Reject nonexistent requested paths during asset discovery

dandi/cli/tests/test_cmd_validate.py

@@ -37,6 +37,15 @@ def redirected_logdir(monkeypatch: pytest.MonkeyPatch, tmp_path: Path) -> Path:
     return logdir
 
 
+@pytest.mark.ai_generated
+def test_validate_missing_path(tmp_path: Path) -> None:
+    missing = tmp_path / "missing.nwb"
+    result = CliRunner().invoke(validate, [str(missing)])
+    assert result.exit_code == 2
+    assert "does not exist" in result.output
+    assert missing.name in result.output
+
+
 @pytest.mark.parametrize(
     "ds_name, expected_err_location",
     [

dandi/files/__init__.py

@@ -13,6 +13,7 @@
 
 from collections import deque
 from collections.abc import Iterator
+import errno
 import os.path
 from pathlib import Path
 
@@ -84,6 +85,10 @@ def find_dandi_files(
     they are of a type represented by a `LocalDirectoryAsset` subclass, in
     which case they are not recursed into.
 
+    All requested paths are checked before any files are yielded.  A missing
+    path raises `FileNotFoundError`; a broken symlink is still discovered so
+    that the caller can apply its missing-content policy.
+
     :param dandiset_path:
         The path to the root of the Dandiset in which the paths are located.
         All paths in ``paths`` must be equal to or subpaths of
@@ -111,6 +116,9 @@ def find_dandi_files(
                 raise ValueError(
                     f"Path {str(p)!r} is not inside Dandiset path {str(dandiset_path)!r}"
                 )
+        # Do not follow symlinks: unfetched DataLad content can be a broken link.
+        if not os.path.lexists(p):
+            raise FileNotFoundError(errno.ENOENT, os.strerror(errno.ENOENT), str(p))
         path_queue.append((Path(p), None))
     bids_roots = []
     while path_queue:

dandi/tests/test_files.py

@@ -51,6 +51,41 @@ def mkpaths(root: Path, *paths: str) -> None:
             pp.touch()
 
 
+@pytest.mark.ai_generated
+@pytest.mark.parametrize(
+    "name", ["missing.nwb", "missing.txt", "missing.zarr", ".missing"]
+)
+@pytest.mark.parametrize("allow_all", [False, True])
+def test_find_dandi_files_missing_path(
+    tmp_path: Path, name: str, allow_all: bool
+) -> None:
+    existing = tmp_path / "existing.nwb"
+    existing.touch()
+    missing = tmp_path / name
+    discovery = find_dandi_files(
+        existing, missing, dandiset_path=tmp_path, allow_all=allow_all
+    )
+    with pytest.raises(FileNotFoundError) as exc:
+        next(discovery)
+    assert exc.value.filename == str(missing)
+
+
+@pytest.mark.ai_generated
+@pytest.mark.skipif(os.name == "nt", reason="Creating symlinks requires Windows privileges")
+def test_find_dandi_files_broken_symlink(tmp_path: Path) -> None:
+    link = tmp_path / "unfetched.nwb"
+    link.symlink_to(tmp_path / "absent.nwb")
+    assets = list(find_dandi_files(link))
+    assert len(assets) == 1
+    assert isinstance(assets[0], NWBAsset)
+    assert assets[0].filepath == link
+
+
+@pytest.mark.ai_generated
+def test_find_dandi_files_empty_directory(tmp_path: Path) -> None:
+    assert list(find_dandi_files(tmp_path, dandiset_path=tmp_path)) == []
+
+
 def test_find_dandi_files(tmp_path: Path) -> None:
     mkpaths(
         tmp_path,