def test_temp_blocked(self):
utcnow = util.utcnow()
bad_wifi = WifiObservationFactory.build()
good_wifi = WifiObservationFactory.build()
WifiShardFactory(mac=bad_wifi.mac,
lat=None,
lon=None,
created=utcnow,
block_first=utcnow.date() - timedelta(days=10),
block_last=utcnow.date(),
block_count=1)
obs = [good_wifi, bad_wifi, good_wifi]
self.session.commit()
self._queue_and_update(obs)
shard = WifiShard.shard_model(good_wifi.mac)
wifis = (self.session.query(shard).filter(
shard.mac == good_wifi.mac)).all()
self.assertEqual(len(wifis), 1)
self.assertTrue(wifis[0].lat is not None)
self.assertTrue(wifis[0].lon is not None)
self.assertEqual(wifis[0].samples, 2)
shard = WifiShard.shard_model(bad_wifi.mac)
wifis = (self.session.query(shard).filter(
shard.mac == bad_wifi.mac)).all()
self.assertEqual(len(wifis), 1)
self.assertTrue(wifis[0].block_first < utcnow.date())
self.assertTrue(wifis[0].lat is None)
self.assertTrue(wifis[0].lon is None)
self.check_statcounter(StatKey.wifi, 2)
self.check_statcounter(StatKey.unique_wifi, 1)
def test_temp_blocked(self):
utcnow = util.utcnow()
bad_wifi = WifiObservationFactory.build()
good_wifi = WifiObservationFactory.build()
WifiShardFactory(
mac=bad_wifi.mac,
lat=None,
lon=None,
created=utcnow,
block_first=utcnow.date() - timedelta(days=10),
block_last=utcnow.date(),
block_count=1)
obs = [good_wifi, bad_wifi, good_wifi]
self.session.commit()
self._queue_and_update(obs)
shard = WifiShard.shard_model(good_wifi.mac)
wifis = (self.session.query(shard)
.filter(shard.mac == good_wifi.mac)).all()
self.assertEqual(len(wifis), 1)
self.assertTrue(wifis[0].lat is not None)
self.assertTrue(wifis[0].lon is not None)
self.assertEqual(wifis[0].samples, 2)
shard = WifiShard.shard_model(bad_wifi.mac)
wifis = (self.session.query(shard)
.filter(shard.mac == bad_wifi.mac)).all()
self.assertEqual(len(wifis), 1)
self.assertTrue(wifis[0].block_first < utcnow.date())
self.assertTrue(wifis[0].lat is None)
self.assertTrue(wifis[0].lon is None)
self.check_statcounter(StatKey.wifi, 2)
self.check_statcounter(StatKey.unique_wifi, 1)
def test_update(self):
utcnow = util.utcnow()
obs = []
obs_factory = WifiObservationFactory
# first wifi
wifi1 = WifiShardFactory(lat=None, lon=None, samples=3)
new_pos = WifiShardFactory.build()
mac1, lat1, lon1 = (wifi1.mac, new_pos.lat, new_pos.lon)
obs.extend([
obs_factory(lat=lat1,
lon=lon1, key=mac1),
obs_factory(lat=lat1 + 0.002,
lon=lon1 + 0.003, key=mac1),
obs_factory(lat=lat1 + 0.004,
lon=lon1 + 0.006, key=mac1),
])
# second wifi
wifi2 = WifiShardFactory(
lat=lat1 + 1.0, lon=lon1 + 1.0,
max_lat=lat1 + 1.0, min_lat=lat1 + 0.999,
max_lon=lon1 + 1.0, min_lon=None,
radius=20, samples=2,
created=utcnow - timedelta(10),
modified=utcnow - timedelta(10))
mac2, lat2, lon2 = (wifi2.mac, wifi2.lat, wifi2.lon)
obs.extend([
obs_factory(lat=lat2 + 0.002,
lon=lon2 + 0.004, key=mac2),
obs_factory(lat=lat2 + 0.002,
lon=lon2 + 0.004, key=mac2),
])
self.session.commit()
self._queue_and_update(obs)
shard = WifiShard.shard_model(mac1)
found = self.session.query(shard).filter(shard.mac == mac1).one()
self.assertAlmostEqual(found.lat, lat1 + 0.002)
self.assertAlmostEqual(found.max_lat, lat1 + 0.004)
self.assertAlmostEqual(found.min_lat, lat1)
self.assertAlmostEqual(found.lon, lon1 + 0.003)
self.assertAlmostEqual(found.max_lon, lon1 + 0.006)
self.assertAlmostEqual(found.min_lon, lon1)
self.assertEqual(found.modified.date(), utcnow.date())
self.assertEqual(found.radius, 304)
self.assertEqual(found.region, 'GB')
self.assertEqual(found.samples, 6)
shard = WifiShard.shard_model(mac2)
found = self.session.query(shard).filter(shard.mac == mac2).one()
self.assertAlmostEqual(found.lat, lat2 + 0.001)
self.assertAlmostEqual(found.max_lat, lat2 + 0.002)
self.assertAlmostEqual(found.min_lat, lat2 - 0.001)
self.assertAlmostEqual(found.lon, lon2 + 0.002)
self.assertAlmostEqual(found.max_lon, lon2 + 0.004)
self.assertAlmostEqual(found.min_lon, lon2)
self.assertEqual(found.created.date(), utcnow.date() - timedelta(10))
self.assertEqual(found.modified.date(), utcnow.date())
self.assertEqual(found.radius, 260)
self.assertEqual(found.region, 'GB')
self.assertEqual(found.samples, 4)
def test_update(self):
utcnow = util.utcnow()
obs = []
obs_factory = WifiObservationFactory
# first wifi
wifi1 = WifiShardFactory(lat=None, lon=None, samples=3)
new_pos = WifiShardFactory.build()
mac1, lat1, lon1 = (wifi1.mac, new_pos.lat, new_pos.lon)
obs.extend([
obs_factory(lat=lat1, lon=lon1, key=mac1),
obs_factory(lat=lat1 + 0.002, lon=lon1 + 0.003, key=mac1),
obs_factory(lat=lat1 + 0.004, lon=lon1 + 0.006, key=mac1),
])
# second wifi
wifi2 = WifiShardFactory(lat=lat1 + 1.0,
lon=lon1 + 1.0,
max_lat=lat1 + 1.0,
min_lat=lat1 + 0.999,
max_lon=lon1 + 1.0,
min_lon=None,
radius=20,
samples=2,
created=utcnow - timedelta(10),
modified=utcnow - timedelta(10))
mac2, lat2, lon2 = (wifi2.mac, wifi2.lat, wifi2.lon)
obs.extend([
obs_factory(lat=lat2 + 0.002, lon=lon2 + 0.004, key=mac2),
obs_factory(lat=lat2 + 0.002, lon=lon2 + 0.004, key=mac2),
])
self.session.commit()
self._queue_and_update(obs)
shard = WifiShard.shard_model(mac1)
found = self.session.query(shard).filter(shard.mac == mac1).one()
self.assertAlmostEqual(found.lat, lat1 + 0.002)
self.assertAlmostEqual(found.max_lat, lat1 + 0.004)
self.assertAlmostEqual(found.min_lat, lat1)
self.assertAlmostEqual(found.lon, lon1 + 0.003)
self.assertAlmostEqual(found.max_lon, lon1 + 0.006)
self.assertAlmostEqual(found.min_lon, lon1)
self.assertEqual(found.modified.date(), utcnow.date())
