Add conversion to TimeSeries methods

Author: shovnik

Diff for: exporter/opentelemetry-exporter-prometheus-remote-write/CHANGELOG.md

@@ -7,3 +7,5 @@
   ((#180)[https://github.com/open-telemetry/opentelemetry-python-contrib/pull/180])
 - Add Exporter constructor validation methods
   ((#206)[https://github.com/open-telemetry/opentelemetry-python-contrib/pull/206])
+- Add conversion to TimeSeries methods
+  ((#207)[https://github.com/open-telemetry/opentelemetry-python-contrib/pull/207])

Diff for: exporter/opentelemetry-exporter-prometheus-remote-write/src/opentelemetry/exporter/prometheus_remote_write/__init__.py

@@ -12,8 +12,12 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+import re
 from typing import Dict, Sequence
 
+from opentelemetry.exporter.prometheus_remote_write.gen.remote_pb2 import (
+    WriteRequest,
+)
 from opentelemetry.exporter.prometheus_remote_write.gen.types_pb2 import (
     Label,
     Sample,
@@ -24,6 +28,13 @@
     MetricsExporter,
     MetricsExportResult,
 )
+from opentelemetry.sdk.metrics.export.aggregate import (
+    HistogramAggregator,
+    LastValueAggregator,
+    MinMaxSumCountAggregator,
+    SumAggregator,
+    ValueObserverAggregator,
+)
 
 
 class PrometheusRemoteWriteMetricsExporter(MetricsExporter):
@@ -145,31 +156,88 @@ def shutdown(self) -> None:
     def convert_to_timeseries(
         self, export_records: Sequence[ExportRecord]
     ) -> Sequence[TimeSeries]:
-        raise NotImplementedError()
+        converter_map = {
+            MinMaxSumCountAggregator: self.convert_from_min_max_sum_count,
+            SumAggregator: self.convert_from_sum,
+            HistogramAggregator: self.convert_from_histogram,
+            LastValueAggregator: self.convert_from_last_value,
+            ValueObserverAggregator: self.convert_from_last_value,
+        }
+        timeseries = []
+        for export_record in export_records:
+            aggregator_type = type(export_record.aggregator)
+            converter = converter_map.get(aggregator_type)
+            if not converter:
+                raise ValueError(
+                    str(aggregator_type) + " conversion is not supported"
+                )
+            timeseries.extend(converter(export_record))
+        return timeseries
 
     def convert_from_sum(self, sum_record: ExportRecord) -> TimeSeries:
-        raise NotImplementedError()
+        name = sum_record.instrument.name
+        value = sum_record.aggregator.checkpoint
+        return [self.create_timeseries(sum_record, name, value)]
 
     def convert_from_min_max_sum_count(
         self, min_max_sum_count_record: ExportRecord
     ) -> TimeSeries:
-        raise NotImplementedError()
+        timeseries = []
+        agg_types = ["min", "max", "sum", "count"]
+        for agg_type in agg_types:
+            name = min_max_sum_count_record.instrument.name + "_" + agg_type
+            value = getattr(
+                min_max_sum_count_record.aggregator.checkpoint, agg_type
+            )
+            timeseries.append(
+                self.create_timeseries(min_max_sum_count_record, name, value)
+            )
+        return timeseries
 
     def convert_from_histogram(
         self, histogram_record: ExportRecord
     ) -> TimeSeries:
-        raise NotImplementedError()
+        count = 0
+        timeseries = []
+        for bound in histogram_record.aggregator.checkpoint.keys():
+            bb = "+Inf" if bound == float("inf") else str(bound)
+            name = (
+                histogram_record.instrument.name + '_bucket{le="' + bb + '"}'
+            )
+            value = histogram_record.aggregator.checkpoint[bound]
+            timeseries.append(
+                self.create_timeseries(histogram_record, name, value)
+            )
+            count += value
+        name = histogram_record.instrument.name + "_count"
+        timeseries.append(
+            self.create_timeseries(histogram_record, name, float(count))
+        )
+        return timeseries
 
