Chore: Fix scripts_bazel tests not running

- Fixed tests in scripts_bazel
- Expanded tests in scripts_bazel
- Adapted workflow to tests all of docs

Author: MaximilianSoerenPollak

.github/workflows/test.yml

@@ -36,7 +36,7 @@ jobs:
       - name: Run test targets
         run: |
           bazel run --lockfile_mode=error //:ide_support
-          bazel test --lockfile_mode=error //src/... //score_pytest/...
+          bazel test --lockfile_mode=error //...
 
       - name: Prepare bundled consumer report
         if: always()

scripts_bazel/tests/generate_sourcelinks_cli_test.py

@@ -11,17 +11,68 @@
 # SPDX-License-Identifier: Apache-2.0
 # *******************************************************************************
 
+# ╓                                                          ╖
+# ║ Some parts are generated by Gemini                       ║
+# ╙                                                          ╜
+
 """Tests for generate_sourcelinks_cli.py"""
 
 import json
-import subprocess
 import sys
 from pathlib import Path
 
+import pytest
+
+import scripts_bazel.generate_sourcelinks_cli
+from src.extensions.score_source_code_linker.needlinks import is_metadata
+
 _MY_PATH = Path(__file__).parent
 
 
-def test_generate_sourcelinks_cli_basic(tmp_path: Path) -> None:
+def assert_json_internal_types(input: list[dict[str, str | int]]):
+    for entry in input:
+        assert "file" in entry
+        assert "line" in entry
+        assert "tag" in entry
+        assert "need" in entry
+        assert "full_line" in entry
+
+        assert isinstance(entry["file"], str)
+        assert isinstance(entry["line"], int)
+        assert isinstance(entry["tag"], str)
+        assert isinstance(entry["need"], str)
+        assert isinstance(entry["full_line"], str)
+
+
+# Unit test generated by Gemini.
+@pytest.mark.parametrize(
+    "input_path, expected_output",
+    [
+        # Case 1: Path starts with "external/" and has a project name
+        (Path("external/score_docs_as_code+/docs/index.md"), Path("docs/index.md")),
+        # Case 2: Path does NOT start with "external/"
+        (Path("src/main.py"), Path("src/main.py")),
+        # Case 3: Path has "external" elsewhere in the string (should not be removed)
+        (Path("my_external_data/file.txt"), Path("my_external_data/file.txt")),
+        # Case 4: Deeply nested path inside an external prefix
+        (Path("external/repo/subfolder/file.py"), Path("subfolder/file.py")),
+        # Case 5: Path is exactly "external/" (edge case, returns empty path based on split logic)
+        (Path("external/"), Path("external")),
+    ],
+)
+def test_clean_external_prefix(input_path: Path, expected_output: Path):
+    output = scripts_bazel.generate_sourcelinks_cli.clean_external_prefix(input_path)
+    print(f"{input_path} => {output}")
+
+    assert (
+        scripts_bazel.generate_sourcelinks_cli.clean_external_prefix(input_path)
+        == expected_output
+    )
+
+
+def test_generate_sourcelinks_cli_basic(
+    tmp_path: Path, monkeypatch: pytest.MonkeyPatch
+) -> None:
     """Test basic functionality of generate_sourcelinks_cli."""
     # Create a test source file with a traceability tag
     test_file = tmp_path / "test_source.py"
@@ -36,39 +87,116 @@ def some_function():
 
     output_file = tmp_path / "output.json"
 
-    # Execute the script
-    result = subprocess.run(
-        [
-            sys.executable,
-            _MY_PATH.parent / "generate_sourcelinks_cli.py",
-            "--output",
-            str(output_file),
-            str(test_file),
-        ],
-    )
-
-    assert result.returncode == 0
-    assert output_file.exists()
+    test_args: list[Path | str] = [
+        _MY_PATH.parent
+        / "generate_sourcelinks_cli.py",  # sys.argv[0] is always the script name
+        "--output",
+        str(output_file),
+        str(test_file),
+    ]
+    monkeypatch.setattr(sys, "argv", test_args)
+    result = scripts_bazel.generate_sourcelinks_cli.main()
+    assert result == 0
 