self.assertEqual(found.radius, 304)
self.assertEqual(found.region, 'GB')
self.assertEqual(found.samples, 6)
shard = WifiShard.shard_model(mac2)
found = self.session.query(shard).filter(shard.mac == mac2).one()
self.assertAlmostEqual(found.lat, lat2 + 0.001)
self.assertAlmostEqual(found.max_lat, lat2 + 0.002)
self.assertAlmostEqual(found.min_lat, lat2 - 0.001)
self.assertAlmostEqual(found.lon, lon2 + 0.002)
self.assertAlmostEqual(found.max_lon, lon2 + 0.004)
self.assertAlmostEqual(found.min_lon, lon2)
self.assertEqual(found.created.date(), utcnow.date() - timedelta(10))
self.assertEqual(found.modified.date(), utcnow.date())
self.assertEqual(found.radius, 260)
self.assertEqual(found.region, 'GB')
self.assertEqual(found.samples, 4)
def query_wifis(query, raven_client):
macs = [lookup.mac for lookup in query.wifi]
if not macs: # pragma: no cover
return []
result = []
today = util.utcnow().date()
temp_blocked = today - TEMPORARY_BLOCKLIST_DURATION
try:
# load all fields used in score calculation and those we
# need for the position
load_fields = ('lat', 'lon', 'radius', 'created', 'modified',
'samples')
shards = defaultdict(list)
for mac in macs:
shards[WifiShard.shard_model(mac)].append(mac)
for shard, shard_macs in shards.items():
rows = (query.session.query(shard).filter(
shard.mac.in_(shard_macs)).filter(
shard.lat.isnot(None)).filter(
shard.lon.isnot(None)).filter(
or_(
shard.block_count.is_(None), shard.block_count
< PERMANENT_BLOCKLIST_THRESHOLD)).filter(
or_(shard.block_last.is_(None),
shard.block_last <
temp_blocked)).options(
load_only(*load_fields))).all()
result.extend(list(rows))
except Exception:
raven_client.captureException()
return result
def test_weighted_update(self):
wifi = WifiShardFactory(samples=2, weight=3.0)
wifi_lat = wifi.lat
wifi_lon = wifi.lon
obs_factory = WifiObservationFactory
obs = [
obs_factory(lat=wifi.lat, lon=wifi.lon - 0.002,
accuracy=20.0, signal=-30, mac=wifi.mac),
obs_factory(lat=wifi.lat, lon=wifi.lon - 0.004,
accuracy=40.0, signal=-60, mac=wifi.mac),
]
self.session.commit()
self._queue_and_update_wifi(obs)
shard = WifiShard.shard_model(wifi.mac)
wifis = self.session.query(shard).all()
self.assertEqual(len(wifis), 1)
wifi = wifis[0]
self.assertAlmostEqual(wifi.lat, wifi_lat)
self.assertAlmostEqual(wifi.max_lat, wifi_lat)
self.assertAlmostEqual(wifi.min_lat, wifi_lat)
self.assertAlmostEqual(wifi.lon, wifi_lon - 0.0017709, 7)
self.assertAlmostEqual(wifi.max_lon, wifi_lon)
self.assertAlmostEqual(wifi.min_lon, wifi_lon - 0.004)
self.assertEqual(wifi.radius, 154)
self.assertEqual(wifi.samples, 4)
self.assertAlmostEqual(wifi.weight, 15.53, 2)
def test_temp_blocked_admitted_again(self):
now = util.utcnow()
last_week = now - TEMPORARY_BLOCKLIST_DURATION - timedelta(days=1)
obs = WifiObservationFactory()
WifiShardFactory(
mac=obs.mac,
lat=None,
lon=None,
samples=0,
created=last_week,
modified=last_week,
block_first=last_week.date(),
block_last=last_week.date(),
block_count=1)
self.session.flush()
# add a new entry for the previously blocked wifi
self.data_queue.enqueue([obs])
self.assertEqual(self.data_queue.size(), 1)
update_wifi.delay().get()
# the wifi was inserted again
shard = WifiShard.shard_model(obs.mac)
wifis = self.session.query(shard).all()
self.assertEqual(len(wifis), 1)
wifi = wifis[0]
self.assertEqual(wifi.created.date(), last_week.date())
self.assertAlmostEqual(wifi.lat, obs.lat)
self.assertAlmostEqual(wifi.lon, obs.lon)
self.assertEqual(wifi.country, 'GB')
self.assertEqual(wifi.samples, 1)
self.check_statcounter(StatKey.unique_wifi, 0)
def configure_data(redis_client):
"""
Configure fixed set of data queues.
"""
data_queues = {
'update_cell':
DataQueue('update_cell', redis_client, queue_key='update_cell'), # BBB
'update_cellarea':
DataQueue('update_cellarea', redis_client,
queue_key='update_cellarea'),
'update_cellarea_ocid':
DataQueue('update_cellarea_ocid',
redis_client,
queue_key='update_cellarea_ocid'),
'update_score':
DataQueue('update_score', redis_client, queue_key='update_score'),
}
for shard_id in DataMap.shards().keys():
name = 'update_datamap_' + shard_id
data_queues[name] = DataQueue(name, redis_client, queue_key=name)
for shard_id in CellShard.shards().keys():
name = 'update_cell_' + shard_id
data_queues[name] = DataQueue(name, redis_client, queue_key=name)
for shard_id in WifiShard.shards().keys():
name = 'update_wifi_' + shard_id
data_queues[name] = DataQueue(name, redis_client, queue_key=name)
return data_queues
def test_temp_blocked_admitted_again(self):
now = util.utcnow()
last_week = now - TEMPORARY_BLOCKLIST_DURATION - timedelta(days=1)
obs = WifiObservationFactory()
WifiShardFactory(mac=obs.mac,
lat=None,
lon=None,
samples=0,
created=last_week,
modified=last_week,
block_first=last_week.date(),
block_last=last_week.date(),
block_count=1)
self.session.commit()
# add a new entry for the previously blocked wifi
self._queue_and_update([obs])
# the wifi was inserted again
shard = WifiShard.shard_model(obs.mac)
wifis = self.session.query(shard).all()
self.assertEqual(len(wifis), 1)
wifi = wifis[0]
self.assertEqual(wifi.block_first, last_week.date())
self.assertEqual(wifi.block_last, last_week.date())
self.assertEqual(wifi.created.date(), last_week.date())
self.assertAlmostEqual(wifi.lat, obs.lat)
self.assertAlmostEqual(wifi.lon, obs.lon)
self.assertEqual(wifi.region, 'GB')
self.assertEqual(wifi.samples, 1)
self.check_statcounter(StatKey.unique_wifi, 0)
def configure_data(redis_client):
"""
Configure fixed set of data queues.
