else:
thr_name = 'DASCore:TaskManager'
self.taskmgr = TaskManager(nworkers=nworkers, name=thr_name)
else:
self.taskmgr = None
if logger:
self.logger = logger
else:
self.logger = PrintManager('DASCore', self.verbose)
# define Mapping/Analytics/Parser in this order since Parser depends
# on first two
dasmapping = DASMapping(dasconfig)
dasconfig['dasmapping'] = dasmapping
self.mapping = dasmapping
self.analytics = DASAnalytics(dasconfig)
dasconfig['dasanalytics'] = self.analytics
self.mongoparser = ql_manager(dasconfig)
dasconfig['mongoparser'] = self.mongoparser
self.keylearning = DASKeyLearning(dasconfig)
dasconfig['keylearning'] = self.keylearning
# init DAS cache
self.rawcache = DASMongocache(dasconfig)
dasconfig['rawcache'] = self.rawcache
# plug-in architecture: loop over registered data-services in
# dasconfig; load appropriate module/class; register data
# service with DASCore.
self.systems = dasmapping.list_systems()
# pointer to the DAS top level directory
dasroot = '/'.join(__file__.split('/')[:-3])
for name in self.systems:
try:
klass = 'DAS/services/%s/%s_service.py' \
% (name, name)
srvfile = os.path.join(dasroot, klass)
with file(srvfile) as srvclass:
for line in srvclass:
if line.find('(DASAbstractService)') != -1:
klass = line.split('(DASAbstractService)')[0]
klass = klass.split('class ')[-1]
break
mname = 'DAS.services.%s.%s_service' % (name, name)
module = __import__(mname, fromlist=[klass])
obj = getattr(module, klass)(dasconfig)
setattr(self, name, obj)
SERVICES[name] = obj
except IOError as err:
if debug > 1:
# we have virtual services, so IOError can be correct
print_exc(err)
try:
mname = 'DAS.services.generic_service'
module = __import__(mname, fromlist=['GenericService'])
obj = module.GenericService(name, dasconfig)
setattr(self, name, obj)
except Exception as exc:
print_exc(exc)
msg = "Unable to load %s data-service plugin" % name
raise Exception(msg)
except Exception as exc:
print_exc(exc)
msg = "Unable to load %s data-service plugin" % name
raise Exception(msg)
# loop over systems and get system keys, add mapping keys to final list
self.service_keys = {}
self.service_parameters = {}
for name in self.systems:
skeys = getattr(self, name).keys()
self.service_keys[getattr(self, name).name] = skeys
sparams = getattr(self, name).parameters()
self.service_parameters[getattr(self, name).name] = sparams
self.service_keys['special'] = das_special_keys()
self.dasconfig = dasconfig
das_timer('DASCore::init', self.verbose)
def keys(self):
"""
Return map of data service keys
"""
return self.service_keys
def das_keys(self):
"""
Return map of data service keys
"""
_keys = ['records']
for values in self.service_keys.values():
for key in values:
if key not in _keys:
_keys.append(key)
return _keys
def result(self, query, idx=0, limit=None):
"""
Get results either from cache or from explicit call
"""
self.logger.info('input query=%s' % query)
results = []
dasquery = DASQuery(query, mongoparser=self.mongoparser)
dasquery.add_to_analytics()
query = dasquery.mongo_query
# check if we have any service which cover the query
# otherwise decompose it into list of queries
service_map = dasquery.service_apis_map()
if not service_map:
msg = 'no APIs found to answer input query, will decompose it'
self.logger.info(msg)
skeys = query['fields']
if not skeys:
skeys = []
for key in skeys:
newquery = DASQuery(dict(fields=[key], spec=query['spec']),
mongoparser=self.mongoparser)
self.call(newquery) # process query
else:
self.call(dasquery) # process query
# lookup provided query in a cache
if not self.noresults:
results = self.get_from_cache(dasquery, idx, limit)
return results
def remove_from_cache(self, dasquery):
"""
Delete in cache entries about input query
"""
self.rawcache.remove_from_cache(dasquery)
def get_status(self, dasquery):
"""
Look-up status of provided query in a cache.
Return status of the query request and its hash.