     # Check the output content
     with open(output_file) as f:
         data: list[dict[str, str | int]] = json.load(f)
     assert isinstance(data, list)
-    assert len(data) > 0
+    # The first dictionary has to be metadata
+    assert len(data) == 2
+    assert is_metadata(data[0])
+    assert data[0]["repo_name"] == "local_repo"
+    # hash & url can not be set in this script therefore HAVE to be empty
+    assert data[0]["hash"] == ""
+    assert data[0]["url"] == ""
 
     # Verify schema of each entry
-    for entry in data:
-        assert "file" in entry
-        assert "line" in entry
-        assert "tag" in entry
-        assert "need" in entry
-        assert "full_line" in entry
+    assert_json_internal_types(data[1:])
 
-        # Verify types
-        assert isinstance(entry["file"], str)
-        assert isinstance(entry["line"], int)
-        assert isinstance(entry["tag"], str)
-        assert isinstance(entry["need"], str)
-        assert isinstance(entry["full_line"], str)
+    assert data[1]["need"] == "tool_req__docs_arch_types"
+
+
+def test_generate_sourcelinks_cli_parse_external_module(
+    tmp_path: Path, monkeypatch: pytest.MonkeyPatch
+):
+    # 1. Create the 'external' directory inside the temp path
+    external_root = tmp_path / "external"
+    external_root.mkdir()
+
+    # 2. Create your file inside that 'external' directory
+    test_file = external_root / "score_baselibs+" / "src" / "source_file1.py"
+    test_file.parent.mkdir(parents=True)
+    test_file.write_text("content")
 
-    assert any(entry["need"] == "tool_req__docs_arch_types" for entry in data)
+    # 3. Create the path relative to tmp_path so it starts with 'external/'
+    # Use .relative_to(tmp_path) to get 'external/score_docs_as_code+/...'
+    test_file.write_text(
+        """
+# Some code here
+# req-Id: tool_req__docs_arch_types
+def some_function():
+    pass
+"""
+    )
+    output_file = tmp_path / "output.json"
+    monkeypatch.chdir(tmp_path)
+    relative_test_file = test_file.relative_to(tmp_path)
+    test_args: list[Path | str] = [
+        _MY_PATH.parent
+        / "generate_sourcelinks_cli.py",  # sys.argv[0] is always the script name
+        "--output",
+        str(output_file),
+        str(relative_test_file),
+    ]
+    monkeypatch.setattr(sys, "argv", test_args)
+    result = scripts_bazel.generate_sourcelinks_cli.main()
+    assert result == 0
+    with open(output_file) as f:
+        data: list[dict[str, str | int]] = json.load(f)
+    assert isinstance(data, list)
+    assert len(data) == 2
+    # The first dictionary has to be metadata
+    assert is_metadata(data[0])
+    assert data[0]["repo_name"] == "score_baselibs"
+    # hash & url can not be set in this script therefore HAVE to be empty
+    assert data[0]["hash"] == ""
+    assert data[0]["url"] == ""
+
+    # Verify schema of each entry
+    assert_json_internal_types(data[1:])
+
+
+def test_generate_sourcelinks_cli_file_not_exists(
+    tmp_path: Path, monkeypatch: pytest.MonkeyPatch
+):
+    external_root = tmp_path / "external"
+    external_root.mkdir()
+
+    # 2. Create your file inside that 'external' directory
+    test_file = external_root / "score_baselibs+" / "src" / "source_file1.py"
+    test_file.parent.mkdir(parents=True)
+    test_file.write_text("content")
+
+    # 3. Create the path relative to tmp_path so it starts with 'external/'
+    # Use .relative_to(tmp_path) to get 'external/score_docs_as_code+/...'
+    test_file.write_text(
+        """
+# Some code here
+# req-Id: tool_req__docs_arch_types
+def some_function():
+    pass
+"""
+    )
+    output_file = tmp_path / "output.json"
+    # BY not changing directory (like above) we can FORCE the file to not exists
+    relative_test_file = test_file.relative_to(tmp_path)
+    test_args: list[Path | str] = [
+        _MY_PATH.parent
+        / "generate_sourcelinks_cli.py",  # sys.argv[0] is always the script name
+        "--output",
+        str(output_file),
+        str(relative_test_file),
+    ]
+    monkeypatch.setattr(sys, "argv", test_args)
+    with pytest.raises(AssertionError):
+        scripts_bazel.generate_sourcelinks_cli.main()