"""
data_queues = {
'update_cell': DataQueue('update_cell', redis_client,
queue_key='update_cell'), # BBB
'update_cellarea': DataQueue('update_cellarea', redis_client,
queue_key='update_cellarea'),
'update_cellarea_ocid': DataQueue('update_cellarea_ocid', redis_client,
queue_key='update_cellarea_ocid'),
'update_score': DataQueue('update_score', redis_client,
queue_key='update_score'),
}
for shard_id in DataMap.shards().keys():
name = 'update_datamap_' + shard_id
data_queues[name] = DataQueue(name, redis_client, queue_key=name)
for shard_id in CellShard.shards().keys():
name = 'update_cell_' + shard_id
data_queues[name] = DataQueue(
name, redis_client, queue_key=name)
for shard_id in WifiShard.shards().keys():
name = 'update_wifi_' + shard_id
data_queues[name] = DataQueue(
name, redis_client, queue_key=name)
return data_queues
def test_position_invalid(self, celery, session, metricsmock):
self.add_reports(celery,
1,
cell_factor=0,
wifi_factor=1,
wifi_key="000000123456",
lat=-90.1)
self.add_reports(celery,
1,
cell_factor=0,
wifi_factor=1,
wifi_key="000000234567")
self._update_all(session)
shard = WifiShard.shards()["0"]
assert session.query(shard).count() == 1
metricsmock.assert_incr_once("data.report.upload",
value=2,
tags=["key:test"])
metricsmock.assert_incr_once("data.report.drop",
value=1,
tags=["key:test"])
metricsmock.assert_incr_once("data.observation.insert",
value=1,
tags=["type:wifi"])
metricsmock.assert_incr_once("data.observation.upload",
tags=["type:wifi", "key:test"])
def query_database(query, raven_client):
macs = [lookup.mac for lookup in query.wifi]
if not macs: # pragma: no cover
return []
result = []
today = util.utcnow().date()
temp_blocked = today - TEMPORARY_BLOCKLIST_DURATION
try:
load_fields = ('lat', 'lon', 'radius')
shards = defaultdict(list)
for mac in macs:
shards[WifiShard.shard_model(mac)].append(mac)
for shard, shard_macs in shards.items():
rows = (
query.session.query(shard)
.filter(shard.mac.in_(shard_macs))
.filter(shard.lat.isnot(None))
.filter(shard.lon.isnot(None))
.filter(or_(
shard.block_count.is_(None),
shard.block_count <
PERMANENT_BLOCKLIST_THRESHOLD))
.filter(or_(
shard.block_last.is_(None),
shard.block_last < temp_blocked))
.options(load_only(*load_fields))
).all()
result.extend(list(rows))
except Exception:
raven_client.captureException()
return result
def test_new(self):
utcnow = util.utcnow()
obs = WifiObservationFactory.build()
self.data_queue.enqueue([obs])
self.assertEqual(self.data_queue.size(), 1)
update_wifi.delay().get()
shard = WifiShard.shard_model(obs.mac)
wifis = self.session.query(shard).all()
self.assertEqual(len(wifis), 1)
wifi = wifis[0]
self.assertAlmostEqual(wifi.lat, obs.lat)
self.assertAlmostEqual(wifi.max_lat, obs.lat)
self.assertAlmostEqual(wifi.min_lat, obs.lat)
self.assertAlmostEqual(wifi.lon, obs.lon)
self.assertAlmostEqual(wifi.max_lon, obs.lon)
self.assertAlmostEqual(wifi.min_lon, obs.lon)
self.assertEqual(wifi.country, 'GB')
self.assertEqual(wifi.radius, 0)
self.assertEqual(wifi.samples, 1)
self.assertEqual(wifi.created.date(), utcnow.date())
self.assertEqual(wifi.modified.date(), utcnow.date())
self.assertEqual(wifi.block_first, None)
self.assertEqual(wifi.block_last, None)
self.assertEqual(wifi.block_count, None)
def configure_data(redis_client):
"""
Configure fixed set of data queues.
"""
data_queues = {
# *_incoming need to be the exact same as in webapp.config
'update_incoming': DataQueue('update_incoming', redis_client,
batch=100, compress=True),
'transfer_incoming': DataQueue('transfer_incoming', redis_client,
batch=100, compress=True),
}
for key in ('update_cellarea', ):
data_queues[key] = DataQueue(key, redis_client, batch=100, json=False)
for shard_id in BlueShard.shards().keys():
key = 'update_blue_' + shard_id
data_queues[key] = DataQueue(key, redis_client, batch=500)
for shard_id in DataMap.shards().keys():
key = 'update_datamap_' + shard_id
data_queues[key] = DataQueue(key, redis_client, batch=500, json=False)
for shard_id in CellShard.shards().keys():
key = 'update_cell_' + shard_id
data_queues[key] = DataQueue(key, redis_client, batch=500)
for shard_id in WifiShard.shards().keys():
key = 'update_wifi_' + shard_id
data_queues[key] = DataQueue(key, redis_client, batch=500)
return data_queues
def test_weighted_update(self):
wifi = WifiShardFactory(samples=2, weight=3.0)
wifi_lat = wifi.lat
wifi_lon = wifi.lon
obs_factory = WifiObservationFactory
obs = [
obs_factory(lat=wifi.lat,
lon=wifi.lon - 0.002,
accuracy=20.0,
signal=-30,
mac=wifi.mac),
obs_factory(lat=wifi.lat,
lon=wifi.lon - 0.004,
accuracy=40.0,
signal=-60,
mac=wifi.mac),
]
self.session.commit()
self._queue_and_update_wifi(obs)
shard = WifiShard.shard_model(wifi.mac)
wifis = self.session.query(shard).all()
self.assertEqual(len(wifis), 1)
wifi = wifis[0]
self.assertAlmostEqual(wifi.lat, wifi_lat)
self.assertAlmostEqual(wifi.max_lat, wifi_lat)
self.assertAlmostEqual(wifi.min_lat, wifi_lat)
self.assertAlmostEqual(wifi.lon, wifi_lon - 0.0017709, 7)
self.assertAlmostEqual(wifi.max_lon, wifi_lon)
self.assertAlmostEqual(wifi.min_lon, wifi_lon - 0.004)
self.assertEqual(wifi.radius, 154)
self.assertEqual(wifi.samples, 4)
self.assertAlmostEqual(wifi.weight, 15.53, 2)
def _update_all(self):
schedule_export_reports.delay().get()
for shard_id in CellShard.shards().keys():
update_cell.delay(shard_id=shard_id).get()
for shard_id in WifiShard.shards().keys():
update_wifi.delay(shard_id=shard_id).get()
def _update_all(self):
schedule_export_reports.delay().get()
for shard_id in CellShard.shards().keys():
update_cell.delay(shard_id=shard_id).get()
for shard_id in WifiShard.shards().keys():
update_wifi.delay(shard_id=shard_id).get()
def _queue_and_update(self, obs):
sharded_obs = defaultdict(list)
for ob in obs:
sharded_obs[WifiShard.shard_id(ob.mac)].append(ob)
for shard_id, values in sharded_obs.items():
queue = self.celery_app.data_queues['update_wifi_' + shard_id]
queue.enqueue(values)
update_wifi.delay(shard_id=shard_id).get()
def _shard_observations(self, observations):
sharded_obs = {}
for obs in observations:
if obs is not None:
shard = WifiShard.shard_model(obs.mac)
if shard not in sharded_obs:
sharded_obs[shard] = defaultdict(list)
sharded_obs[shard][obs.mac].append(obs)
return sharded_obs
def _queue_and_update(self, obs):
sharded_obs = defaultdict(list)
for ob in obs:
sharded_obs[WifiShard.shard_id(ob.mac)].append(ob)
for shard_id, values in sharded_obs.items():
queue = self.celery_app.data_queues['update_wifi_' + shard_id]
queue.enqueue(values)
update_wifi.delay(shard_id=shard_id).get()
def _update_all(self):
update_incoming.delay().get()
for shard_id in BlueShard.shards().keys():
update_blue.delay(shard_id=shard_id).get()
for shard_id in CellShard.shards().keys():
update_cell.delay(shard_id=shard_id).get()