     def convert_from_last_value(
         self, last_value_record: ExportRecord
     ) -> TimeSeries:
-        raise NotImplementedError()
+        name = last_value_record.instrument.name
+        value = last_value_record.aggregator.checkpoint
+        return [self.create_timeseries(last_value_record, name, value)]
 
     def convert_from_value_observer(
         self, value_observer_record: ExportRecord
     ) -> TimeSeries:
-        raise NotImplementedError()
-
+        timeseries = []
+        agg_types = ["min", "max", "sum", "count", "last"]
+        for agg_type in agg_types:
+            name = value_observer_record.instrument.name + "_" + agg_type
+            value = getattr(
+                value_observer_record.aggregator.checkpoint, agg_type
+            )
+            timeseries.append(
+                self.create_timeseries(value_observer_record, name, value)
+            )
+        return timeseries
+
+    # TODO: Implement convert from quantile once supported by SDK for Prometheus Summaries
     def convert_from_quantile(
         self, summary_record: ExportRecord
     ) -> TimeSeries:
@@ -179,13 +247,37 @@ def convert_from_quantile(
     def create_timeseries(
         self, export_record: ExportRecord, name, value: float
     ) -> TimeSeries:
-        raise NotImplementedError()
+        timeseries = TimeSeries()
+        # Add name label, record labels and resource labels
+        timeseries.labels.append(self.create_label("__name__", name))
+        resource_attributes = export_record.resource.attributes
+        for label_name, label_value in resource_attributes.items():
+            timeseries.labels.append(
+                self.create_label(label_name, label_value)
+            )
+        for label in export_record.labels:
+            if label[0] not in resource_attributes.keys():
+                timeseries.labels.append(self.create_label(label[0], label[1]))
+        # Add sample
+        timeseries.samples.append(
+            self.create_sample(
+                export_record.aggregator.last_update_timestamp, value
+            )
+        )
+        return timeseries
 
     def create_sample(self, timestamp: int, value: float) -> Sample:
-        raise NotImplementedError()
+        sample = Sample()
+        sample.timestamp = int(timestamp / 1000000)
+        sample.value = value
+        return sample
 
     def create_label(self, name: str, value: str) -> Label:
-        raise NotImplementedError()
+        label = Label()
+        # Label name must contain only alphanumeric characters and underscores
+        label.name = re.sub("[^0-9a-zA-Z_]+", "_", name)
+        label.value = value
+        return label
 
     def build_message(self, timeseries: Sequence[TimeSeries]) -> bytes:
         raise NotImplementedError()

Diff for: exporter/opentelemetry-exporter-prometheus-remote-write/tests/test_prometheus_remote_write_exporter.py

@@ -13,10 +13,25 @@
 # limitations under the License.
 
 import unittest
+from unittest import mock
 
 from opentelemetry.exporter.prometheus_remote_write import (
     PrometheusRemoteWriteMetricsExporter,
 )
+from opentelemetry.exporter.prometheus_remote_write.gen.types_pb2 import (
+    TimeSeries,
+)
+from opentelemetry.sdk.metrics import Counter
+from opentelemetry.sdk.metrics.export import ExportRecord, MetricsExportResult
+from opentelemetry.sdk.metrics.export.aggregate import (
+    HistogramAggregator,
+    LastValueAggregator,
+    MinMaxSumCountAggregator,
+    SumAggregator,
+    ValueObserverAggregator,
+)
+from opentelemetry.sdk.resources import Resource
+from opentelemetry.sdk.util import get_dict_as_key
 
