fixed error with file deletion permission

Author: ecrum19

test/test_vcf_rdfizer_unit.py

@@ -601,6 +601,69 @@ def fake_run(cmd, cwd=None, env=None):
             self.assertTrue((out_dir / "sample" / "sample.nt").exists())
             self.assertTrue((out_dir / "hdt" / "sample.hdt").exists())
 
+    def test_main_full_mode_deletes_nt_with_docker_fallback_on_permission_error(self):
+        """Full mode falls back to Docker-based removal when .nt unlink raises PermissionError."""
+        with tempfile.TemporaryDirectory() as td:
+            tmp_path = Path(td)
+            input_dir, rules_path = prepare_inputs(tmp_path)
+            out_dir = tmp_path / "out"
+            target_nt = out_dir / "sample" / "sample.nt"
+            target_nt_resolved = target_nt.resolve()
+            commands = []
+            original_unlink = Path.unlink
+
+            def fake_run(cmd, cwd=None, env=None):
+                commands.append(cmd)
+                if "/opt/vcf-rdfizer/run_conversion.sh" in cmd:
+                    target_nt.parent.mkdir(parents=True, exist_ok=True)
+                    target_nt.write_text("<s> <p> <o> .\n")
+                if "/opt/vcf-rdfizer/compression.sh" in cmd:
+                    hdt_dir = out_dir / "hdt"
+                    hdt_dir.mkdir(parents=True, exist_ok=True)
+                    (hdt_dir / "sample.hdt").write_text("fake-hdt\n")
+                if isinstance(cmd, list) and cmd[-1].startswith("rm -f ") and "/data/out/sample/sample.nt" in cmd[-1]:
+                    if target_nt_resolved.exists():
+                        original_unlink(target_nt_resolved)
+                return 0
+
+            def unlink_side_effect(path_obj, *args, **kwargs):
+                if path_obj.resolve() == target_nt_resolved:
+                    raise PermissionError(13, "Permission denied", str(path_obj))
+                return original_unlink(path_obj, *args, **kwargs)
+
+            old_cwd = os.getcwd()
+            os.chdir(tmp_path)
+            try:
+                with mock.patch.object(vcf_rdfizer, "run", side_effect=fake_run), mock.patch.object(
+                    vcf_rdfizer, "check_docker", return_value=True
+                ), mock.patch.object(
+                    vcf_rdfizer, "docker_image_exists", return_value=True
+                ), mock.patch.object(
+                    vcf_rdfizer, "discover_tsv_triplets", return_value=mocked_triplets()
+                ), mock.patch("pathlib.Path.unlink", autospec=True, side_effect=unlink_side_effect):
+                    rc = invoke_main(
+                        [
+                            "--input",
+                            str(input_dir),
+                            "--rules",
+                            str(rules_path),
+                            "--out",
+                            str(out_dir),
+                            "--keep-tsv",
+                        ]
+                    )
+            finally:
+                os.chdir(old_cwd)
+
+            self.assertEqual(rc, 0)
+            self.assertFalse(target_nt.exists())
+            self.assertTrue(
+                any(
+                    isinstance(cmd, list) and cmd[-1].startswith("rm -f ") and "/data/out/sample/sample.nt" in cmd[-1]
+                    for cmd in commands
+                )
+            )
+
     def test_main_ignores_unrelated_existing_tsv_triplets(self):
         """Wrapper converts only triplets that match the CLI-selected VCF snapshot."""
         with tempfile.TemporaryDirectory() as td:

vcf_rdfizer.py

@@ -189,6 +189,50 @@ def print_nt_hdt_summary(*, output_root: Path, nt_path: Path, hdt_path: Path):
         print(f"  - HDT (.hdt): {format_bytes(hdt_size)} ({hdt_path})")
 
 
+def remove_file_with_docker_fallback(
+    *,
+    path: Path,
+    mount_root: Path,
+    mount_point: str,
+    image_ref: str,
+    wrapper_log_path: Path,
+) -> bool:
+    if not path.exists():
+        return True
+
+    try:
+        path.unlink()
+        return True
+    except PermissionError:
+        pass
+
+    try:
+        rel = path.relative_to(mount_root)
+    except ValueError:
+        eprint(f"Error: cannot remove file outside mounted root: {path}")
+        eprint(f"See log for details: {wrapper_log_path}")
+        return False
+
+    container_path = f"{mount_point}/{rel.as_posix()}"
+    rm_cmd = [
+        "sudo",
+        "docker",
+        "run",
+        "--rm",
+        "-v",
+        f"{str(mount_root)}:{mount_point}",
+        image_ref,
+        "bash",
+        "-lc",
+        f"rm -f {shlex.quote(container_path)}",
+    ]
+    if run(rm_cmd) != 0:
+        eprint(f"Error: failed to remove file with Docker fallback: {path}")
+        eprint(f"See log for details: {wrapper_log_path}")
+        return False
+    return True
+
+
 def existing_parent(path: Path) -> Path:
     cur = path
     while not cur.exists():
@@ -531,7 +575,14 @@ def run_full_mode(
 
         if not keep_rdf and selected_methods:
             if nt_path.exists():
-                nt_path.unlink()
+                if not remove_file_with_docker_fallback(
+                    path=nt_path,
+                    mount_root=out_dir,
+                    mount_point="/data/out",
+                    image_ref=image_ref,
+                    wrapper_log_path=wrapper_log_path,
+                ):
+                    return 1
                 print(f"    * Removed N-Triples file (set --keep-rdf to retain): {nt_path}")
             else:
                 print(f"    * N-Triples cleanup skipped (not found): {nt_path}")
@@ -541,7 +592,14 @@ def run_full_mode(
         if not keep_tsv:
             for tsv_path in (triplet["records"], triplet["headers"], triplet["metadata"]):
                 if tsv_path.exists():
-                    tsv_path.unlink()
+                    if not remove_file_with_docker_fallback(
+                        path=tsv_path,
+                        mount_root=tsv_dir,
+                        mount_point="/data/tsv",
+                        image_ref=image_ref,
+                        wrapper_log_path=wrapper_log_path,
+                    ):
+                        return 1
 
     if not keep_tsv:
         if not tsv_existed:
@@ -731,9 +789,10 @@ def main():
     )
     parser.add_argument(
         "-q",
+        "--nq",
         "--nt",
         "--rdf",
-        dest="nt",
+        dest="nq",
         default=None,
         help="Input RDF file (.nt or .nq) for --mode compress",
     )