"""
if dasquery and dasquery.mongo_query.has_key('fields'):
fields = dasquery.mongo_query['fields']
if fields and isinstance(fields, list) and 'queries' in fields:
return 'ok', dasquery.qhash
status = 0
record = self.rawcache.find(dasquery)
try:
if record and record.has_key('das') and \
record['das'].has_key('status'):
status = record['das']['status']
return status, record['qhash']
except:
pass
similar_dasquery = self.rawcache.similar_queries(dasquery)
if similar_dasquery:
record = self.rawcache.find(similar_dasquery)
if record and record.has_key('das') and \
record['das'].has_key('status'):
similar_query_status = record['das']['status']
return similar_query_status, record['qhash']
return status, 0
def worker(self, srv, dasquery):
"""Main worker function which calls data-srv call function"""
self.logger.info('##### %s ######\n' % srv)
das_timer(srv, self.verbose)
getattr(getattr(self, srv), 'call')(dasquery)
das_timer(srv, self.verbose)
def call(self, query, add_to_analytics=True, **kwds):
"""
Top level DAS api which execute a given query using underlying
data-services. It follows the following steps:
- parse input query
- identify data-sercices based on selection keys
and where clause conditions
- construct DAS workflow and execute data-service
API calls. At this step individual
data-services store results into DAS cache.
Return status 0/1 depending on success of the calls, can be
used by workers on cache server.
kwds is provided for compatibility with web layer, e.g. it
may invoke this method with additional pid parameter.
"""
self.logger.info('input query=%s' % query)
das_timer('DASCore::call', self.verbose)
services = []
if isinstance(query, object) and hasattr(query, '__class__')\
and query.__class__.__name__ == 'DASQuery':
dasquery = query
else:
dasquery = DASQuery(query, mongoparser=self.mongoparser)
if add_to_analytics:
dasquery.add_to_analytics()
query = dasquery.mongo_query
if dasquery.mongo_query.has_key('system'):
system = query['system']
if isinstance(system, str) or isinstance(system, unicode):
services = [system]
elif isinstance(system, list):
services = system
else:
msg = 'Unsupported system=%s type=%s in DAS query' \
% (system, type(system))
raise Exception(msg)
spec = query.get('spec')
fields = query.get('fields')
if fields == ['records']:
msg = 'look-up all records in cache'
self.logger.info(msg)
return 'in cache'
if spec == dict(records='*'):
self.logger.info("look-up everything in cache")
return 'in cache'
for record in self.rawcache.find_specs(dasquery):
status = record['das']['status']
msg = 'found query %s in cache, status=%s\n' \
% (record['query'], status)
self.logger.info(msg)
return status
similar_dasquery = self.rawcache.similar_queries(dasquery)
if similar_dasquery:
for record in self.rawcache.find_specs(similar_dasquery):
if record:
try:
status = record['das']['status']
except:
status = 'N/A'
msg = 'Fail to look-up das.status, record=%s' % record
self.logger.info(msg)
msg = 'found SIMILAR query in cache,'
msg += 'query=%s, status=%s\n' % (record['query'], status)
self.logger.info(msg)
return status
self.logger.info(dasquery)
params = dasquery.params()
if not services:
services = params['services']
self.logger.info('services = %s' % services)
das_timer('das_record', self.verbose)
# initial expire tstamp 1 day (long enough to be overwriten by data-srv)
expire = expire_timestamp(time.time()+1*24*60*60)
header = dasheader("das", dasquery, expire)
header['lookup_keys'] = []
self.rawcache.insert_query_record(dasquery, header)
das_timer('das_record', self.verbose)
try:
if self.multitask:
jobs = []
for srv in services:
jobs.append(self.taskmgr.spawn(self.worker, srv, dasquery))
self.taskmgr.joinall(jobs)
else:
for srv in services:
self.worker(srv, dasquery)
except Exception as exc:
print_exc(exc)
return 'fail'
self.logger.info('\n##### merging ######\n')
self.rawcache.update_query_record(dasquery, 'merging')
das_timer('merge', self.verbose)
self.rawcache.merge_records(dasquery)
das_timer('merge', self.verbose)
self.rawcache.update_query_record(dasquery, 'ok')
self.rawcache.add_to_record(\
dasquery, {'das.timer': get_das_timer()}, system='das')
das_timer('DASCore::call', self.verbose)
return 'ok'
def nresults(self, dasquery, coll='merge'):
"""
Return total number of results (count) for provided query
Code should match body of get_from_cache method.
