Improve Cargo testing with more cases

Signed-off-by: ziad hany <[email protected]>

Author: ziadhany

minecode_pipelines/miners/cargo.py

@@ -45,7 +45,10 @@ def process_cargo_packages(cargo_index_repo, cloned_data_repo, config_repo, logg
 
         try:
             next_commit = get_commit_at_distance_ahead(
-                cargo_index_repo, checkpoints_last_commit, num_commits_ahead=COMMIT_BATCH_SIZE, branch_name="master"
+                cargo_index_repo,
+                checkpoints_last_commit,
+                num_commits_ahead=COMMIT_BATCH_SIZE,
+                branch_name="master",
             )
         except ValueError as e:
             logger(str(e))

minecode_pipelines/pipes/cargo.py

@@ -29,4 +29,4 @@ def store_cargo_packages(packages, repo):
     ppath = hashid.get_package_purls_yml_file_path(base_purl)
     purl_file_full_path = Path(repo.working_dir) / ppath
     write_data_to_yaml_file(path=purl_file_full_path, data=updated_purls)
-    return purl_file_full_path, base_purl
+    return purl_file_full_path, base_purl

minecode_pipelines/tests/pipes/test_cargo.py

@@ -1,3 +1,12 @@
+#
+# Copyright (c) nexB Inc. and others. All rights reserved.
+# purldb is a trademark of nexB Inc.
+# SPDX-License-Identifier: Apache-2.0
+# See http://www.apache.org/licenses/LICENSE-2.0 for the license text.
+# See https://github.com/aboutcode-org/purldb for support or download.
+# See https://aboutcode.org for more information about nexB OSS projects.
+#
+
 import json
 import tempfile
 from pathlib import Path
@@ -6,7 +15,7 @@
 import saneyaml
 from django.test import TestCase
 
-from minecode_pipelines.pipes import write_packageurls_to_file
+from minecode_pipelines.pipes import write_data_to_yaml_file
 from minecode_pipelines.pipes.cargo import store_cargo_packages
 
 DATA_DIR = Path(__file__).parent.parent / "test_data" / "cargo"
@@ -27,22 +36,24 @@ def test_collect_packages_from_cargo_calls_write(self, mock_write):
         with open(expected_file, encoding="utf-8") as f:
             expected = saneyaml.load(f)
 
-        repo = Mock()
-        store_cargo_packages(packages, repo)
+        with tempfile.TemporaryDirectory() as tmpdir:
+            repo = Mock()
+            repo.working_dir = tmpdir
 
-        mock_write.assert_called_once()
-        args, kwargs = mock_write.call_args
-        called_repo, base_purl, written_packages = args
+            store_cargo_packages(packages, repo)
 
-        self.assertEqual(called_repo, repo)
+            mock_write.assert_called_once()
+            args, kwargs = mock_write.call_args
+            base_purl, written_packages = kwargs["path"], kwargs["data"]
 
-        expected_base_purl = "aboutcode-packages-cargo-0/cargo/c5store/purls.yml"
-        self.assertEqual(str(base_purl), str(expected_base_purl))
-        self.assertEqual(written_packages, expected)
+            expected_base_purl = (
+                Path(tmpdir) / "aboutcode-packages-cargo-0" / "cargo" / "c5store" / "purls.yml"
+            )
 
-    def test_add_purl_result_with_mock_repo(self):
-        purls = [{"purl": "pkg:pypi/[email protected]"}, {"purl": "pkg:pypi/[email protected]"}]
+            self.assertEqual(str(base_purl), str(expected_base_purl))
+            self.assertEqual(written_packages, expected)
 
+    def _assert_purls_written(self, purls):
         with tempfile.TemporaryDirectory() as tmpdir:
             repo_dir = Path(tmpdir)
 
@@ -52,11 +63,23 @@ def test_add_purl_result_with_mock_repo(self):
 
             purls_file = repo_dir / "purls.yaml"
 
-            relative_path = write_packageurls_to_file(mock_repo, purls_file, purls)
+            write_data_to_yaml_file(purls_file, purls)
 
-            written_file = repo_dir / relative_path
-            self.assertTrue(written_file.exists())
+            self.assertTrue(purls_file.exists())
 
-            with open(written_file, encoding="utf-8") as f:
+            with open(purls_file, encoding="utf-8") as f:
                 content = saneyaml.load(f)
+
             self.assertEqual(content, purls)
+
+    def test_add_purl_result_with_mock_repo(self):
+        self._assert_purls_written(
+            [{"purl": "pkg:pypi/[email protected]"}, {"purl": "pkg:pypi/[email protected]"}]
+        )
+
+    def test_add_empty_purl_result_with_mock_repo(self):
+        self._assert_purls_written([])
+
+    def test_add_invalid_purl_with_mock_repo(self):
+        # invalid but still written as empty file
+        self._assert_purls_written([{"purl": "pkg:pypi/django"}])

minecode_pipelines/tests/pipes/test_pipes.py

@@ -0,0 +1,65 @@
+#
+# Copyright (c) nexB Inc. and others. All rights reserved.
+# purldb is a trademark of nexB Inc.
+# SPDX-License-Identifier: Apache-2.0
+# See http://www.apache.org/licenses/LICENSE-2.0 for the license text.
+# See https://github.com/aboutcode-org/purldb for support or download.
+# See https://aboutcode.org for more information about nexB OSS projects.
+#
+import tempfile
+from pathlib import Path
+from unittest import TestCase
+from git import Repo
+
+from minecode_pipelines.pipes import get_commit_at_distance_ahead
+
+
+class GetCommitAtDistanceAheadIntegrationTests(TestCase):
+    def setUp(self):
+        # Create a temporary directory and init a repo
+        self.tmpdir = tempfile.TemporaryDirectory()
+        self.repo_path = Path(self.tmpdir.name)
+        self.repo = Repo.init(self.repo_path)
+
+        # Configure identity (needed for commits)
+        with self.repo.config_writer() as cw:
+            cw.set_value("user", "name", "Test User")
+            cw.set_value("user", "email", "[email protected]")
+
+        # Create 5 commits
+        self.commits = []
+        for i in range(5):
+            file_path = self.repo_path / f"file{i}.txt"
+            file_path.write_text(f"content {i}")
+            self.repo.index.add([str(file_path)])
+            commit = self.repo.index.commit(f"commit {i}")
+            self.commits.append(commit.hexsha)
+
+        # By construction, self.commits[0] = first commit, self.commits[-1] = latest commit
+
+    def tearDown(self):
+        self.tmpdir.cleanup()
+
+    def test_get_commit_at_distance_none_current_commit(self):
+        # If current_commit is None, it should start from the empty tree hash
+        result = get_commit_at_distance_ahead(
+            self.repo, None, num_commits_ahead=3, branch_name="master"
+        )
+        # Should return the 3rd commit in history
+        self.assertEqual(result, self.commits[2])
+
+    def test_get_commit_at_distance(self):
+        # current_commit = first commit, ask for 3 commits ahead
+        result = get_commit_at_distance_ahead(
+            self.repo, self.commits[0], num_commits_ahead=3, branch_name="master"
+        )
+        # Should return the 3rd commit from start (self.commits[3])
+        self.assertEqual(result, self.commits[3])
+
+    def test_raises_if_not_enough_commits(self):
+        # From latest commit, ask for 10 ahead (only 0 available)
+        with self.assertRaises(ValueError) as cm:
+            get_commit_at_distance_ahead(
+                self.repo, self.commits[-1], num_commits_ahead=10, branch_name="master"
+            )
+        self.assertIn("Not enough commits ahead; only 0 available.", str(cm.exception))