ensure correct alignment between metadata and raw data

data/clip_metadata.py

@@ -7,4 +7,5 @@ df = pd.read_parquet("data/dataset_audio_" + _SPLIT +".parquet.gzip")
 clipped_df = df.filter(["label_string", "probability", "probability_vector", "prediction",
                 "prediction_string", "embedding_reduced"], axis=1)
 
+clipped_df = clipped_df.sort_values(by=["label_string"])
 clipped_df.to_parquet("data/dataset_audio_" + _SPLIT +"_clipped.parquet.gzip")

data/dataset_audio_test_clipped.parquet.gzip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:334324791b5c45096971f7bbf6a05afa60160d9ff5a9fc966762b9c2923cff02
-size 642187
+oid sha256:f547f9d8c86950a42baf9ddf1a5a52cd04a8edc9200529ddf61cea4902dca30c
+size 659677

data/dataset_audio_train_clipped.parquet.gzip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b3fed567b3d361fcfff79ad59a2d45233ef663d965ef17a4d9f42d3bc9662693
-size 7864308
+oid sha256:434345cc785bb9d400d8764edbe413df27c5ade323ce7f4545372f6e7932106c
+size 8172206

data/dataset_audio_validation_clipped.parquet.gzip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f0dfba31f1d0587ebb4b9cbeb96fd10ae93cac9036dc482f4802b1495ab6806c
-size 1443196
+oid sha256:58aecf14451ee7a32c6723c9954e79324ff7c937ed8b739eb6f2ff277f9f8285
+size 1482039

speech_commands_enriched.py

@@ -229,11 +229,13 @@ class SpeechCommands(datasets.GeneratorBasedBuilder):
         # HINT: metadata should already be the split-specific metadata
 
         pathlist = Path(archive_path).glob('**/*.wav')
+        # sort _silence_ after all the other paths; aligns metadata to data as data is not sorted by #
+        # ascending order
+        pathlist = sorted(pathlist, key=lambda d: str(d).lower().replace("_", "{"))
 
         for path, row in zip(pathlist, metadata.iterrows()): 
             
             # row is a tuple containg an index and a pandas series    
-
             pathcomponents = str(path).split("/")
             word = pathcomponents[-2]
             audio_filename = pathcomponents[-1]
@@ -268,6 +270,6 @@ class SpeechCommands(datasets.GeneratorBasedBuilder):
             }
 
 #for debugging, comment out after
-if __name__ == "__main__":
-    ds = datasets.load_dataset("speech_commands_enriched.py", 'v0.01', split="test", 
-                               streaming=False)
+#if __name__ == "__main__":
+    #ds = datasets.load_dataset("speech_commands_enriched.py", 'v0.01', split="test", 
+                               #streaming=False)