Revert Convert dataset to Parquet

README.md

@@ -1,4 +1,5 @@
 ---
+pretty_name: Arabic Speech Corpus
 annotations_creators:
 - expert-generated
 language_creators:
@@ -9,6 +10,7 @@ license:
 - cc-by-4.0
 multilinguality:
 - monolingual
+paperswithcode_id: arabic-speech-corpus
 size_categories:
 - 1K<n<10K
 source_datasets:
@@ -16,10 +18,22 @@ source_datasets:
 task_categories:
 - automatic-speech-recognition
 task_ids: []
-paperswithcode_id: arabic-speech-corpus
-pretty_name: Arabic Speech Corpus
+train-eval-index:
+- config: clean
+  task: automatic-speech-recognition
+  task_id: speech_recognition
+  splits:
+    train_split: train
+    eval_split: test
+  col_mapping:
+    file: path
+    text: text
+  metrics:
+  - type: wer
+    name: WER
+  - type: cer
+    name: CER
 dataset_info:
-  config_name: clean
   features:
   - name: file
     dtype: string
@@ -33,38 +47,16 @@ dataset_info:
     dtype: string
   - name: orthographic
     dtype: string
+  config_name: clean
   splits:
   - name: train
-    num_bytes: 1527815416.966
+    num_bytes: 1002365
     num_examples: 1813
   - name: test
-    num_bytes: 99851729.0
+    num_bytes: 65784
     num_examples: 100
-  download_size: 1347643373
-  dataset_size: 1627667145.966
-configs:
-- config_name: clean
-  data_files:
-  - split: train
-    path: clean/train-*
-  - split: test
-    path: clean/test-*
-  default: true
-train-eval-index:
-- config: clean
-  task: automatic-speech-recognition
-  task_id: speech_recognition
-  splits:
-    train_split: train
-    eval_split: test
-  col_mapping:
-    file: path
-    text: text
-  metrics:
-  - type: wer
-    name: WER
-  - type: cer
-    name: CER
+  download_size: 1192302846
+  dataset_size: 1068149
 ---
 