 
 class TestValidation(unittest.TestCase):
@@ -111,35 +126,151 @@ def test_invalid_tls_config_key_only_param(self):
 class TestConversion(unittest.TestCase):
     # Initializes test data that is reused across tests
     def setUp(self):
-        pass
+        self._test_metric = Counter(
+            "testname", "testdesc", "testunit", int, None
+        )
+        self._exporter = PrometheusRemoteWriteMetricsExporter(
+            endpoint="/prom/test_endpoint"
+        )
+
+        def generate_record(aggregator_type):
+            return ExportRecord(
+                self._test_metric, None, aggregator_type(), Resource({}),
+            )
+
+        self._generate_record = generate_record
+
+        def converter_method(record, name, value):
+            return (type(record.aggregator), name, value)
+
+        self._converter_mock = mock.MagicMock(return_value=converter_method)
 
     # Ensures conversion to timeseries function works with valid aggregation types
     def test_valid_convert_to_timeseries(self):
-        pass
+        timeseries_mock_method = mock.Mock(return_value=["test_value"])
+        self._exporter.convert_from_sum = timeseries_mock_method
+        self._exporter.convert_from_min_max_sum_count = timeseries_mock_method
+        self._exporter.convert_from_histogram = timeseries_mock_method
+        self._exporter.convert_from_last_value = timeseries_mock_method
+        self._exporter.convert_from_value_observer = timeseries_mock_method
+        test_records = [
+            self._generate_record(SumAggregator),
+            self._generate_record(MinMaxSumCountAggregator),
+            self._generate_record(HistogramAggregator),
+            self._generate_record(LastValueAggregator),
+            self._generate_record(ValueObserverAggregator),
+        ]
+        data = self._exporter.convert_to_timeseries(test_records)
+        self.assertEqual(len(data), 5)
+        for timeseries in data:
+            self.assertEqual(timeseries, "test_value")
+
+        no_type_records = [self._generate_record(lambda: None)]
+        with self.assertRaises(ValueError):
+            self._exporter.convert_to_timeseries(no_type_records)
 
     # Ensures conversion to timeseries fails for unsupported aggregation types
     def test_invalid_convert_to_timeseries(self):
-        pass
+        no_type_records = [self._generate_record(lambda: None)]
+        with self.assertRaises(ValueError):
+            self._exporter.convert_to_timeseries(no_type_records)
 
     # Ensures sum aggregator is correctly converted to timeseries
     def test_convert_from_sum(self):
-        pass
+        sum_record = self._generate_record(SumAggregator)
+        sum_record.aggregator.update(3)
+        sum_record.aggregator.update(2)
+        sum_record.aggregator.take_checkpoint()
+
+        self._exporter.create_timeseries = self._converter_mock()
+        timeseries = self._exporter.convert_from_sum(sum_record)
+        self.assertEqual(timeseries[0], (SumAggregator, "testname", 5))
 
     # Ensures sum min_max_count aggregator is correctly converted to timeseries
     def test_convert_from_min_max_sum_count(self):
-        pass
+        min_max_sum_count_record = self._generate_record(
+            MinMaxSumCountAggregator
+        )
+        min_max_sum_count_record.aggregator.update(5)
+        min_max_sum_count_record.aggregator.update(1)
+        min_max_sum_count_record.aggregator.take_checkpoint()
+
+        self._exporter.create_timeseries = self._converter_mock()
+        timeseries = self._exporter.convert_from_min_max_sum_count(
+            min_max_sum_count_record
+        )
+        self.assertEqual(
+            timeseries[0], (MinMaxSumCountAggregator, "testname_min", 1)
+        )
+        self.assertEqual(
+            timeseries[1], (MinMaxSumCountAggregator, "testname_max", 5)
+        )
+        self.assertEqual(
+            timeseries[2], (MinMaxSumCountAggregator, "testname_sum", 6)
+        )
+        self.assertEqual(
+            timeseries[3], (MinMaxSumCountAggregator, "testname_count", 2)
+        )
 
