fix: validate_dataset checks stem-name pairing between audio and label files

src/voxkit/storage/datasets.py

@@ -510,6 +510,7 @@ def validate_dataset(dataset_path: Path) -> Tuple[bool, str]:
     - Each speaker directory contains audio files (.wav, .flac, .mp3, .ogg, .m4a)
     - Each speaker directory contains label files (.lab, .txt)
     - Number of audio files matches number of label files per speaker
+    - Each audio file has a matching label file with the same stem name
 
     Expected structure:
 
@@ -562,15 +563,9 @@ def validate_dataset(dataset_path: Path) -> Tuple[bool, str]:
         audio_files = [
             f
             for f in os.listdir(speaker_path)
-            if f.endswith(".wav")
-            or f.endswith(".flac")
-            or f.endswith(".mp3")
-            or f.endswith(".ogg")
-            or f.endswith(".m4a")
-        ]
-        label_files = [
-            f for f in os.listdir(speaker_path) if f.endswith(".lab") or f.endswith(".txt")
+            if f.endswith((".wav", ".flac", ".mp3", ".ogg", ".m4a"))
         ]
+        label_files = [f for f in os.listdir(speaker_path) if f.endswith((".lab", ".txt"))]
 
         if not audio_files:
             return False, f"No audio files found in speaker directory '{speaker_path}'."
@@ -585,4 +580,14 @@ def validate_dataset(dataset_path: Path) -> Tuple[bool, str]:
                 f"directory '{speaker_path}'.",
             )
 
+        audio_stems = {Path(f).stem for f in audio_files}
+        label_stems = {Path(f).stem for f in label_files}
+        unmatched = audio_stems.symmetric_difference(label_stems)
+        if unmatched:
+            return (
+                False,
+                f"Unpaired audio/label files in speaker directory '{speaker_path}': "
+                f"{', '.join(sorted(unmatched))}.",
+            )
+
     return True, "Dataset is valid."

tests/storage/test_datasets.py

@@ -813,3 +813,19 @@ def test_validate_dataset_mismatched_counts(self, monkeypatch):
 
             assert is_valid is False
             assert "Mismatch" in msg
+
+        def test_validate_dataset_unpaired_stems(self, monkeypatch):
+            from voxkit.storage.datasets import validate_dataset
+
+            # Create a dataset where counts match but stems do not
+            # (e.g. recording_A.wav paired with recording_B.lab)
+            unpaired_path = mock_get_storage_root() / "fake_datasets" / "unpaired_stems"
+            speaker_path = unpaired_path / "speaker_1"
+            speaker_path.mkdir(parents=True, exist_ok=True)
+            (speaker_path / "recording_A.wav").touch()
+            (speaker_path / "recording_B.lab").touch()
+
+            is_valid, msg = validate_dataset(unpaired_path)
+
+            assert is_valid is False
+            assert "Unpaired" in msg