for shard_id in WifiShard.shards().keys():
update_wifi.delay(shard_id=shard_id).get()
def test_wifi(self, celery, session):
reports = self.add_reports(celery, cell_factor=0, wifi_factor=1)
self._update_all(session)
position = reports[0]['position']
wifi_data = reports[0]['wifiAccessPoints'][0]
shard = WifiShard.shard_model(wifi_data['macAddress'])
wifis = session.query(shard).all()
assert len(wifis) == 1
wifi = wifis[0]
assert wifi.lat == position['latitude']
assert wifi.lon == position['longitude']
assert wifi.mac == wifi_data['macAddress']
assert wifi.samples == 1
def test_wifi(self, celery, session):
reports = self.add_reports(celery, cell_factor=0, wifi_factor=1)
self._update_all(session)
position = reports[0]["position"]
wifi_data = reports[0]["wifiAccessPoints"][0]
shard = WifiShard.shard_model(wifi_data["macAddress"])
wifis = session.query(shard).all()
assert len(wifis) == 1
wifi = wifis[0]
assert wifi.lat == position["latitude"]
assert wifi.lon == position["longitude"]
assert wifi.mac == wifi_data["macAddress"]
assert wifi.samples == 1
def __call__(self):
cells = (self.session.query(CellArea.region,
CellArea.radio,
func.sum(CellArea.num_cells))
.filter(CellArea.region.isnot(None))
.group_by(CellArea.region, CellArea.radio)).all()
default = {'gsm': 0, 'wcdma': 0, 'lte': 0, 'wifi': 0}
stats = {}
for region, radio, num in cells:
if region not in stats:
stats[region] = default.copy()
stats[region][radio.name] = int(num)
for shard in WifiShard.shards().values():
wifis = (self.session.query(shard.region, func.count())
.filter(shard.region.isnot(None))
.group_by(shard.region)).all()
for region, num in wifis:
if region not in stats:
stats[region] = default.copy()
stats[region]['wifi'] += int(num)
if not stats:
return
region_stats = dict(self.session.query(RegionStat.region,
RegionStat).all())
for region, values in stats.items():
if region in region_stats:
region_stats[region].gsm = values['gsm']
region_stats[region].wcdma = values['wcdma']
region_stats[region].lte = values['lte']
region_stats[region].wifi = values['wifi']
else:
self.session.add(RegionStat(
region=region,
gsm=values['gsm'],
wcdma=values['wcdma'],
lte=values['lte'],
wifi=values['wifi'],
))
obsolete_regions = list(set(region_stats.keys()) - set(stats.keys()))
if obsolete_regions:
(self.session.query(RegionStat)
.filter(RegionStat.region.in_(obsolete_regions))
).delete(synchronize_session=False)
def __call__(self):
cells = (self.session.query(CellArea.region, CellArea.radio,
func.sum(CellArea.num_cells)).filter(
CellArea.region.isnot(None)).group_by(
CellArea.region,
CellArea.radio)).all()
default = {'gsm': 0, 'wcdma': 0, 'lte': 0, 'wifi': 0}
stats = {}
for region, radio, num in cells:
if region not in stats:
stats[region] = default.copy()
stats[region][radio.name] = int(num)
for shard in WifiShard.shards().values():
wifis = (self.session.query(shard.region, func.count()).filter(
shard.region.isnot(None)).group_by(shard.region)).all()
for region, num in wifis:
if region not in stats:
stats[region] = default.copy()
stats[region]['wifi'] += int(num)
if not stats:
return
region_stats = dict(
self.session.query(RegionStat.region, RegionStat).all())
for region, values in stats.items():
if region in region_stats:
region_stats[region].gsm = values['gsm']
region_stats[region].wcdma = values['wcdma']
region_stats[region].lte = values['lte']
region_stats[region].wifi = values['wifi']
else:
self.session.add(
RegionStat(
region=region,
gsm=values['gsm'],
wcdma=values['wcdma'],
lte=values['lte'],
wifi=values['wifi'],
))
obsolete_regions = list(set(region_stats.keys()) - set(stats.keys()))
if obsolete_regions:
(self.session.query(RegionStat).filter(
RegionStat.region.in_(obsolete_regions))).delete(
synchronize_session=False)
def test_position_invalid(self, celery, session, stats):
self.add_reports(celery, 1, cell_factor=0, wifi_factor=1,
wifi_key='000000123456', lat=-90.1)
self.add_reports(celery, 1, cell_factor=0, wifi_factor=1,
wifi_key='000000234567')
self._update_all(session)
shard = WifiShard.shards()['0']
assert session.query(shard).count() == 1
stats.check(counter=[
('data.report.upload', 1, 2, ['key:test']),
('data.report.drop', 1, 1, ['key:test']),
('data.observation.insert', 1, 1, ['type:wifi']),
('data.observation.upload', 1, 1, ['type:wifi', 'key:test']),
])
def test_wifi(self):
reports = self.add_reports(cell_factor=0, wifi_factor=1)
self._update_all()
position = reports[0]['position']
wifi_data = reports[0]['wifiAccessPoints'][0]
mac = wifi_data['macAddress']
shard = WifiShard.shard_model(mac)
wifis = self.session.query(shard).all()
self.assertEqual(len(wifis), 1)
wifi = wifis[0]
self.assertEqual(wifi.lat, position['latitude'])
self.assertEqual(wifi.lon, position['longitude'])
self.assertEqual(wifi.mac, wifi_data['macAddress'])
self.assertEqual(wifi.samples, 1)
def test_wifi(self):
reports = self.add_reports(cell_factor=0, wifi_factor=1)
self._update_all()
position = reports[0]['position']
wifi_data = reports[0]['wifiAccessPoints'][0]
mac = wifi_data['macAddress']
shard = WifiShard.shard_model(mac)
wifis = self.session.query(shard).all()
self.assertEqual(len(wifis), 1)
wifi = wifis[0]
self.assertEqual(wifi.lat, position['latitude'])
self.assertEqual(wifi.lon, position['longitude'])
self.assertEqual(wifi.mac, wifi_data['macAddress'])
self.assertEqual(wifi.samples, 1)
def test_position_invalid(self):
self.add_reports(1, cell_factor=0, wifi_factor=1,
wifi_key='000000123456', lat=-90.1)
self.add_reports(1, cell_factor=0, wifi_factor=1,
wifi_key='000000234567')
self._update_all()
shard = WifiShard.shards()['0']
self.assertEqual(self.session.query(shard).count(), 1)
self.check_stats(counter=[
('data.report.upload', 1, 2, ['key:test']),
('data.report.drop', 1, 1, ['key:test']),
('data.observation.insert', 1, 1, ['type:wifi']),
('data.observation.upload', 1, 1, ['type:wifi', 'key:test']),
])
def test_position_invalid(self):
self.add_reports(1, cell_factor=0, wifi_factor=1,
wifi_key='000000123456', lat=-90.1)
self.add_reports(1, cell_factor=0, wifi_factor=1,
wifi_key='000000234567')
self._update_all()
shard = WifiShard.shards()['0']
self.assertEqual(self.session.query(shard).count(), 1)
self.check_stats(counter=[
('data.report.upload', 1, 2, ['key:test']),
('data.report.drop', 1, 1, ['reason:malformed', 'key:test']),
('data.observation.insert', 1, 1, ['type:wifi']),
('data.observation.upload', 1, 1, ['type:wifi', 'key:test']),
])
def _update_all(self, session, datamap_only=False):
ExportConfigFactory(name='internal', batch=0, schema='internal')
session.flush()
update_incoming.delay().get()
if datamap_only:
return
for shard_id in BlueShard.shards().keys():
update_blue.delay(shard_id=shard_id).get()
for shard_id in CellShard.shards().keys():
update_cell.delay(shard_id=shard_id).get()
