Initial commit

README.md

@@ -0,0 +1,168 @@
+---
+annotations_creators:
+- expert-generated
+language:
+- zh
+language_creators:
+- crowdsourced
+license:
+- cc-by-nc-nd-4.0
+multilinguality:
+- monolingual
+pretty_name: MAGICDATA_Mandarin_Chinese_Read_Speech_Corpus
+size_categories:
+- 10K<n<100K
+source_datasets:
+- original
+tags: []
+task_categories:
+- automatic-speech-recognition
+task_ids: []
+---
+
+# Dataset Card for MMCRSC
+
+## Table of Contents
+- [Table of Contents](#table-of-contents)
+- [Dataset Description](#dataset-description)
+  - [Dataset Summary](#dataset-summary)
+  - [Supported Tasks and Leaderboards](#supported-tasks-and-leaderboards)
+  - [Languages](#languages)
+- [Dataset Structure](#dataset-structure)
+  - [Data Instances](#data-instances)
+  - [Data Fields](#data-fields)
+  - [Data Splits](#data-splits)
+- [Dataset Creation](#dataset-creation)
+  - [Curation Rationale](#curation-rationale)
+  - [Source Data](#source-data)
+  - [Annotations](#annotations)
+  - [Personal and Sensitive Information](#personal-and-sensitive-information)
+- [Considerations for Using the Data](#considerations-for-using-the-data)
+  - [Social Impact of Dataset](#social-impact-of-dataset)
+  - [Discussion of Biases](#discussion-of-biases)
+  - [Other Known Limitations](#other-known-limitations)
+- [Additional Information](#additional-information)
+  - [Dataset Curators](#dataset-curators)
+  - [Licensing Information](#licensing-information)
+  - [Citation Information](#citation-information)
+  - [Contributions](#contributions)
+
+## Dataset Description
+
+- **Homepage:** [MAGICDATA Mandarin Chinese Read Speech Corpus](https://openslr.org/68/)
+- **Repository:**
+- **Paper:** 
+- **Leaderboard:**
+- **Point of Contact:** 
+
+### Dataset Summary
+
+MAGICDATA Mandarin Chinese Read Speech Corpus was developed by MAGIC DATA Technology Co., Ltd. and freely published for non-commercial use.
+The contents and the corresponding descriptions of the corpus include:
+
+The corpus contains 755 hours of speech data, which is mostly mobile recorded data.
+1080 speakers from different accent areas in China are invited to participate in the recording.
+The sentence transcription accuracy is higher than 98%.
+Recordings are conducted in a quiet indoor environment.
+The database is divided into training set, validation set, and testing set in a ratio of 51: 1: 2.
+Detail information such as speech data coding and speaker information is preserved in the metadata file.
+The domain of recording texts is diversified, including interactive Q&A, music search, SNS messages, home command and control, etc.
+Segmented transcripts are also provided.
+The corpus aims to support researchers in speech recognition, machine translation, speaker recognition, and other speech-related fields. Therefore, the corpus is totally free for academic use.
+The corpus is a subset of a much bigger data ( 10566.9 hours Chinese Mandarin Speech Corpus ) set which was recorded in the same environment. Please feel free to contact us via business@magicdatatech.com for more details.
+
+### Supported Tasks and Leaderboards
+
+[More Information Needed]
+
+### Languages
+
+zh-CN
+
+## Dataset Structure
+
+### Data Instances
+
+```json
+{
+  'file': '14_3466_20170826171404.wav',
+  'audio': {
+    'path': '14_3466_20170826171404.wav',
+    'array': array([0., 0., 0., ..., 0., 0., 0.]),
+    'sampling_rate': 16000
+  },
+  'text': '请搜索我附近的超市',
+  'speaker_id': 143466,
+  'id': '14_3466_20170826171404.wav'
+}
+```
+
+### Data Fields
+
+- file: A path to the downloaded audio file in .wav format.
+- audio: A dictionary containing the path to the downloaded audio file, the decoded audio array, and the sampling rate. Note that when accessing the audio column: `dataset[0]["audio"]` the audio file is automatically decoded and resampled to `dataset.features["audio"].sampling_rate`. Decoding and resampling of a large number of audio files might take a significant amount of time. Thus it is important to first query the sample index before the `"audio"` column, *i.e.* `dataset[0]["audio"]` should **always** be preferred over `dataset["audio"][0]`.
+- text: the transcription of the audio file.
+- id: unique id of the data sample.
+- speaker_id: unique id of the speaker. The same speaker id can be found for multiple data samples.
+
+### Data Splits
+
+[More Information Needed]
+
+## Dataset Creation
+
+### Curation Rationale
+
+[More Information Needed]
+
+### Source Data
+
+#### Initial Data Collection and Normalization
+
+[More Information Needed]
+
+#### Who are the source language producers?
+
+[More Information Needed]
+
+### Annotations
+
+#### Annotation process
+
+[More Information Needed]
+
+#### Who are the annotators?
+
+[More Information Needed]
+
+### Personal and Sensitive Information
+
+[More Information Needed]
+
+## Considerations for Using the Data
+
+### Social Impact of Dataset
+
+[More Information Needed]
+
+### Discussion of Biases
+
+[More Information Needed]
+
+### Other Known Limitations
+
+[More Information Needed]
+
+## Additional Information
+
+### Dataset Curators
+
+[More Information Needed]
+
+### Licensing Information
+
+[More Information Needed]
+
+### Citation Information
+
+Please cite the corpus as "Magic Data Technology Co., Ltd., "http://www.imagicdatatech.com/index.php/home/dataopensource/data_info/id/101", 05/2019".

