Give a more helpful message when endpoint not found in source.

.github/workflows/continuous-integration.yml

@@ -6,7 +6,7 @@ on:
 jobs:
   build:
 
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-22.04
     strategy:
       matrix:
         python-version: [3.8, 3.9]

digital_land/collection.py

@@ -168,6 +168,11 @@ def load(
             organisations = set()
             datasets = set()
             for endpoint in resource["endpoints"]:
+                if endpoint not in source.records:
+                    raise RuntimeError(
+                        f"Endpoint '{endpoint}' not found in source. Check the endpoint.csv and source.csv files."
+                    )
+
                 for entry in source.records[endpoint]:
                     organisations.add(entry["organisation"])
                     datasets = set(

tests/unit/test_collection.py

@@ -1,5 +1,6 @@
+import pytest
 from digital_land.register import hash_value, Item
-from digital_land.collection import Collection, LogStore
+from digital_land.collection import Collection, LogStore, ResourceLogStore, CSVStore
 from digital_land.schema import Schema
 from datetime import datetime
 
@@ -90,3 +91,30 @@ def test_format_date():
         for date in dates:
             print(date)
             assert Collection.format_date(date) == check
+
+
+def test_endpoint_source_mismatch():
+    log = LogStore(Schema("log"))
+    log.add_entry(
+        {
+            "endpoint": "abc123",
+            "entry-date": "2025-01-06",
+            "resource": "aaa",
+            "bytes": "1024",
+        }
+    )
+
+    # This one matches
+    source = CSVStore(Schema("source"))
+    source.add_entry({"endpoint": "abc123", "organisation": "test-org"})
+
+    resource = ResourceLogStore(Schema("resource"))
+    resource.load(log, source)
+
+    # This one doesn't
+    source = CSVStore(Schema("source"))
+    source.add_entry({"endpoint": "abc124", "organisation": "test-org"})
+
+    resource = ResourceLogStore(Schema("resource"))
+    with pytest.raises(RuntimeError):
+        resource.load(log, source)