Fix tests for explicit default dataset

Author: MaxGhenis

policyengine_uk/tests/conftest.py

@@ -0,0 +1,28 @@
+import os
+
+import pytest
+
+DEFAULT_TEST_DATASET_URL = (
+    "hf://policyengine/policyengine-uk-data-private/enhanced_frs_2023_24.h5@1.40.3"
+)
+
+if os.environ.get("HUGGING_FACE_TOKEN") and not os.environ.get(
+    "POLICYENGINE_UK_DEFAULT_DATASET"
+):
+    os.environ["POLICYENGINE_UK_DEFAULT_DATASET"] = DEFAULT_TEST_DATASET_URL
+
+
+def pytest_collection_modifyitems(config, items):
+    has_default_dataset = bool(os.environ.get("POLICYENGINE_UK_DEFAULT_DATASET"))
+    if has_default_dataset:
+        return
+
+    skip_microsimulation = pytest.mark.skip(
+        reason=(
+            "Requires POLICYENGINE_UK_DEFAULT_DATASET or HUGGING_FACE_TOKEN "
+            "for microsimulation dataset access"
+        )
+    )
+    for item in items:
+        if "microsimulation" in item.keywords:
+            item.add_marker(skip_microsimulation)

policyengine_uk/tests/test_behavioral_responses.py

@@ -16,8 +16,11 @@
 from policyengine_uk import Microsimulation
 from policyengine_uk.model_api import Scenario
 
-# Check if HF token is available for data-dependent tests
-HF_TOKEN_AVAILABLE = bool(os.environ.get("HUGGING_FACE_TOKEN"))
+# Check if a default dataset is available for data-dependent tests
+HF_TOKEN_AVAILABLE = bool(
+    os.environ.get("HUGGING_FACE_TOKEN")
+    or os.environ.get("POLICYENGINE_UK_DEFAULT_DATASET")
+)
 requires_hf_data = pytest.mark.skipif(
     not HF_TOKEN_AVAILABLE,
     reason="Requires HUGGING_FACE_TOKEN for private data access",