fix(fal_client): handle missing metrics

.github/workflows/tests-fal-client.yml

@@ -0,0 +1,28 @@
+name: Run fal-client tests
+
+on:
+  push:
+    branches:
+    - main
+  pull_request:
+  workflow_dispatch:
+
+jobs:
+  tests:
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v3
+      with:
+        fetch-depth: 0
+    - uses: actions/setup-python@v4
+      with:
+        python-version: "3.12"
+    - name: Install dependencies
+      run: |
+        pip install --upgrade pip wheel
+        pip install -e 'projects/fal_client[test]'
+    - name: Run tests
+      env:
+        FAL_KEY: ${{ secrets.FAL_KEY_PROD }}
+      run: |
+        pytest projects/fal_client/tests

projects/fal_client/pyproject.toml

@@ -23,6 +23,7 @@ dependencies = [
 test = [
     "pytest",
     "pytest-asyncio",
+    "pillow",
 ]
 dev = [
     "fal_client[test]",

projects/fal_client/src/fal_client/client.py

@@ -72,7 +72,9 @@ def _parse_status(self, data: AnyJSON) -> Status:
         elif data["status"] == "IN_PROGRESS":
             return InProgress(logs=data["logs"])
         elif data["status"] == "COMPLETED":
-            return Completed(logs=data["logs"], metrics=data["metrics"])
+            # NOTE: legacy apps might not return metrics
+            metrics = data.get("metrics", {})
+            return Completed(logs=data["logs"], metrics=metrics)
         else:
             raise ValueError(f"Unknown status: {data['status']}")
 

projects/fal_client/tests/unit/test_client.py

@@ -0,0 +1,43 @@
+import pytest
+
+from fal_client.client import Queued, InProgress, Completed, _BaseRequestHandle
+
+
+@pytest.mark.parametrize(
+    "data, result, raised",
+    [
+        (
+            {"status": "IN_QUEUE", "queue_position": 123},
+            Queued(position=123),
+            False,
+        ),
+        (
+            {"status": "IN_PROGRESS", "logs": [{"msg": "foo"}, {"msg": "bar"}]},
+            InProgress(logs=[{"msg": "foo"}, {"msg": "bar"}]),
+            False,
+        ),
+        (
+            {"status": "COMPLETED", "logs": [{"msg": "foo"}, {"msg": "bar"}]},
+            Completed(logs=[{"msg": "foo"}, {"msg": "bar"}], metrics={}),
+            False,
+        ),
+        (
+            {"status": "COMPLETED", "logs": [{"msg": "foo"}, {"msg": "bar"}], "metrics": {"m1": "v1", "m2": "v2"}},
+            Completed(logs=[{"msg": "foo"}, {"msg": "bar"}], metrics={"m1": "v1", "m2": "v2"}),
+            False,
+        ),
+        (
+            {"status": "FOO"},
+            ValueError,
+            True,
+        )
+    ]
+)
+def test_parse_status(data, result, raised):
+    handle = _BaseRequestHandle("foo", "bar", "baz", "qux")
+
+    if raised:
+        with pytest.raises(result):
+            handle._parse_status(data)
+    else:
+        assert handle._parse_status(data) == result