def test_should_add_summary_to_logger_instance(summary_patched):
metrics_logger = MetricsLogger("some_id")
metric_name = "some_summary"
metrics_logger.add_summary(metric_name)
assert metric_name in metrics_logger._metrics
def test_should_add_counter_to_logger_instance(counter_patched):
metrics_logger = MetricsLogger("some_id")
metric_name = "some_counter"
metrics_logger.add_counter(metric_name)
assert metric_name in metrics_logger._metrics
def test_should_add_gauge_to_logger_instance(gauge_patched):
metrics_logger = MetricsLogger("some_id")
metric_name = "some_gauge"
metrics_logger.add_gauge(metric_name)
assert metric_name in metrics_logger._metrics
def test_should_add_info_to_logger_instance(info_patched):
metrics_logger = MetricsLogger("some_id")
metric_name = "some_info"
metrics_logger.add_info(metric_name)
assert metric_name in metrics_logger._metrics
def test_should_add_histogram_to_logger_instance(histogram_patched):
metrics_logger = MetricsLogger("some_id")
metric_name = "some_histogram"
metrics_logger.add_histogram(metric_name)
assert metric_name in metrics_logger._metrics
def test_should_update_summary_value(summary_patched):
metrics_logger = MetricsLogger("some_id")
metric_name = "some_summary"
metrics_logger.add_summary(metric_name)
metrics_logger[metric_name].observe(2)
summary_patched().labels("some_id", None).observe.assert_called_once_with(2)
def test_should_update_counter_value(counter_patched):
metrics_logger = MetricsLogger("some_id")
metric_name = "some_counter"
metrics_logger.add_counter(metric_name)
metrics_logger[metric_name].inc(2)
counter_patched().labels("some_id", None).inc.assert_called_once_with(2)
def test_should_update_gauge_value(gauge_patched):
metrics_logger = MetricsLogger("some_id")
metric_name = "some_gauge"
metrics_logger.add_gauge(metric_name)
metrics_logger[metric_name].set(2)
gauge_patched().labels("some_id", None).set.assert_called_once_with(2)
def test_should_update_info_value(info_patched):
metrics_logger = MetricsLogger("some_id")
metric_name = "some_info"
metrics_logger.add_info(metric_name)
metrics_logger[metric_name].info(2)
info_patched().labels("some_id", None).info.assert_called_once_with(2)
def test_should_update_histogram_value(histogram_patched):
metrics_logger = MetricsLogger("some_id")
metric_name = "some_histogram"
metrics_logger.add_histogram(metric_name)
metrics_logger[metric_name].observe(2)
histogram_patched().labels("some_id", None).observe.assert_called_once_with(2)
def test_should_use_push_to_gateway_to_log_metrics(push_to_gateway_patched):
registry = CollectorRegistry()
metrics_logger = MetricsLogger(
"some_id",
registry=registry,
)
metrics_logger.push_metrics()
push_to_gateway_patched.assert_called_once_with(
gateway=get_metric_pushgateway(),
job="some_id",
registry=registry,
grouping_key={
'label_metrics_experiment_handle': 'some_id',
'step': None},
timeout=30,
)
def test_should_create_metrics_logger_instance_with_required_parameters(
get_workload_id_patched, get_metric_pushgateway_patched):
metrics_logger = MetricsLogger()
assert metrics_logger.id == "some_id"
assert metrics_logger._grouping_key == {
'label_metrics_experiment_handle': 'some_id',
'step': None}
assert metrics_logger._push_gateway == "some_gateway"
assert isinstance(metrics_logger._registry, CollectorRegistry)
def test_should_create_metrics_logger_instance_with_all_parameters(
get_workload_id_patched, get_metric_pushgateway_patched):
registry = CollectorRegistry()
push_gateway = "some.url"
metrics_logger = MetricsLogger(
"some_id",
registry=registry,
push_gateway=push_gateway,
)
assert metrics_logger.id == "some_id"
assert metrics_logger._push_gateway == push_gateway
assert metrics_logger._registry is registry
get_workload_id_patched.assert_not_called()
get_metric_pushgateway_patched.assert_not_called()
import json
from detectron2.engine import DefaultPredictor
from detectron2.config import get_cfg
from detectron2.utils.visualizer import Visualizer
from detectron2.utils.visualizer import ColorMode
from detectron2 import model_zoo
from detectron2.data import MetadataCatalog, DatasetCatalog
from detectron2.modeling import build_model
import torch
import numpy as np
from PIL import Image
from gradient_utils.metrics import MetricsLogger
try:
logger = MetricsLogger()
logger.add_counter("inference_count")
except:
logger = None
class Detector:
def __init__(self):
# set model and test set
self.model = 'mask_rcnn_R_50_FPN_3x.yaml'
# obtain detectron2's default config
self.cfg = get_cfg()