Handle null timestamps

If either started_at or completed_at timestamp is null, use the current
timestamp as the event time and set duration to zero.

Signed-off-by: Michi Mutsuzaki <michi@isovalent.com>

.github/workflows/main.yml

@@ -15,8 +15,15 @@ jobs:
       # you must check out the repository
       - name: Checkout
         uses: actions/checkout@v3
+      - uses: actions/setup-python@v4
+        with:
+          python-version: '3.10.7'
       - name: Run formatter
         uses: psf/black@stable
+      - name: Run unit tests
+        run: |
+          pip install -r requirements.txt
+          python -m unittest test_push.py
       - name: Step 1
         run: echo step 1
       - name: Step 2

push.py

@@ -36,7 +36,12 @@ def workflow_run_to_stream(workflow_run, branch):
 def job_to_stream(workflow_run, job, branch):
     duration = 0
     completed_at = datetime.datetime.now(datetime.timezone.utc)
-    if "started_at" in job and "completed_at" in job:
+    if (
+        "started_at" in job
+        and job["started_at"] is not None
+        and "completed_at" in job
+        and job["completed_at"] is not None
+    ):
         started_at = dateutil.parser.isoparse(job["started_at"])
         completed_at = dateutil.parser.isoparse(job["completed_at"])
         duration = int(datetime.timedelta.total_seconds(completed_at - started_at))
@@ -63,7 +68,12 @@ def job_to_stream(workflow_run, job, branch):
 def step_to_stream(workflow_run, job, step, branch):
     duration = 0
     completed_at = datetime.datetime.now(datetime.timezone.utc)
-    if "started_at" in step and "completed_at" in step:
+    if (
+        "started_at" in step
+        and step["started_at"] is not None
+        and "completed_at" in step
+        and step["completed_at"] is not None
+    ):
         started_at = dateutil.parser.isoparse(step["started_at"])
         completed_at = dateutil.parser.isoparse(step["completed_at"])
         duration = int(datetime.timedelta.total_seconds(completed_at - started_at))
@@ -87,6 +97,15 @@ def step_to_stream(workflow_run, job, step, branch):
     }
 
 
+def jobs_to_stream(workflow_run, jobs, branch):
+    streams = []
+    for job in jobs["jobs"]:
+        streams.append(job_to_stream(workflow_run, job, branch))
+        for step in job["steps"]:
+            streams.append(step_to_stream(workflow_run, job, step, branch))
+    return streams
+
+
 def main(loki_endpoint, loki_username, loki_password, workflow_run_url, github_token):
     # Get workflow run.
     headers = {
@@ -113,12 +132,7 @@ def main(loki_endpoint, loki_username, loki_password, workflow_run_url, github_t
     if "pull_requests" in workflow_run and workflow_run["pull_requests"]:
         branch = workflow_run["pull_requests"][0]["base"]["ref"]
 
-    streams = []
-    for job in jobs["jobs"]:
-        streams.append(job_to_stream(workflow_run, job, branch))
-        for step in job["steps"]:
-            streams.append(step_to_stream(workflow_run, job, step, branch))
-
+    streams = jobs_to_stream(workflow_run, jobs, branch)
     streams.append(workflow_run_to_stream(workflow_run, branch))
 
     body = {"streams": streams}

test_push.py

@@ -0,0 +1,14 @@
+import json
+from unittest import TestCase
+from push import jobs_to_stream
+
+
+class Test(TestCase):
+    def test_jobs_to_stream(self):
+        with open("testdata/run.json") as run_json, open(
+            "testdata/jobs.json"
+        ) as jobs_json:
+            run = json.load(run_json)
+            jobs = json.load(jobs_json)
+            # Make sure it can handle null timestamps
+            jobs_to_stream(run, jobs, "main")