for shard_id in WifiShard.shards().keys():
update_wifi.delay(shard_id=shard_id).get()
def _update_all(self, session, datamap_only=False):
ExportConfigFactory(name="internal", batch=0, schema="internal")
session.flush()
update_incoming.delay().get()
if datamap_only:
return
for shard_id in BlueShard.shards().keys():
update_blue.delay(shard_id=shard_id).get()
for shard_id in CellShard.shards().keys():
update_cell.delay(shard_id=shard_id).get()
for shard_id in WifiShard.shards().keys():
update_wifi.delay(shard_id=shard_id).get()
def test_wifi(self):
reports = self.add_reports(cell_factor=0, wifi_factor=1)
schedule_export_reports.delay().get()
for i in range(16):
update_wifi.delay(shard_id='%x' % i).get()
position = reports[0]['position']
wifi_data = reports[0]['wifiAccessPoints'][0]
mac = wifi_data['macAddress']
shard = WifiShard.shard_model(mac)
wifis = self.session.query(shard).all()
self.assertEqual(len(wifis), 1)
wifi = wifis[0]
self.assertEqual(wifi.lat, position['latitude'])
self.assertEqual(wifi.lon, position['longitude'])
self.assertEqual(wifi.mac, wifi_data['macAddress'])
self.assertEqual(wifi.samples, 1)
def test_wifi_duplicated(self):
self.add_reports(cell_factor=0, wifi_factor=1)
# duplicate the wifi entry inside the report
item = self._pop_item()
report = item['report']
wifi = report['wifiAccessPoints'][0]
mac = wifi['macAddress']
report['wifiAccessPoints'].append(wifi.copy())
report['wifiAccessPoints'].append(wifi.copy())
report['wifiAccessPoints'][1]['signalStrength'] += 2
report['wifiAccessPoints'][2]['signalStrength'] -= 2
self._push_item(item)
self._update_all()
shard = WifiShard.shard_model(mac)
wifis = self.session.query(shard).all()
self.assertEqual(len(wifis), 1)
self.assertEqual(wifis[0].samples, 1)
def test_wifi_duplicated(self, celery, session):
self.add_reports(celery, cell_factor=0, wifi_factor=1)
# duplicate the wifi entry inside the report
item = self._pop_item(celery)
report = item['report']
wifi = report['wifiAccessPoints'][0]
mac = wifi['macAddress']
report['wifiAccessPoints'].append(wifi.copy())
report['wifiAccessPoints'].append(wifi.copy())
report['wifiAccessPoints'][1]['signalStrength'] += 2
report['wifiAccessPoints'][2]['signalStrength'] -= 2
self._push_item(celery, item)
self._update_all(session)
shard = WifiShard.shard_model(mac)
wifis = session.query(shard).all()
assert len(wifis) == 1
assert wifis[0].samples == 1
def test_wifi_duplicated(self, celery, session):
self.add_reports(celery, cell_factor=0, wifi_factor=1)
# duplicate the wifi entry inside the report
item = self._pop_item(celery)
report = item["report"]
wifi = report["wifiAccessPoints"][0]
mac = wifi["macAddress"]
report["wifiAccessPoints"].append(wifi.copy())
report["wifiAccessPoints"].append(wifi.copy())
report["wifiAccessPoints"][1]["signalStrength"] += 2
report["wifiAccessPoints"][2]["signalStrength"] -= 2
self._push_item(celery, item)
self._update_all(session)
shard = WifiShard.shard_model(mac)
wifis = session.query(shard).all()
assert len(wifis) == 1
assert wifis[0].samples == 1
def test_wifi_duplicated(self):
self.add_reports(cell_factor=0, wifi_factor=1)
# duplicate the wifi entry inside the report
queue = self.celery_app.export_queues['queue_export_internal']
items = queue.dequeue(queue.queue_key())
report = items[0]['report']
wifi = report['wifiAccessPoints'][0]
mac = wifi['macAddress']
report['wifiAccessPoints'].append(wifi.copy())
report['wifiAccessPoints'].append(wifi.copy())
report['wifiAccessPoints'][1]['signalStrength'] += 2
report['wifiAccessPoints'][2]['signalStrength'] -= 2
queue.enqueue(items, queue.queue_key())
self._update_all()
shard = WifiShard.shard_model(mac)
wifis = self.session.query(shard).all()
self.assertEqual(len(wifis), 1)
self.assertEqual(wifis[0].samples, 1)
def test_wifi_duplicated(self):
self.add_reports(cell_factor=0, wifi_factor=1)
# duplicate the wifi entry inside the report
queue = self.celery_app.export_queues['internal']
items = queue.dequeue(queue.queue_key())
report = items[0]['report']
wifi = report['wifiAccessPoints'][0]
mac = wifi['macAddress']
report['wifiAccessPoints'].append(wifi.copy())
report['wifiAccessPoints'].append(wifi.copy())
report['wifiAccessPoints'][1]['signalStrength'] += 2
report['wifiAccessPoints'][2]['signalStrength'] -= 2
queue.enqueue(items, queue.queue_key())
self._update_all()
shard = WifiShard.shard_model(mac)
wifis = self.session.query(shard).all()
self.assertEqual(len(wifis), 1)
self.assertEqual(wifis[0].samples, 1)
def new_stations(self, name, station_keys):
if len(station_keys) == 0:
return 0
# assume all stations are unknown
unknown_keys = set(station_keys)
if name == 'wifi':
# there is only one combined table structure
shards = defaultdict(list)
for mac in unknown_keys:
shards[WifiShard.shard_model(mac)].append(mac)
for shard, macs in shards.items():
query = (self.session.query(shard.mac)
.filter(shard.mac.in_(macs)))
unknown_keys -= set([r.mac for r in query.all()])
elif name == 'cell':
# first check the station table, which is more likely to contain
# stations
station_iter = Cell.iterkeys(
self.session,
list(unknown_keys),
# only load the columns required for the hashkey
extra=lambda query: query.options(
load_only(*tuple(Cell._hashkey_cls._fields))))
# subtract all stations which are found in the station table
unknown_keys -= set([sta.hashkey() for sta in station_iter])
if len(unknown_keys) == 0: # pragma: no cover
return 0
# Only check the blocklist table for the still unknown keys.
# There is no need to check for the already found keys again.
block_iter = CellBlocklist.iterkeys(
self.session,
list(unknown_keys),
# only load the columns required for the hashkey
extra=lambda query: query.options(
load_only(*tuple(CellBlocklist._hashkey_cls._fields))))
# subtract all stations which are found in the blocklist table
unknown_keys -= set([block.hashkey() for block in block_iter])
return len(unknown_keys)
def test_new(self):
utcnow = util.utcnow()
obs = WifiObservationFactory.build()
self._queue_and_update([obs])
shard = WifiShard.shard_model(obs.mac)
wifis = self.session.query(shard).all()
self.assertEqual(len(wifis), 1)
wifi = wifis[0]
self.assertAlmostEqual(wifi.lat, obs.lat)
self.assertAlmostEqual(wifi.max_lat, obs.lat)
self.assertAlmostEqual(wifi.min_lat, obs.lat)
self.assertAlmostEqual(wifi.lon, obs.lon)
self.assertAlmostEqual(wifi.max_lon, obs.lon)
self.assertAlmostEqual(wifi.min_lon, obs.lon)
self.assertEqual(wifi.radius, 0)
self.assertEqual(wifi.region, 'GB')
self.assertEqual(wifi.samples, 1)
self.assertEqual(wifi.created.date(), utcnow.date())
self.assertEqual(wifi.modified.date(), utcnow.date())
self.assertEqual(wifi.block_first, None)
self.assertEqual(wifi.block_last, None)
self.assertEqual(wifi.block_count, None)
def configure_data(redis_client):
"""
Configure fixed set of data queues.