 # Dataset Card for Arabic Speech Corpus

arabic_speech_corpus.py

@@ -0,0 +1,145 @@
+# coding=utf-8
+# Copyright 2021 The TensorFlow Datasets Authors and the HuggingFace Datasets Authors.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# Lint as: python3
+"""Arabic Speech Corpus"""
+
+
+import os
+
+import datasets
+from datasets.tasks import AutomaticSpeechRecognition
+
+
+_CITATION = """\
+@phdthesis{halabi2016modern,
+  title={Modern standard Arabic phonetics for speech synthesis},
+  author={Halabi, Nawar},
+  year={2016},
+  school={University of Southampton}
+}
+"""
+
+_DESCRIPTION = """\
+This Speech corpus has been developed as part of PhD work carried out by Nawar Halabi at the University of Southampton.
+The corpus was recorded in south Levantine Arabic
+(Damascian accent) using a professional studio. Synthesized speech as an output using this corpus has produced a high quality, natural voice.
+Note that in order to limit the required storage for preparing this dataset, the audio
+is stored in the .flac format and is not converted to a float32 array. To convert, the audio
+file to a float32 array, please make use of the `.map()` function as follows:
+
+
+```python
+import soundfile as sf
+
+def map_to_array(batch):
+    speech_array, _ = sf.read(batch["file"])
+    batch["speech"] = speech_array
+    return batch
+
+dataset = dataset.map(map_to_array, remove_columns=["file"])
+```
+"""
+
+_URL = "http://en.arabicspeechcorpus.com/arabic-speech-corpus.zip"
+
+
+class ArabicSpeechCorpusConfig(datasets.BuilderConfig):
+    """BuilderConfig for ArabicSpeechCorpu."""
+
+    def __init__(self, **kwargs):
+        """
+        Args:
+          data_dir: `string`, the path to the folder containing the files in the
+            downloaded .tar
+          citation: `string`, citation for the data set
+          url: `string`, url for information about the data set
+          **kwargs: keyword arguments forwarded to super.
+        """
+        super(ArabicSpeechCorpusConfig, self).__init__(version=datasets.Version("2.1.0", ""), **kwargs)
+
+
+class ArabicSpeechCorpus(datasets.GeneratorBasedBuilder):
+    """ArabicSpeechCorpus dataset."""
+
+    BUILDER_CONFIGS = [
+        ArabicSpeechCorpusConfig(name="clean", description="'Clean' speech."),
+    ]
+
+    def _info(self):
+        return datasets.DatasetInfo(
+            description=_DESCRIPTION,
+            features=datasets.Features(
+                {
+                    "file": datasets.Value("string"),
+                    "text": datasets.Value("string"),
+                    "audio": datasets.Audio(sampling_rate=48_000),
+                    "phonetic": datasets.Value("string"),
+                    "orthographic": datasets.Value("string"),
+                }
+            ),
+            supervised_keys=("file", "text"),
+            homepage=_URL,
+            citation=_CITATION,
+            task_templates=[AutomaticSpeechRecognition(audio_column="audio", transcription_column="text")],
+        )
+
+    def _split_generators(self, dl_manager):
+        archive_path = dl_manager.download_and_extract(_URL)
+        archive_path = os.path.join(archive_path, "arabic-speech-corpus")
+        return [
+            datasets.SplitGenerator(name="train", gen_kwargs={"archive_path": archive_path}),
+            datasets.SplitGenerator(name="test", gen_kwargs={"archive_path": os.path.join(archive_path, "test set")}),
+        ]
+
+    def _generate_examples(self, archive_path):
+        """Generate examples from a Librispeech archive_path."""
+        lab_dir = os.path.join(archive_path, "lab")
+        wav_dir = os.path.join(archive_path, "wav")
+        if "test set" in archive_path:
+            phonetic_path = os.path.join(archive_path, "phonetic-transcript.txt")
+        else:
+            phonetic_path = os.path.join(archive_path, "phonetic-transcipt.txt")
+
+        orthographic_path = os.path.join(archive_path, "orthographic-transcript.txt")
+
+        phonetics = {}
+        orthographics = {}
+
+        with open(phonetic_path, "r", encoding="utf-8") as f:
+            for line in f:
+                wav_file, phonetic = line.split('"')[1::2]
+                phonetics[wav_file] = phonetic
+
+        with open(orthographic_path, "r", encoding="utf-8") as f:
+            for line in f:
+                wav_file, orthographic = line.split('"')[1::2]
+                orthographics[wav_file] = orthographic
+
+        for _id, lab_name in enumerate(sorted(os.listdir(lab_dir))):
+            lab_path = os.path.join(lab_dir, lab_name)
+            lab_text = open(lab_path, "r", encoding="utf-8").read()
+
+            wav_name = lab_name[:-4] + ".wav"
+            wav_path = os.path.join(wav_dir, wav_name)
+
+            example = {
+                "file": wav_path,
+                "audio": wav_path,
+                "text": lab_text,
+                "phonetic": phonetics[wav_name],
+                "orthographic": orthographics[wav_name],
+            }
+            yield str(_id), example

clean/test-00000-of-00001.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:bd5ea889532615c4ca9c63b5b83fc3bacb94e9fa156c26f5963b8da2c8e87768
-size 90899095

clean/train-00000-of-00004.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:9c3f2931ab19224daf55126c1cf96ff068f3ad442d760c1f5db99805d5a290be
-size 398895011

clean/train-00001-of-00004.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:d02e7e7d080082d1d96929b83e19b924d7c10c8b59a39f190c373245559ea36d
-size 322764456

clean/train-00002-of-00004.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:8ea3385f7d8496bf1e77d9b1a2696fb2bb3769e1ffa060e43fa4fc6c5e25cf06
-size 291793854

clean/train-00003-of-00004.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:91bbec3487d3ba745113c5869be40a6008ef815b9681fe683cf7ab46dd06efcf
-size 243290957