Update ESLO.py

ESLO.py

@@ -1,3 +1,5 @@
+"""Multilingual Librispeech automatic speech recognition dataset."""
+
 import os
 import re
 from ctypes import Array
@@ -24,7 +26,7 @@ _DESCRIPTION = """\
 ESLO dataset, each utterance are taken out individually
 """
 SAMPLING_RATE = 16000
-AUDIO_FOLDER = "audio"
+AUDIO_FOLDER = "/Users/brunohays/datalab/épellations/Scrapping/ortolang_datasets/eslo/downloads/audio"
 AUDIO_FILES = ['ESLO2_ITI_1073.mp4', 'ESLO1_ENT_117.mp4', 'ESLO1_ENT_103.mp4', 'ESLO2_ITI_1098.mp4',
                'ESLO1_ENTCONT_237.mp4', 'ESLO1_ENTCONT_223.mp4', 'ESLO2_RUMEUR_1339.mp4', 'ESLO2_INTPERS_1245.mp4',
                'ESLO1_TEL_364.mp4', 'ESLO2_ITI_1107.mp4', 'ESLO1_CONF_503.mp4', 'ESLO1_MAG_631.mp4',
@@ -261,8 +263,8 @@ class MultilingualLibrispeech(datasets.GeneratorBasedBuilder):
     def _split_generators(self, dl_manager):
 
         transcripts = dl_manager.download({
-            "train": "transcripts_deduplicated_train.zip",
-            "test": "transcripts_deduplicated_test.zip",
+            "train": "/Users/brunohays/datalab/épellations/Scrapping/ortolang_datasets/eslo/downloads/transcripts_deduplicated_train.zip",
+            "test": "/Users/brunohays/datalab/épellations/Scrapping/ortolang_datasets/eslo/downloads/transcripts_deduplicated_test.zip",
         })
         audio_files = dl_manager.download(
             {audio_file: os.path.join(AUDIO_FOLDER, audio_file) for audio_file in AUDIO_FILES})
@@ -297,7 +299,7 @@ class MultilingualLibrispeech(datasets.GeneratorBasedBuilder):
 
     def clean_text(self, text: str) -> str:
         def replace_uppercase(match):
-            """replaces BRUNO spelling by B R U N O"""
+            """replaces BRUNO spelling  by B R U N O"""
             return ' '.join(match.group(1))
 
         text = re.sub(r"\bNPERS\b", "", text)
@@ -321,7 +323,7 @@ class MultilingualLibrispeech(datasets.GeneratorBasedBuilder):
             end_time = float(turn.get('endTime'))
             text = re.sub(r"[\r\n\s]+", " ", MultilingualLibrispeech.extract_text(turn).strip())
             text = self.clean_text(text)
-            if text:
+            if any(c.isalnum() for c in text):
                 utts.append(Utterance(
                     speaker=speaker,
                     sentence=text,
@@ -337,7 +339,7 @@ class MultilingualLibrispeech(datasets.GeneratorBasedBuilder):
         Open an audio file and read as mono waveform, resampling as necessary
         Parameters
         ----------
-        file:The audio file to read
+        file:vThe audio file to read
         sr: int
             The sample rate to resample the audio if necessary
         Returns
@@ -361,6 +363,7 @@ class MultilingualLibrispeech(datasets.GeneratorBasedBuilder):
         return audio[int(round(start_timestamp * SAMPLING_RATE)): int(round(end_timestamp * SAMPLING_RATE)) + 1]
 
     def _generate_examples(self, transcripts_paths, audio_files):
+        """Generate examples from a Multilingual LibriSpeech data dir."""
         for path, file in transcripts_paths:
             file_name = os.path.splitext(os.path.basename(path))[0][:-2]
             audio_path = f"{file_name}.mp4"
@@ -368,14 +371,13 @@ class MultilingualLibrispeech(datasets.GeneratorBasedBuilder):
             for utterance in self.load_one(file):
                 if not self.config.overlap and utterance.overlap:
                     continue
-                if len(utterance.sentence.strip())>0:
-                    yield f"{file_name}_{utterance.start_timestamp}-{utterance.end_timestamp}", {
-                        "file": file_name,
-                        "sentence": utterance.sentence,
-                        "start_timestamp": utterance.start_timestamp,
-                        "end_timestamp": utterance.end_timestamp,
-                        "speaker": utterance.speaker,
-                        "overlap": utterance.overlap,
-                        "audio": {"path": audio_path,
-                                  "array": self._cut_audio(audio, utterance.start_timestamp, utterance.end_timestamp),
-                                  "sampling_rate": 16000}}
+                yield f"{file_name}_{utterance.start_timestamp}-{utterance.end_timestamp}", {
+                    "file": file_name,
+                    "sentence": utterance.sentence,
+                    "start_timestamp": utterance.start_timestamp,
+                    "end_timestamp": utterance.end_timestamp,
+                    "speaker": utterance.speaker,
+                    "overlap": utterance.overlap,
+                    "audio": {"path": audio_path,
+                              "array": self._cut_audio(audio, utterance.start_timestamp, utterance.end_timestamp),
+                              "sampling_rate": 16000}}