Add linting/checks against using "feature branch" references #48

benchmark_scenarios/max_ndvi_composite.json

@@ -7,7 +7,7 @@
     "process_graph": {
       "maxndvi1": {
         "process_id": "max_ndvi_composite",
-        "namespace": "https://raw.githubusercontent.com/ESA-APEx/apex_algorithms/max_ndvi_composite/openeo_udp/examples/max_ndvi_composite/max_ndvi_composite.json",
+        "namespace": "https://raw.githubusercontent.com/ESA-APEx/apex_algorithms/refs/heads/main/openeo_udp/examples/max_ndvi_composite/max_ndvi_composite.json",
         "arguments": {
           "spatial_extent": {
             "west": 5.07,

qa/tools/apex_algorithm_qa_tools/common.py

@@ -1,6 +1,7 @@
 from __future__ import annotations
 
 import logging
+import re
 from pathlib import Path
 from typing import Iterator
 
@@ -33,3 +34,34 @@ def candidates() -> Iterator[Path]:
             return candidate
 
     raise RuntimeError("Could not determine project root directory.")
+
+
+def assert_no_github_feature_branch_refs(href: str) -> None:
+    """
+    Check that GitHub links do not point to (ephemeral) feature branches.
+    """
+    # TODO: automatically suggest commit hash based fix?
+    allowed_branches = {"main", "master"}
+
+    # Check for feature branches with explicit "refs/heads" prefix in the URL
+    if match := re.search("//raw.githubusercontent.com/.*/refs/heads/(.*?)/", href):
+        if match.group(1) not in allowed_branches:
+            # TODO: automatically suggest commit hash based fix?
+            raise ValueError(
+                f"Links should not point to ephemeral feature branches: found {match.group(1)!r} in {href!r}"
+            )
+
+    # Also check shorthand URLs without "refs/heads"
+    if match := re.search(
+        "//raw.githubusercontent.com/ESA-APEx/apex_algorithms/(.*?)/", href
+    ):
+        ref = match.group(1)
+        # Only, allow main/master and commit hashes
+        if (
+            not re.fullmatch("[0-9a-f]+", ref)
+            and ref != "refs"
+            and ref not in allowed_branches
+        ):
+            raise ValueError(
+                f"Links should not point to ephemeral feature branches: found {match.group(1)!r} in {href!r}"
+            )

qa/tools/apex_algorithm_qa_tools/scenarios.py

@@ -9,7 +9,10 @@
 
 import jsonschema
 import requests
-from apex_algorithm_qa_tools.common import get_project_root
+from apex_algorithm_qa_tools.common import (
+    assert_no_github_feature_branch_refs,
+    get_project_root,
+)
 from openeo.util import TimingLogger
 
 _log = logging.getLogger(__name__)
@@ -93,6 +96,7 @@ def lint_benchmark_scenario(scenario: BenchmarkScenario):
                     raise ValueError(
                         f"Invalid github.com based namespace {namespace!r}: should be a raw URL"
                     )
+                assert_no_github_feature_branch_refs(namespace)
                 # Inspect openEO process definition URL
                 resp = requests.get(namespace)
                 resp.raise_for_status()

qa/unittests/tests/test_algorithm_catalog.py

@@ -1,7 +1,10 @@
 import json
 
 import pytest
-from apex_algorithm_qa_tools.common import get_project_root
+from apex_algorithm_qa_tools.common import (
+    assert_no_github_feature_branch_refs,
+    get_project_root,
+)
 from esa_apex_toolbox.algorithms import Algorithm
 
 
@@ -23,6 +26,10 @@ def test_lint_algorithm_catalog_json_file(path):
     assert data["properties"]["type"] == "apex_algorithm"
 
     assert "openeo-process" in {k["rel"] for k in data["links"]}
+
+    for link in data["links"]:
+        assert_no_github_feature_branch_refs(link.get("href"))
+
     # TODO #17 more checks
 
     algo = Algorithm.from_ogc_api_record(path)

qa/unittests/tests/test_common.py

@@ -1,5 +1,40 @@
-from apex_algorithm_qa_tools.common import get_project_root
+import pytest
+from apex_algorithm_qa_tools.common import (
+    assert_no_github_feature_branch_refs,
+    get_project_root,
+)
 
 
 def test_get_project_root():
     assert get_project_root().is_dir()
+
+
+@pytest.mark.parametrize(
+    "href",
+    [
+        "https://example.com/foo/bar",
+        "https://raw.githubusercontent.com/ESA-APEx/apex_algorithms/refs/heads/main/openeo_udp/examples/max_ndvi_composite/max_ndvi_composite.json",
+        "https://raw.githubusercontent.com/ESA-APEx/apex_algorithms/main/openeo_udp/examples/max_ndvi_composite/max_ndvi_composite.json"
+        "https://raw.githubusercontent.com/ESA-APEx/apex_algorithms/3b5a011a90f4a3050ff8fdf69ca5bc2fd1535881/openeo_udp/biopar/biopar.json",
+    ],
+)
+def test_assert_no_github_feature_branch_refs_ok(href):
+    assert_no_github_feature_branch_refs(href)
+
+
+@pytest.mark.parametrize(
+    ["href", "expected_error"],
+    [
+        (
+            "https://raw.githubusercontent.com/ESA-APEx/apex_algorithms/refs/heads/footurebranch/openeo_udp/examples/max_ndvi_composite/max_ndvi_composite.json",
+            "should not point to ephemeral feature branches: found 'footurebranch'",
+        ),
+        (
+            "https://raw.githubusercontent.com/ESA-APEx/apex_algorithms/footurebranch/openeo_udp/examples/max_ndvi_composite/max_ndvi_composite.json",
+            "should not point to ephemeral feature branches: found 'footurebranch'",
+        ),
+    ],
+)
+def test_assert_no_github_feature_branch_refs_not_ok(href, expected_error):
+    with pytest.raises(ValueError, match=expected_error):
+        assert_no_github_feature_branch_refs(href)

qa/unittests/tests/test_openeo_udp.py

@@ -1,7 +1,10 @@
 import json
 
 import pytest
-from apex_algorithm_qa_tools.common import get_project_root
+from apex_algorithm_qa_tools.common import (
+    assert_no_github_feature_branch_refs,
+    get_project_root,
+)
 
 
 @pytest.mark.parametrize(
@@ -19,5 +22,9 @@ def test_lint_openeo_udp_json_file(path):
     assert "description" in data
     assert "parameters" in data
     assert "process_graph" in data
+
+    for link in data.get("links", []):
+        assert_no_github_feature_branch_refs(link.get("href"))
+
     # TODO #17 more checks
     # TODO require a standardized openEO "type"? https://github.com/Open-EO/openeo-api/issues/539