float argument must be a string or a number (#762)

* Use saft_float to parse * Bump version * Add tests
MetaphorData · Jan 23, 2024 · 1ccaa9c · 1ccaa9c
1 parent b630645
commit 1ccaa9c
Show file tree

Hide file tree

Showing 3 changed files with 19 additions and 4 deletions.
diff --git a/metaphor/azure_data_factory/extractor.py b/metaphor/azure_data_factory/extractor.py
@@ -29,7 +29,7 @@
 )
 from metaphor.common.event_util import ENTITY_TYPES
 from metaphor.common.logger import get_logger, json_dump_to_debug_file
-from metaphor.common.utils import unique_list
+from metaphor.common.utils import safe_float, unique_list
 from metaphor.models.crawler_run_metadata import Platform
 from metaphor.models.metadata_change_event import (
     ActivityDependency,
@@ -539,7 +539,7 @@ def _extract_pipeline_metadata(
                 pipeline_name, df_client, factory
             )
             if last_pipeline_run:
-                last_duration_in_ms = float(last_pipeline_run.duration_in_ms)  # type: ignore
+                last_duration_in_ms = safe_float(last_pipeline_run.duration_in_ms)  # type: ignore
                 invoked_by: DfModels.PipelineRunInvokedBy = last_pipeline_run.invoked_by  # type: ignore
                 last_invoke_type = invoked_by.invoked_by_type
                 last_run_message = last_pipeline_run.message

diff --git a/pyproject.toml b/pyproject.toml
@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "metaphor-connectors"
-version = "0.13.109"
+version = "0.13.110"
 license = "Apache-2.0"
 description = "A collection of Python-based 'connectors' that extract metadata from various sources to ingest into the Metaphor app."
 authors = ["Metaphor <[email protected]>"]

diff --git a/tests/azure_data_factory/test_extractor.py b/tests/azure_data_factory/test_extractor.py
@@ -203,7 +203,22 @@ def mock_list_pipelines(factory_name, resource_group_name):
     mock_client.pipelines.list_by_factory.side_effect = mock_list_pipelines
 
     def mock_get_pipeline_runs(factory_name, resource_group_name, filter_parameters):
-        return DfModels.PipelineRunsQueryResponse(value=[])
+        return DfModels.PipelineRunsQueryResponse(
+            value=[
+                DfModels.PipelineRun.deserialize(
+                    {
+                        "runId": "run-id",
+                        "runGroupId": "run-group-id",
+                        "isLatest": True,
+                        "pipelineName": "pipeline-1",
+                        "invokedBy": {
+                            "name": "foo",
+                            "id": "[email protected]",
+                        },
+                    }
+                )
+            ]
+        )
 
     mock_client.pipeline_runs.query_by_factory.side_effect = mock_get_pipeline_runs
     mock_client.pipeline_runs._serialize.body.side_effect = lambda _1, _2: {