     # Ensures histogram aggregator is correctly converted to timeseries
     def test_convert_from_histogram(self):
-        pass
+        histogram_record = self._generate_record(HistogramAggregator)
+        histogram_record.aggregator.update(5)
+        histogram_record.aggregator.update(2)
+        histogram_record.aggregator.update(-1)
+        histogram_record.aggregator.take_checkpoint()
+
+        self._exporter.create_timeseries = self._converter_mock()
+        timeseries = self._exporter.convert_from_histogram(histogram_record)
+        self.assertEqual(
+            timeseries[0], (HistogramAggregator, 'testname_bucket{le="0"}', 1)
+        )
+        self.assertEqual(
+            timeseries[1],
+            (HistogramAggregator, 'testname_bucket{le="+Inf"}', 2),
+        )
+        self.assertEqual(
+            timeseries[2], (HistogramAggregator, "testname_count", 3)
+        )
 
     # Ensures last value aggregator is correctly converted to timeseries
     def test_convert_from_last_value(self):
-        pass
+        last_value_record = self._generate_record(LastValueAggregator)
+        last_value_record.aggregator.update(1)
+        last_value_record.aggregator.update(5)
+        last_value_record.aggregator.take_checkpoint()
+
+        self._exporter.create_timeseries = self._converter_mock()
+        timeseries = self._exporter.convert_from_last_value(last_value_record)
+        self.assertEqual(timeseries[0], (LastValueAggregator, "testname", 5))
 
     # Ensures value observer aggregator is correctly converted to timeseries
     def test_convert_from_value_observer(self):
-        pass
+        value_observer_record = self._generate_record(ValueObserverAggregator)
+        value_observer_record.aggregator.update(5)
+        value_observer_record.aggregator.update(1)
+        value_observer_record.aggregator.update(2)
+        value_observer_record.aggregator.take_checkpoint()
+
+        self._exporter.create_timeseries = self._converter_mock()
+        timeseries = self._exporter.convert_from_value_observer(
+            value_observer_record
+        )
+        self.assertEqual(
+            timeseries[0], (ValueObserverAggregator, "testname_min", 1)
+        )
+        self.assertEqual(
+            timeseries[1], (ValueObserverAggregator, "testname_max", 5)
+        )
+        self.assertEqual(
+            timeseries[2], (ValueObserverAggregator, "testname_sum", 8)
+        )
+        self.assertEqual(
+            timeseries[3], (ValueObserverAggregator, "testname_count", 3)
+        )
+        self.assertEqual(
+            timeseries[4], (ValueObserverAggregator, "testname_last", 2)
+        )
 
     # Ensures quantile aggregator is correctly converted to timeseries
     # TODO: Add test once method is implemented
@@ -148,7 +279,34 @@ def test_convert_from_quantile(self):
 
     # Ensures timeseries produced contains appropriate sample and labels
     def test_create_timeseries(self):
-        pass
+        sum_aggregator = SumAggregator()
+        sum_aggregator.update(5)
+        sum_aggregator.take_checkpoint()
+        sum_aggregator.last_update_timestamp = 10
+        export_record = ExportRecord(
+            self._test_metric,
+            get_dict_as_key({"record_name": "record_value"}),
+            sum_aggregator,
+            Resource({"resource_name": "resource_value"}),
+        )
+
+        expected_timeseries = TimeSeries()
+        expected_timeseries.labels.append(
+            self._exporter.create_label("__name__", "testname")
+        )
+        expected_timeseries.labels.append(
+            self._exporter.create_label("resource_name", "resource_value")
+        )
+        expected_timeseries.labels.append(
+            self._exporter.create_label("record_name", "record_value")
+        )
+        expected_timeseries.samples.append(
+            self._exporter.create_sample(10, 5.0),
+        )
+        timeseries = self._exporter.create_timeseries(
+            export_record, "testname", 5.0,
+        )
+        self.assertEqual(timeseries, expected_timeseries)
 
 
 class TestExport(unittest.TestCase):