"""
fields = dasquery.mongo_query.get('fields', None)
if dasquery.mapreduce:
result = self.rawcache.map_reduce(dasquery.mapreduce, dasquery)
return len([1 for _ in result])
elif dasquery.aggregators:
return len(dasquery.aggregators)
elif isinstance(fields, list) and 'queries' in fields:
return len([1 for _ in self.get_queries(dasquery)])
return self.rawcache.nresults(dasquery, coll)
def incache(self, dasquery, coll='merge'):
"""
Answer the question if given query in DAS cache or not
"""
return self.rawcache.incache(dasquery, collection=coll)
def get_from_cache(self, dasquery, idx=0, limit=0, collection='merge'):
"""
Look-up results from the merge cache and yield them for
further processing.
"""
das_timer('DASCore::get_from_cache', self.verbose)
msg = 'col=%s, query=%s, idx=%s, limit=%s'\
% (collection, dasquery, idx, limit)
self.logger.info(msg)
fields = dasquery.mongo_query.get('fields', None)
if dasquery.mapreduce:
res = self.rawcache.map_reduce(dasquery.mapreduce, dasquery)
elif dasquery.aggregators:
res = []
_id = 0
for func, key in dasquery.aggregators:
rows = self.rawcache.get_from_cache(\
dasquery, collection=collection)
data = getattr(das_aggregator, 'das_%s' % func)(key, rows)
res += \
[{'_id':_id, 'function': func, 'key': key, 'result': data}]
_id += 1
elif isinstance(fields, list) and 'queries' in fields:
res = itertools.islice(self.get_queries(dasquery), idx, idx+limit)
else:
res = self.rawcache.get_from_cache(dasquery, idx, limit, \
collection=collection)
for row in res:
fix_times(row)
yield row
das_timer('DASCore::get_from_cache', self.verbose)
def get_queries(self, dasquery):
"""
Look-up (popular) queries in DAS analytics/logging db
"""
das_timer('DASCore::get_queries', self.verbose)
fields = dasquery.mongo_query.get('fields')
spec = dasquery.mongo_query.get('spec')
if 'popular' in fields:
res = self.analytics.get_popular_queries(spec)
else:
datestamp = spec.get('date')
if isinstance(datestamp, dict):
value = datestamp.get('$in')
res = \
self.analytics.list_queries(after=value[0], before=value[1])
elif isinstance(datestamp, int):
res = self.analytics.list_queries(after=datestamp)
elif not datestamp:
res = self.analytics.list_queries()
else:
msg = 'Unsupported date value: %s' % datestamp
raise Exception(msg)
for row in res:
rid = row.pop('_id')
yield dict(das_query=row, _id=rid)
das_timer('DASCore::get_queries', self.verbose)
class QLManager(object):
"""
DAS QL manager.
"""
def __init__(self, config=None):
if not config:
config = das_readconfig()
self.dasmapping = DASMapping(config)
if not self.dasmapping.check_maps():
msg = "No DAS maps found in MappingDB"
raise Exception(msg)
self.analytics = DASAnalytics(config)
self.dasservices = config['services']
self.daskeysmap = self.dasmapping.daskeys()
self.operators = list(das_operators())
self.daskeys = list(das_special_keys())
self.verbose = config['verbose']
self.logger = PrintManager('QLManger', self.verbose)
for val in self.daskeysmap.values():
for item in val:
self.daskeys.append(item)
parserdir = config['das']['parserdir']
self.parserdir = parserdir
self.enabledb = config['parserdb']['enable']
if self.enabledb:
self.parserdb = DASParserDB(config)
def parse(self, query):
"""
Parse input query and return query in MongoDB form.
Optionally parsed query can be written into analytics DB.
"""
mongo_query = self.mongo_query(query)
self.convert2skeys(mongo_query)
return mongo_query
def add_to_analytics(self, query, mongo_query):
"Add DAS query to analytics DB"
self.analytics.add_query(query, mongo_query)
def get_ply_query(self, query):
"""
Get ply object for given query. Since we rely on PLY package and it may
fail under the load we use couple of trials.