"""
data_queues = {
# update_incoming needs to be the exact same as in webapp.config
'update_incoming': DataQueue('update_incoming', redis_client,
batch=100, compress=True),
}
for key in ('update_cellarea', 'update_cellarea_ocid'):
data_queues[key] = DataQueue(key, redis_client, batch=100, json=False)
for shard_id in BlueShard.shards().keys():
key = 'update_blue_' + shard_id
data_queues[key] = DataQueue(key, redis_client, batch=500)
for shard_id in DataMap.shards().keys():
key = 'update_datamap_' + shard_id
data_queues[key] = DataQueue(key, redis_client, batch=500, json=False)
for shard_id in CellShard.shards().keys():
key = 'update_cell_' + shard_id
data_queues[key] = DataQueue(key, redis_client, batch=500)
for shard_id in WifiShard.shards().keys():
key = 'update_wifi_' + shard_id
data_queues[key] = DataQueue(key, redis_client, batch=500)
return data_queues
def process_reports(self, reports, userid=None):
malformed_reports = 0
positions = set()
observations = {'cell': [], 'wifi': []}
obs_count = {
'cell': {
'upload': 0,
'drop': 0
},
'wifi': {
'upload': 0,
'drop': 0
},
}
new_station_count = {'cell': 0, 'wifi': 0}
for report in reports:
cell, wifi, malformed_obs = self.process_report(report)
if cell:
observations['cell'].extend(cell)
obs_count['cell']['upload'] += len(cell)
if wifi:
observations['wifi'].extend(wifi)
obs_count['wifi']['upload'] += len(wifi)
if (cell or wifi):
positions.add((report['lat'], report['lon']))
else:
malformed_reports += 1
for name in ('cell', 'wifi'):
obs_count[name]['drop'] += malformed_obs[name]
# group by unique station key
for name in ('cell', 'wifi'):
station_keys = set()
for obs in observations[name]:
if name == 'cell':
station_keys.add(obs.cellid)
elif name == 'wifi':
station_keys.add(obs.mac)
# determine scores for stations
new_station_count[name] += self.new_stations(name, station_keys)
if observations['cell']:
sharded_obs = defaultdict(list)
for ob in observations['cell']:
shard_id = CellShard.shard_id(ob.cellid)
sharded_obs[shard_id].append(ob)
for shard_id, values in sharded_obs.items():
cell_queue = self.data_queues['update_cell_' + shard_id]
cell_queue.enqueue(list(values), pipe=self.pipe)
if observations['wifi']:
sharded_obs = defaultdict(list)
for ob in observations['wifi']:
shard_id = WifiShard.shard_id(ob.mac)
sharded_obs[shard_id].append(ob)
for shard_id, values in sharded_obs.items():
wifi_queue = self.data_queues['update_wifi_' + shard_id]
wifi_queue.enqueue(list(values), pipe=self.pipe)
self.process_datamap(positions)
self.process_score(userid, positions, new_station_count)
self.emit_stats(
len(reports),
malformed_reports,
obs_count,
)
def test_blocklist_moving_wifis(self):
now = util.utcnow()
obs = []
obs_factory = WifiObservationFactory
moving = set()
wifis = WifiShardFactory.create_batch(7)
wifis.append(WifiShardFactory.build())
# a wifi without an entry and disagreeing observations
wifi = wifis[-1]
obs.extend([
obs_factory(lat=wifi.lat, lon=wifi.lon, mac=wifi.mac),
obs_factory(lat=wifi.lat + 2.0, lon=wifi.lon, mac=wifi.mac),
])
moving.add(wifi.mac)
# a wifi with an entry but no prior position
wifi = wifis[0]
obs.extend([
obs_factory(lat=wifi.lat + 0.001,
lon=wifi.lon + 0.001, mac=wifi.mac),
obs_factory(lat=wifi.lat + 0.002,
lon=wifi.lon + 0.005, mac=wifi.mac),
obs_factory(lat=wifi.lat + 0.003,
lon=wifi.lon + 0.009, mac=wifi.mac),
])
wifi.lat = None
wifi.lon = None
wifi.samples = None
wifi.weight = None
# a wifi with a prior known position
wifi = wifis[1]
wifi.samples = 1
wifi.weight = 1.0
wifi.lat += 1.0
wifi.lon += 1.0
obs.extend([
obs_factory(lat=wifi.lat + 0.01,
lon=wifi.lon, mac=wifi.mac),
obs_factory(lat=wifi.lat + 0.07,
lon=wifi.lon, mac=wifi.mac),
])
moving.add(wifi.mac)
# a wifi with a very different prior position
wifi = wifis[2]
wifi.samples = 1
wifi.weight = 1.0
obs.extend([
obs_factory(lat=wifi.lat + 2.0,
lon=wifi.lon, mac=wifi.mac),
obs_factory(lat=wifi.lat + 2.002,
lon=wifi.lon, mac=wifi.mac),
])
moving.add(wifi.mac)
# an already blocked wifi
wifi = wifis[3]
wifi.block_last = now.date()
wifi.block_count = 1
obs.extend([
obs_factory(lat=wifi.lat,
lon=wifi.lon, mac=wifi.mac),
obs_factory(lat=wifi.lat + 0.1,
lon=wifi.lon, mac=wifi.mac),
])
moving.add(wifi.mac)
# a permanently blocked wifi
wifi = wifis[4]
wifi_lat, wifi_lon = (wifi.lat, wifi.lon)
wifi.block_last = now.date() - 2 * TEMPORARY_BLOCKLIST_DURATION
wifi.block_count = PERMANENT_BLOCKLIST_THRESHOLD
for col in ('lat', 'lon', 'max_lat', 'min_lat', 'max_lon', 'min_lon'):
setattr(wifi, col, None)
obs.extend([
obs_factory(lat=wifi_lat, lon=wifi_lon, mac=wifi.mac),
])
moving.add(wifi.mac)
# a no longer blocked wifi
wifi = wifis[5]
wifi_lat, wifi_lon = (wifi.lat, wifi.lon)
wifi.block_last = now.date() - 2 * TEMPORARY_BLOCKLIST_DURATION
wifi.block_count = 2
for col in ('lat', 'lon', 'max_lat', 'min_lat', 'max_lon', 'min_lon'):
setattr(wifi, col, None)
obs.extend([
obs_factory(lat=wifi_lat, lon=wifi_lon, mac=wifi.mac),
])
# a no longer blocked wifi with disagreeing observations
wifi = wifis[6]
wifi_lat, wifi_lon = (wifi.lat, wifi.lon)
wifi.block_last = now.date() - 2 * TEMPORARY_BLOCKLIST_DURATION
wifi.block_count = 2
for col in ('lat', 'lon', 'max_lat', 'min_lat', 'max_lon', 'min_lon'):
setattr(wifi, col, None)
obs.extend([
obs_factory(lat=wifi_lat, lon=wifi_lon, mac=wifi.mac),
obs_factory(lat=wifi_lat + 2.0, lon=wifi_lon, mac=wifi.mac),
])
moving.add(wifi.mac)
self.session.commit()
self._queue_and_update_wifi(obs)
shards = set()
for mac in moving:
shards.add(WifiShard.shard_model(mac))
blocks = []
for shard in shards:
for row in self.session.query(shard).all():
if row.blocked():
blocks.append(row)
self.assertEqual(set([b.mac for b in blocks]), moving)
def celerybeat_schedule(app_config):
"""Return the celery beat schedule as a dictionary."""
sections = app_config.sections()
schedule = {
# Monitoring
'monitor-queue-size': {
'task': 'ichnaea.data.tasks.monitor_queue_size',
'schedule': timedelta(seconds=60),
'options': {'expires': 57},
},
'monitor-api-users': {
'task': 'ichnaea.data.tasks.monitor_api_users',
'schedule': timedelta(seconds=600),
'options': {'expires': 570},
},
'monitor-api-key-limits': {
'task': 'ichnaea.data.tasks.monitor_api_key_limits',
'schedule': timedelta(seconds=600),
'options': {'expires': 570},
},
# Statistics
'update-statcounter': {
'task': 'ichnaea.data.tasks.update_statcounter',
'args': (1, ),
'schedule': crontab(minute=3),
'options': {'expires': 2700},
},
'update-statregion': {
'task': 'ichnaea.data.tasks.update_statregion',
'schedule': timedelta(seconds=3600 * 6),
'options': {'expires': 3600 * 5},
},
# Data Pipeline
'schedule-export-reports': {
'task': 'ichnaea.data.tasks.schedule_export_reports',
'schedule': timedelta(seconds=8),
'options': {'expires': 15},
},
'update-cellarea': {
'task': 'ichnaea.data.tasks.update_cellarea',
'schedule': timedelta(seconds=8),
'args': (100, ),
'options': {'expires': 15},
},
'update-cellarea-ocid': {
'task': 'ichnaea.data.tasks.update_cellarea_ocid',
'schedule': timedelta(seconds=9),
'args': (100, ),
'options': {'expires': 15},
},
'update-score': {
'task': 'ichnaea.data.tasks.update_score',
'args': (250, ),
'schedule': timedelta(seconds=9),
'options': {'expires': 10},
},
}
for shard_id in CellShard.shards().keys():
schedule.update({
'update-cell-' + shard_id: {
'task': 'ichnaea.data.tasks.update_cell',
'schedule': timedelta(seconds=7),
'args': (500, shard_id),
'options': {'expires': 10},
}
})
for shard_id in DataMap.shards().keys():
schedule.update({
'update-datamap-' + shard_id: {
'task': 'ichnaea.data.tasks.update_datamap',
'args': (500, shard_id),
'schedule': timedelta(seconds=14),
'options': {'expires': 20},
},
})
for shard_id in WifiShard.shards().keys():
schedule.update({
'update-wifi-' + shard_id: {
'task': 'ichnaea.data.tasks.update_wifi',
'schedule': timedelta(seconds=6),
'args': (500, shard_id),
'options': {'expires': 10},
}
})
if 'assets' in sections and app_config.get('assets', 'bucket', None):
# only configure tasks if target bucket is configured
schedule.update({
'cell-export-full': {
'task': 'ichnaea.data.tasks.cell_export_full',
'schedule': crontab(hour=0, minute=13),
'options': {'expires': 39600},
},
'cell-export-diff': {
'task': 'ichnaea.data.tasks.cell_export_diff',
'schedule': crontab(minute=3),
'options': {'expires': 2700},
},
})
if 'import:ocid' in sections:
schedule.update({
'monitor-ocid-import': {
'task': 'ichnaea.data.tasks.monitor_ocid_import',
'schedule': timedelta(seconds=600),
'options': {'expires': 570},
},
'cell-import-external': {
'task': 'ichnaea.data.tasks.cell_import_external',
'args': (True, ),
'schedule': crontab(minute=52),
'options': {'expires': 2700},
},
})
return schedule
def celerybeat_schedule(app_config):
"""Return the celery beat schedule as a dictionary."""
