fix: always validate permissions_obtained

also add a unittest

everyvoice/config/preprocessing_config.py

@@ -99,6 +99,7 @@ class Dataset(PartialLoadConfig):
     permissions_obtained: bool = Field(
         False,
         description="An attestation that permission has been obtained to use this data. You may not use EveryVoice to build a TTS system with data that you do not have permission to use and there are serious possible consequences for doing so. Finding data online does not constitute permission. The speaker should be aware and consent to their data being used in this way.",
+        validate_default=True,
     )
     data_dir: PossiblyRelativePath = Field(
         Path("/please/create/a/path/to/your/dataset/data"),

everyvoice/tests/data/relative/config/everyvoice-shared-data.yaml

@@ -11,6 +11,7 @@ source_data:
   filelist: ../r-filelist.psv
   filelist_loader: everyvoice.utils.generic_psv_filelist_reader
   label: dataset_0
+  permissions_obtained: true
   sox_effects:
   - [channel, '1']
 train_split: 0.9

everyvoice/tests/test_configs.py

@@ -316,7 +316,12 @@ def test_shared_sox(self) -> None:
         vocoder_config = VocoderConfig(
             contact=self.contact,
             preprocessing=PreprocessingConfig(
-                source_data=[Dataset(), Dataset(), Dataset(), Dataset()]
+                source_data=[
+                    Dataset(permissions_obtained=True),
+                    Dataset(permissions_obtained=True),
+                    Dataset(permissions_obtained=True),
+                    Dataset(permissions_obtained=True),
+                ]
             ),
         )
         config: EveryVoiceConfig = EveryVoiceConfig(

everyvoice/tests/test_preprocessing.py

@@ -59,6 +59,14 @@ def test_run_doctest(self):
     def test_read_filelist(self):
         self.assertEqual(self.filelist[0]["basename"], "LJ050-0269")
 
+    def test_no_permissions(self):
+        no_permissions_args = self.fp_config.model_dump()
+        no_permissions_args["preprocessing"]["source_data"][0][
+            "permissions_obtained"
+        ] = False
+        with self.assertRaises(ValueError):
+            FeaturePredictionConfig(**no_permissions_args)
+
     def test_process_audio_for_alignment(self):
         config = AlignerConfig(contact=self.contact)
         for entry in self.filelist[1:]:
@@ -392,9 +400,9 @@ def test_text_processing(self):
                     preprocessed_dir.mkdir(parents=True, exist_ok=True)
                     output_filelist = preprocessed_dir / "preprocessed_filelist.psv"
                     shutil.copyfile(filelist_test_info["path"], output_filelist)
-                    fp_config.preprocessing.source_data[0].filelist = (
-                        filelist_test_info["path"]
-                    )
+                    fp_config.preprocessing.source_data[
+                        0
+                    ].filelist = filelist_test_info["path"]
                     fp_config.preprocessing.save_dir = preprocessed_dir
                     preprocessor = Preprocessor(fp_config)
                     with capture_stdout() as output, mute_logger(
@@ -489,9 +497,13 @@ def test_incremental_preprocess(self):
         with tempfile.TemporaryDirectory(
             prefix="test_incremental_preprocess", dir="."
         ) as tmpdir:
-            fp_config, lj_filelist, full_filelist, partial_filelist, to_process = (
-                self.get_simple_config(tmpdir)
-            )
+            (
+                fp_config,
+                lj_filelist,
+                full_filelist,
+                partial_filelist,
+                to_process,
+            ) = self.get_simple_config(tmpdir)
 
             fp_config.preprocessing.source_data[0].filelist = partial_filelist
             with capture_stdout() as output, mute_logger("everyvoice.preprocessor"):