"""
ply_query = ply_parse_query(query, self.daskeys, self.dasservices,
self.parserdir, self.verbose)
return ply_query
def mongo_query(self, query):
"""
Return mongo query for provided input query
"""
mongo_query = None
if self.verbose:
ply_output(query, self.daskeys, self.dasservices,
self.parserdir, self.verbose)
parse_again = True
if self.enabledb:
status, value = self.parserdb.lookup_query(query)
if status == PARSERCACHE_VALID and \
len(last_key_pattern.findall(query)) == 0:
mongo_query = value
parse_again = False
elif status == PARSERCACHE_INVALID:
# we unable to find query in parserdb, so will parse again
parse_again = True
else:
ply_query = self.get_ply_query(query)
if ply_query:
try:
mongo_query = ply2mongo(ply_query)
parse_again = False
except Exception as exc:
msg = "Fail in ply2mongo, query=%s, ply_query=%s" \
% (query, ply_query)
print msg
try:
self.parserdb.insert_valid_query(query, mongo_query)
except Exception as exc:
msg = "Fail to insert into parserdb, exception=%s" \
% str(exc)
print_exc(msg, print_traceback=True)
if parse_again:
try:
ply_query = self.get_ply_query(query)
mongo_query = ply2mongo(ply_query)
except Exception as exc:
msg = "Fail to parse query='%s'" % query
print_exc(msg, print_traceback=False)
raise exc
if set(mongo_query.keys()) & set(['fields', 'spec']) != \
set(['fields', 'spec']):
raise Exception('Invalid MongoDB query %s' % mongo_query)
if not mongo_query['fields'] and len(mongo_query['spec'].keys()) > 1:
raise Exception(ambiguous_msg(query, mongo_query['spec'].keys()))
for key, val in mongo_query['spec'].iteritems():
if isinstance(val, list):
raise Exception(ambiguos_val_msg(query, key, val))
return mongo_query
def convert2skeys(self, mongo_query):
"""
Convert DAS input keys into DAS selection keys.
"""
if not mongo_query['spec']:
for key in mongo_query['fields']:
for system in self.dasmapping.list_systems():
mapkey = self.dasmapping.find_mapkey(system, key)
if mapkey:
mongo_query['spec'][mapkey] = '*'
return
spec = mongo_query['spec']
to_replace = []
for key, val in spec.iteritems():
for system in self.dasmapping.list_systems():
mapkey = self.dasmapping.find_mapkey(system, key, val)
if mapkey and mapkey != key and \
key in mongo_query['spec']:
to_replace.append((key, mapkey))
continue
for key, mapkey in to_replace:
if key in mongo_query['spec']:
mongo_query['spec'][mapkey] = mongo_query['spec'][key]
del mongo_query['spec'][key]
def services(self, query):
"""Find out DAS services to use for provided query"""
skeys, cond = decompose(query)
if not skeys:
skeys = []
if isinstance(skeys, str):
skeys = [skeys]
slist = []
# look-up services from Mapping DB
for key in skeys + [i for i in cond.keys()]:
for service, keys in self.daskeysmap.iteritems():
if service not in self.dasservices:
continue
value = cond.get(key, None)
daskeys = self.dasmapping.find_daskey(service, key, value)
if set(keys) & set(daskeys) and service not in slist:
slist.append(service)
# look-up special key condition
requested_system = query.get('system', None)
if requested_system:
if isinstance(requested_system, basestring):
requested_system = [requested_system]
return list( set(slist) & set(requested_system) )
return slist
def service_apis_map(self, query):
"""
Find out which APIs correspond to provided query.
Return a map of found services and their apis.
"""
skeys, cond = decompose(query)
if not skeys:
skeys = []
if isinstance(skeys, str):
skeys = [skeys]
adict = {}
mapkeys = [key for key in cond.keys() if key not in das_special_keys()]
services = self.services(query)
for srv in services:
alist = self.dasmapping.list_apis(srv)
for api in alist:
daskeys = self.dasmapping.api_info(srv, api)['das_map']
maps = [r['rec_key'] for r in daskeys]
if set(mapkeys) & set(maps) == set(mapkeys):
if srv in adict:
new_list = adict[srv] + [api]
adict[srv] = list( set(new_list) )
else:
adict[srv] = [api]
return adict
def params(self, query):
"""
Return dictionary of parameters to be used in DAS Core:
selection keys, conditions and services.
"""
skeys, cond = decompose(query)
services = []
for srv in self.services(query):
if srv not in services:
services.append(srv)
return dict(selkeys=skeys, conditions=cond, services=services)