sections = app_config.sections()
schedule = {
# Monitoring
'monitor-queue-size': {
'task': 'ichnaea.data.tasks.monitor_queue_size',
'schedule': timedelta(seconds=60),
'options': {
'expires': 57
},
},
'monitor-api-users': {
'task': 'ichnaea.data.tasks.monitor_api_users',
'schedule': timedelta(seconds=600),
'options': {
'expires': 570
},
},
'monitor-api-key-limits': {
'task': 'ichnaea.data.tasks.monitor_api_key_limits',
'schedule': timedelta(seconds=600),
'options': {
'expires': 570
},
},
# Statistics
'update-statcounter': {
'task': 'ichnaea.data.tasks.update_statcounter',
'args': (1, ),
'schedule': crontab(minute=3),
'options': {
'expires': 2700
},
},
'update-statregion': {
'task': 'ichnaea.data.tasks.update_statregion',
'schedule': crontab(minute=5),
'options': {
'expires': 2700
},
},
# Data Pipeline
'schedule-export-reports': {
'task': 'ichnaea.data.tasks.schedule_export_reports',
'schedule': timedelta(seconds=8),
'options': {
'expires': 15
},
},
'update-cellarea': {
'task': 'ichnaea.data.tasks.update_cellarea',
'schedule': timedelta(seconds=8),
'args': (100, ),
'options': {
'expires': 15
},
},
'update-cellarea-ocid': {
'task': 'ichnaea.data.tasks.update_cellarea_ocid',
'schedule': timedelta(seconds=9),
'args': (100, ),
'options': {
'expires': 15
},
},
'update-score': {
'task': 'ichnaea.data.tasks.update_score',
'args': (250, ),
'schedule': timedelta(seconds=9),
'options': {
'expires': 10
},
},
}
for shard_id in CellShard.shards().keys():
schedule.update({
'update-cell-' + shard_id: {
'task': 'ichnaea.data.tasks.update_cell',
'schedule': timedelta(seconds=7),
'args': (500, shard_id),
'options': {
'expires': 10
},
}
})
for shard_id in DataMap.shards().keys():
schedule.update({
'update-datamap-' + shard_id: {
'task': 'ichnaea.data.tasks.update_datamap',
'args': (500, shard_id),
'schedule': timedelta(seconds=14),
'options': {
'expires': 20
},
},
})
for shard_id in WifiShard.shards().keys():
schedule.update({
'update-wifi-' + shard_id: {
'task': 'ichnaea.data.tasks.update_wifi',
'schedule': timedelta(seconds=6),
'args': (500, shard_id),
'options': {
'expires': 10
},
}
})
if 'assets' in sections and app_config.get('assets', 'bucket', None):
# only configure tasks if target bucket is configured
schedule.update({
'cell-export-full': {
'task': 'ichnaea.data.tasks.cell_export_full',
'schedule': crontab(hour=0, minute=13),
'options': {
'expires': 39600
},
},
'cell-export-diff': {
'task': 'ichnaea.data.tasks.cell_export_diff',
'schedule': crontab(minute=3),
'options': {
'expires': 2700
},
},
})
if 'import:ocid' in sections:
schedule.update({
'monitor-ocid-import': {
'task': 'ichnaea.data.tasks.monitor_ocid_import',
'schedule': timedelta(seconds=600),
'options': {
'expires': 570
},
},
'cell-import-external': {
'task': 'ichnaea.data.tasks.cell_import_external',
'args': (True, ),
'schedule': crontab(minute=52),
'options': {
'expires': 2700
},
},
})
return schedule
def process_reports(self, reports, userid=None):
malformed_reports = 0
positions = set()
observations = {'cell': [], 'wifi': []}
obs_count = {
'cell': {'upload': 0, 'drop': 0},
'wifi': {'upload': 0, 'drop': 0},
}
new_station_count = {'cell': 0, 'wifi': 0}
for report in reports:
cell, wifi, malformed_obs = self.process_report(report)
if cell:
observations['cell'].extend(cell)
obs_count['cell']['upload'] += len(cell)
if wifi:
observations['wifi'].extend(wifi)
obs_count['wifi']['upload'] += len(wifi)
if (cell or wifi):
positions.add((report['lat'], report['lon']))
else:
malformed_reports += 1
for name in ('cell', 'wifi'):
obs_count[name]['drop'] += malformed_obs[name]
# group by unique station key
for name in ('cell', 'wifi'):
station_keys = set()
for obs in observations[name]:
if name == 'cell':
station_keys.add(obs.cellid)
elif name == 'wifi':
station_keys.add(obs.mac)
# determine scores for stations
new_station_count[name] += self.new_stations(name, station_keys)
if observations['cell']:
sharded_obs = defaultdict(list)
for ob in observations['cell']:
shard_id = CellShard.shard_id(ob.cellid)
sharded_obs[shard_id].append(ob)
for shard_id, values in sharded_obs.items():
cell_queue = self.data_queues['update_cell_' + shard_id]
cell_queue.enqueue(list(values), pipe=self.pipe)
if observations['wifi']:
sharded_obs = defaultdict(list)
for ob in observations['wifi']:
shard_id = WifiShard.shard_id(ob.mac)
sharded_obs[shard_id].append(ob)
for shard_id, values in sharded_obs.items():
wifi_queue = self.data_queues['update_wifi_' + shard_id]
wifi_queue.enqueue(list(values), pipe=self.pipe)
self.process_datamap(positions)
self.process_score(userid, positions, new_station_count)
self.emit_stats(
len(reports),
malformed_reports,
obs_count,
)
def test_blocklist_moving_wifis(self):
now = util.utcnow()
obs = []
obs_factory = WifiObservationFactory
moving = set()
wifis = WifiShardFactory.create_batch(7)
wifis.append(WifiShardFactory.build())
# a wifi without an entry and disagreeing observations
wifi = wifis[-1]
obs.extend([
obs_factory(lat=wifi.lat, lon=wifi.lon, key=wifi.mac),
obs_factory(lat=wifi.lat + 2.0, lon=wifi.lon, key=wifi.mac),
])
moving.add(wifi.mac)
# a wifi with an entry but no prior position
wifi = wifis[0]
obs.extend([
obs_factory(lat=wifi.lat + 0.001,
lon=wifi.lon + 0.001,
key=wifi.mac),
obs_factory(lat=wifi.lat + 0.002,
lon=wifi.lon + 0.005,
key=wifi.mac),
obs_factory(lat=wifi.lat + 0.003,
lon=wifi.lon + 0.009,
key=wifi.mac),
])
wifi.lat = None
wifi.lon = None
wifi.samples = 0
# a wifi with a prior known position
wifi = wifis[1]
wifi.samples = 1
wifi.lat += 1.0
wifi.lon += 1.0
obs.extend([
