fix vp_nel

data/slue-vp_nel_blind.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f889fae454a5de7ee71f430172057e8ddfb5f4c2c3d34623cf6e0e7ec75c6679
-size 168701758
+oid sha256:62a1fd6d01a19a08042bd7e1ae8549c08db67001cecf0fd43c96d4da3154e1d7
+size 168568066

slue-phase-2.py

@@ -58,9 +58,9 @@ For questions from the other 4 datasets, their question texts, answer strings, a
 
 SLUE-SQA-5 also contains a subset of Spoken Wikipedia, including the audios placed in “document” directories and their transcripts (document_text and normalized_document_text column in .tsv files). Additionally, we provide the text-to-speech alignments (.txt files in “word2time” directories).These contents are licensed with the same Creative Commons (CC BY-SA 4.0) license as Spoken Wikipedia.
 =======================================================
-SLUE-VP-NEL Dataset
+SLUE-vp_nel Dataset
 
-SLUE-VP-NEL includes word-level time stamps for dev and test splits of the SLUE-voxpopuli corpus.
+SLUE-vp_nel includes word-level time stamps for dev and test splits of the SLUE-voxpopuli corpus.
 For the dev split, the dataset also contains named entity annotations and corresponding time-stamps in a tsv format. 
 =======================================================
 
@@ -106,6 +106,8 @@ def load_word2time(word2time_file):
 
 def parse_nel_time_spans(nel_timestamps):
     nel_timestamps = ast.literal_eval(nel_timestamps)
+    if nel_timestamps is None:
+        return []
     return [
         {
             "ne_label": ne,
@@ -157,7 +159,7 @@ class SLUE2(datasets.GeneratorBasedBuilder):
         ),
         SLUE2Config(
             name="vp_nel",
-            description="SLUE-VP-NEL set with named entity labels and time-stamps.",
+            description="SLUE-vp_nel set with named entity labels and time-stamps.",
         ),
     ]
 
@@ -209,10 +211,9 @@ class SLUE2(datasets.GeneratorBasedBuilder):
         elif self.config.name == "vp_nel":
             features = {
                 "id": datasets.Value("string"),
-                "split": datasets.Value("string"),
                 "audio": datasets.Audio(sampling_rate=16_000),
                 "speaker_id": datasets.Value("string"),
-                "normalized_text": datasets.Value("string"),
+                "text": datasets.Value("string"),
                 "word_timestamps": datasets.Sequence(
                     {
                         "word": datasets.Value("string"),
@@ -220,7 +221,7 @@ class SLUE2(datasets.GeneratorBasedBuilder):
                         "end_sec": datasets.Value("float64"),
                     }
                 ),
-                "normalized_nel": datasets.Sequence(
+                "ne_timestamps": datasets.Sequence(
                     {
                         "ne_label": datasets.Value("string"),
                         "start_char_idx": datasets.Value("int32"),
@@ -247,7 +248,6 @@ class SLUE2(datasets.GeneratorBasedBuilder):
 
         dl_dir = dl_manager.download_and_extract(_DL_URLS[config_name])
         data_dir = os.path.join(dl_dir, config_name)
-        print(data_dir)
 
         splits = []
         if self.config.name in ["hvb", "sqa5"]:
@@ -352,7 +352,7 @@ class SLUE2(datasets.GeneratorBasedBuilder):
                             "word2time": load_word2time(word2time_file),
                             "answer_spans": parse_qa_answer_spans(row.get("answer_spans", "[]")),
                     }
-                elif self.config.name == "slue_nel":
+                elif self.config.name == "vp_nel":
                     split = "test" if "test" in filepath else "dev"
                     utt_id = row["id"]
                     word_alignments_fn = os.path.join(
@@ -360,6 +360,7 @@ class SLUE2(datasets.GeneratorBasedBuilder):
                     )
                     audio_file = os.path.join(
                         data_dir,
+                        'audio',
                         split,
                         f"{utt_id}.ogg",
                     )