def connect_if_not_connected(cls):
"""Connect to Celery if not already connected, do it safely as we have concurrent server"""
if cls._connected:
return
with cls._setup_lock:
if not cls._connected:
# We do not have to have result backend on server since we are just pushing tasks to message queue
init_celery(result_backend=False)
init_selinon()
def __init__(self):
self.log = logging.getLogger(__name__)
formatter = logging.Formatter(
'%(asctime)s - %(name)s - %(levelname)s - %(message)s')
logging_handler = logging.StreamHandler(sys.stdout)
logging_handler.setFormatter(formatter)
self.log.addHandler(logging_handler)
self.log.level = logging.DEBUG
if ENABLE_SCHEDULING:
init_celery(result_backend=False)
init_selinon()
def __init__(self):
"""Initialize Relase Monitor."""
logger.info("Starting the monitor service")
# Create PyPi monitor
self.pypi_monitor = PypiMonitor()
# Create NPM monitor
self.npm_monitor = NPMMonitor()
# Initialize Selinon if we want to run in production
if ENABLE_SCHEDULING:
init_celery(result_backend=False)
init_selinon()
def server_run_flow(flow_name, flow_args):
"""Run a flow.
:param flow_name: name of flow to be run as stated in YAML config file
:param flow_args: arguments for the flow
:return: dispatcher ID handling flow
"""
logger.info('Running flow {}'.format(flow_name))
start = datetime.datetime.now()
init_celery(result_backend=False)
dispacher_id = run_flow(flow_name, flow_args)
elapsed_seconds = (datetime.datetime.now() - start).total_seconds()
logger.info("It took {t} seconds to start {f} flow.".format(
t=elapsed_seconds, f=flow_name))
return dispacher_id
def server_run_flow(flow_name, flow_args):
"""Run a flow.
:param flow_name: name of flow to be run as stated in YAML config file
:param flow_args: arguments for the flow
:return: dispatcher ID handling flow
"""
logger.debug('Running flow %s', flow_name)
start = datetime.datetime.now()
init_celery(result_backend=False)
dispacher_id = run_flow(flow_name, flow_args)
logger.debug('It took %f seconds to start %s flow.',
(datetime.datetime.now() - start).total_seconds(), flow_name)
return dispacher_id
def construct_analyses_report(ecosystem, from_date=None, to_date=None):
"""Construct analyses state report.
:param ecosystem: name of the ecosystem
:param from_date: datetime limitation
:type from_date: datetime.datetime
:param to_date: datetime limitation
:type to_date: datetime.datetime
:return: a dict describing the current system state
:rtype: dict
"""
report = {
'report': {},
'from_date': str(from_date) if from_date else None,
'to_date': str(to_date) if to_date else None,
'now': str(datetime.now())
}
# TODO: init only Selinon
# there is required only Selinon configuration, we don't need to connect to queues,
# but let's stick with this for now
init_celery(result_backend=False)
db = StoragePool.get_connected_storage('BayesianPostgres')
finished_analyses = _get_finished_analyses_count(db, ecosystem, from_date,
to_date)
unfinished_analyses = _get_unfinished_analyses_count(
db, ecosystem, from_date, to_date)
report['report']['ecosystem'] = ecosystem
report['report']['analyses'] = finished_analyses + unfinished_analyses
report['report']['analyses_finished'] = finished_analyses
report['report']['analyses_unfinished'] = unfinished_analyses
report['report'][
'analyses_finished_unique'] = _get_unique_finished_analyses_count(
db, ecosystem, from_date, to_date)
report['report']['analyses_unique'] = _get_unique_analyses_count(
db, ecosystem, from_date, to_date)
report['report']['packages'] = _get_packages_count(db, ecosystem,
from_date, to_date)
report['report']['packages_finished'] = _get_finished_packages_count(
db, ecosystem, from_date, to_date)
report['report']['versions'] = _get_versions_count(db, ecosystem,
from_date, to_date)
return report
def __init__(self):
"""Constructor."""
self.log = logging.getLogger(__name__)
formatter = logging.Formatter('%(asctime)s - '
'%(name)s - %(levelname)s'
' - %(message)s')
logging_handler = logging.StreamHandler(sys.stdout)
logging_handler.setFormatter(formatter)
self.log.addHandler(logging_handler)
self.log.level = logging.DEBUG
self.old_npm_feed = None
self.npm_feed = feedparser.parse(NPM_URL + "-/rss")
self.old_pypi_feed = None
self.pypi_feed = feedparser.parse(PYPI_URL + "rss/updates.xml")
self.create_liveness_probe()
if ENABLE_SCHEDULING:
init_celery(result_backend=False)
init_selinon()
#!/usr/bin/env python
"""Start the application."""
import celery
import os
from f8a_worker.setup_celery import init_celery, init_selinon
import raven
from raven.contrib.celery import register_signal, register_logger_signal
class SentryCelery(celery.Celery):
"""Celery class to configure sentry."""
def on_configure(self):
"""Set up sentry client."""
dsn = os.environ.get("SENTRY_DSN")
client = raven.Client(dsn)
register_logger_signal(client)
register_signal(client)
app = SentryCelery('tasks')
init_celery(app)
init_selinon(app)
init_logging(logger)
app.add_api(defaults.SWAGGER_YAML_PATH)
app.add_api(defaults.SWAGGER_INGESTION_YAML_PATH)
# Expose for uWSGI
application.json_encoder = SafeJSONEncoder
manager = Manager(application)
# Needed for session
application.secret_key = defaults.APP_SECRET_KEY
oauth.init_app(application)
# Initializing Selinon and Celery while starting the application
logger.debug("Initializing Selinon")
init_celery(result_backend=False)
init_selinon()
logger.debug("Selinon initialized successfully")
@app.route('/')
def base_url():
"""Redirect client to the Swagger UI web page."""
# Be nice with user access
return redirect('api/v1/ui')
@app.route('/api/v1')
def api_v1():
"""Accept and respont to all REST API calls."""
paths = []
def __init__(self):
"""Create this backend."""
super().__init__()
init_celery(result_backend=False)
init_selinon()
def _init_celery(self):
""" Initialize celery and connect to the broker """
if not self._initialized_celery:
init_celery(result_backend=False)
self._initialized_celery = True
"""Get expiration date after which GitHub data are considered outdated."""
return datetime.utcnow() - timedelta(days=REFRESH_INTERVAL)
def get_limit(db):
"""Get limit how many GitHub statistics can be updated at once."""
return math.ceil(_get_count(db) / REFRESH_INTERVAL / 2)
def is_dry_run():
"""Return True if this is a dry run."""
return os.environ.get('DRY_RUN', 'false').lower() in ('1', 'yes', 'true')
def _get_count(db):
"""Count how many GitHub data are outdated."""
return db.query(Upstream).filter(Upstream.updated_at < get_expiration_date())\
.filter(Upstream.url.like('%github.com%'), Upstream.deactivated_at == None).count()
def _get_outdated(db, limit):
"""Get EPs with outdated GitHub data."""
return db.query(Upstream).join(Package).join(Ecosystem).filter(
Upstream.url.like('%github.com%'), Upstream.deactivated_at == None).limit(limit).all()
if __name__ == '__main__':
init_celery()
init_selinon()
run(sessionmaker(bind=create_engine(CONN_STR))())
def run_server_flow(flow_name, flow_args):
"""To run the worker flow via selinon."""
init_celery(result_backend=False)
init_selinon()
return run_flow(flow_name, flow_args)