obs_factory(lat=wifi.lat + 0.01, lon=wifi.lon, key=wifi.mac),
obs_factory(lat=wifi.lat + 0.07, lon=wifi.lon, key=wifi.mac),
])
moving.add(wifi.mac)
# a wifi with a very different prior position
wifi = wifis[2]
wifi.samples = 1
obs.extend([
obs_factory(lat=wifi.lat + 2.0, lon=wifi.lon, key=wifi.mac),
obs_factory(lat=wifi.lat + 2.002, lon=wifi.lon, key=wifi.mac),
])
moving.add(wifi.mac)
# an already blocked wifi
wifi = wifis[3]
wifi.block_last = now.date()
wifi.block_count = 1
obs.extend([
obs_factory(lat=wifi.lat, lon=wifi.lon, key=wifi.mac),
obs_factory(lat=wifi.lat + 0.1, lon=wifi.lon, key=wifi.mac),
])
moving.add(wifi.mac)
# a permanently blocked wifi
wifi = wifis[4]
wifi_lat, wifi_lon = (wifi.lat, wifi.lon)
wifi.block_last = now.date() - 2 * TEMPORARY_BLOCKLIST_DURATION
wifi.block_count = PERMANENT_BLOCKLIST_THRESHOLD
for col in ('lat', 'lon', 'max_lat', 'min_lat', 'max_lon', 'min_lon'):
setattr(wifi, col, None)
obs.extend([
obs_factory(lat=wifi_lat, lon=wifi_lon, key=wifi.mac),
])
moving.add(wifi.mac)
# a no longer blocked wifi
wifi = wifis[5]
wifi_lat, wifi_lon = (wifi.lat, wifi.lon)
wifi.block_last = now.date() - 2 * TEMPORARY_BLOCKLIST_DURATION
wifi.block_count = 2
for col in ('lat', 'lon', 'max_lat', 'min_lat', 'max_lon', 'min_lon'):
setattr(wifi, col, None)
obs.extend([
obs_factory(lat=wifi_lat, lon=wifi_lon, key=wifi.mac),
])
# a no longer blocked wifi with disagreeing observations
wifi = wifis[6]
wifi_lat, wifi_lon = (wifi.lat, wifi.lon)
wifi.block_last = now.date() - 2 * TEMPORARY_BLOCKLIST_DURATION
wifi.block_count = 2
for col in ('lat', 'lon', 'max_lat', 'min_lat', 'max_lon', 'min_lon'):
setattr(wifi, col, None)
obs.extend([
obs_factory(lat=wifi_lat, lon=wifi_lon, key=wifi.mac),
obs_factory(lat=wifi_lat + 2.0, lon=wifi_lon, key=wifi.mac),
])
moving.add(wifi.mac)
self.session.commit()
self._queue_and_update(obs)
shards = set()
for mac in moving:
shards.add(WifiShard.shard_model(mac))
blocks = []
for shard in shards:
for row in self.session.query(shard).all():
if row.blocked():
blocks.append(row)
self.assertEqual(set([b.mac for b in blocks]), moving)
def test_blocklist_moving_wifis(self):
now = util.utcnow()
obs = []
obs_factory = WifiObservationFactory
moving = set()
wifis = WifiShardFactory.create_batch(8)
wifis.append(WifiShardFactory.build())
# a wifi without an entry and disagreeing observations
wifi = wifis[-1]
obs.extend([
obs_factory(lat=wifi.lat, lon=wifi.lon, key=wifi.mac),
obs_factory(lat=wifi.lat + 2.0, lon=wifi.lon, key=wifi.mac),
])
moving.add(wifi.mac)
# a wifi with an entry but no prior position
wifi = wifis[0]
obs.extend([
obs_factory(lat=wifi.lat + 0.001,
lon=wifi.lon + 0.001, key=wifi.mac),
obs_factory(lat=wifi.lat + 0.002,
lon=wifi.lon + 0.005, key=wifi.mac),
obs_factory(lat=wifi.lat + 0.003,
lon=wifi.lon + 0.009, key=wifi.mac),
])
wifi.lat = None
wifi.lon = None
wifi.samples = 0
# a wifi with a prior known position
wifi = wifis[1]
wifi.samples = 1
wifi.lat += 1.0
wifi.lon += 1.0
obs.extend([
obs_factory(lat=wifi.lat + 0.01,
lon=wifi.lon, key=wifi.mac),
obs_factory(lat=wifi.lat + 0.07,
lon=wifi.lon, key=wifi.mac),
])
moving.add(wifi.mac)
# a wifi with a very different prior position
wifi = wifis[2]
wifi.samples = 1
obs.extend([
obs_factory(lat=wifi.lat + 2.0,
lon=wifi.lon, key=wifi.mac),
obs_factory(lat=wifi.lat + 2.002,
lon=wifi.lon, key=wifi.mac),
])
moving.add(wifi.mac)
# another wifi with a prior known position (and negative lat)
wifi = wifis[3]
wifi.samples = 1
wifi.lat *= -1.0
obs.extend([
obs_factory(lat=wifi.lat - 0.1,
lon=wifi.lon, key=wifi.mac),
obs_factory(lat=wifi.lat - 0.16,
lon=wifi.lon, key=wifi.mac),
])
moving.add(wifi.mac)
# an already blocked wifi
wifi = wifis[4]
wifi.block_last = now.date()
wifi.block_count = 1
obs.extend([
obs_factory(lat=wifi.lat,
lon=wifi.lon, key=wifi.mac),
obs_factory(lat=wifi.lat + 0.1,
lon=wifi.lon, key=wifi.mac),
])
moving.add(wifi.mac)
# a permanently blocked wifi
wifi = wifis[5]
wifi_lat, wifi_lon = (wifi.lat, wifi.lon)
wifi.block_last = now.date() - 2 * TEMPORARY_BLOCKLIST_DURATION
wifi.block_count = PERMANENT_BLOCKLIST_THRESHOLD
for col in ('lat', 'lon', 'max_lat', 'min_lat', 'max_lon', 'min_lon'):
setattr(wifi, col, None)
obs.extend([
obs_factory(lat=wifi_lat, lon=wifi_lon, key=wifi.mac),
])
moving.add(wifi.mac)
# a no longer blocked wifi
wifi = wifis[6]
wifi_lat, wifi_lon = (wifi.lat, wifi.lon)
wifi.block_last = now.date() - 2 * TEMPORARY_BLOCKLIST_DURATION
wifi.block_count = 2
for col in ('lat', 'lon', 'max_lat', 'min_lat', 'max_lon', 'min_lon'):
setattr(wifi, col, None)
obs.extend([
obs_factory(lat=wifi_lat, lon=wifi_lon, key=wifi.mac),
])
# a no longer blocked wifi with disagreeing observations
wifi = wifis[7]
wifi_lat, wifi_lon = (wifi.lat, wifi.lon)
wifi.block_last = now.date() - 2 * TEMPORARY_BLOCKLIST_DURATION
wifi.block_count = 2
for col in ('lat', 'lon', 'max_lat', 'min_lat', 'max_lon', 'min_lon'):
setattr(wifi, col, None)
obs.extend([
obs_factory(lat=wifi_lat, lon=wifi_lon, key=wifi.mac),
obs_factory(lat=wifi_lat + 2.0, lon=wifi_lon, key=wifi.mac),
])
moving.add(wifi.mac)
self.data_queue.enqueue(obs)
self.session.commit()
update_wifi.delay().get()
shards = set()
for mac in moving:
shards.add(WifiShard.shard_model(mac))
blocks = []
for shard in shards:
for row in self.session.query(shard).all():
if row.blocked():
blocks.append(row)
self.assertEqual(set([b.mac for b in blocks]), moving)
self.check_stats(counter=[
('data.observation.drop', 1, 3,
['type:wifi', 'reason:blocklisted']),
('data.station.blocklist', 1, 5,
['type:wifi', 'action:add', 'reason:moving']),
])