def _get_aggregates(self, ms_id):
keyset = set()
for key in request.args.getlist('keys'):
if key not in MetricsKey.to_dict():
raise exceptions.BadRequest(
'unknown metrics key {}'.format(key))
keyset.add(MetricsKey.lookup_by_name(key))
ms = MetricSetDao().get(ms_id)
if not ms:
raise exceptions.NotFound(
'metric set "{}" not found'.format(ms_id))
if not keyset.issubset(METRIC_SET_KEYS[ms.metricSetType]):
raise exceptions.BadRequest(
'unexpected metric keys for metric set of type {}: {}'.format(
ms.metricSetType.name, [
k.name
for k in keyset - METRIC_SET_KEYS[ms.metricSetType]
]))
aggs = AggregateMetricsDao().get_all_for_metric_set(ms_id)
if keyset:
aggs = [agg for agg in aggs if agg.metricsKey in keyset]
return {'metrics': AggregateMetricsDao.to_client_json(aggs)}
def recalculate_public_metrics():
logging.info('generating public metrics')
aggs = PublicMetricsExport.export(LIVE_METRIC_SET_ID)
client_aggs = AggregateMetricsDao.to_client_json(aggs)
# summing all counts for one metric yields a total qualified participant count
participant_count = 0
if len(client_aggs) > 0:
participant_count = sum([a['count'] for a in client_aggs[0]['values']])
logging.info('persisted public metrics: {} aggregations over '
'{} participants'.format(len(client_aggs), participant_count))
# Same format returned by the metric sets API.
return json.dumps({'metrics': client_aggs})