mmcrsc.py

@@ -0,0 +1,157 @@
+# 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
+"""MAGICDATA Mandarin Chinese Read Speech Corpus."""
+
+
+import os
+
+import datasets
+from datasets.tasks import AutomaticSpeechRecognition
+
+
+_CITATION = """\
+ @misc{magicdata_2019, 
+ title={MAGICDATA Mandarin Chinese Read Speech Corpus}, 
+ url={https://openslr.org/68/}, 
+ publisher={Magic Data Technology Co., Ltd.},  
+ year={2019}, 
+ month={May}} 
+"""
+
+_DESCRIPTION = """\
+The corpus by Magic Data Technology Co., Ltd. , containing 755 hours of scripted read speech data 
+from 1080 native speakers of the Mandarin Chinese spoken in mainland China. 
+The sentence transcription accuracy is higher than 98%.
+"""
+
+_URL = "https://openslr.org/68/"
+_DL_URL = "http://www.openslr.org/resources/68/"
+
+
+_DL_URLS = {
+  "train": _DL_URL + "train_set.tar.gz",
+  "dev": _DL_URL + "dev_set.tar.gz",
+  "test": _DL_URL + "test_set.tar.gz",
+}
+
+
+class MMCRSCConfig(datasets.BuilderConfig):
+  """BuilderConfig for MMCRSC."""
+
+  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.
+    """
+    # version history
+    # 0.1.0: First release on Huggingface
+    super(MMCRSCConfig, self).__init__(version=datasets.Version("0.1.0", ""), **kwargs)
+
+
+class MMCRSC(datasets.GeneratorBasedBuilder):
+  """MMCRSC dataset."""
+
+  DEFAULT_WRITER_BATCH_SIZE = 256
+  DEFAULT_CONFIG_NAME = "all"
+
+  def _info(self):
+    return datasets.DatasetInfo(
+      description=_DESCRIPTION,
+      features=datasets.Features(
+        {
+          "file": datasets.Value("string"),
+          "audio": datasets.Audio(sampling_rate=16_000),
+          "text": datasets.Value("string"),
+          "speaker_id": datasets.Value("int64"),
+          "id": 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(_DL_URLS)
+    # (Optional) In non-streaming mode, we can extract the archive locally to have actual local audio files:
+    local_extracted_archive = dl_manager.extract(archive_path) if not dl_manager.is_streaming else {}
+
+    return  [
+      datasets.SplitGenerator(
+        name=datasets.Split.TRAIN,
+        gen_kwargs={
+          "local_extracted_archive": local_extracted_archive.get("train"),
+          "files": dl_manager.iter_archive(archive_path["train"]),
+        },
+      ),
+      datasets.SplitGenerator(
+        name=datasets.Split.VALIDATION,
+        gen_kwargs={
+          "local_extracted_archive": local_extracted_archive.get("dev"),
+          "files": dl_manager.iter_archive(archive_path["dev"]),
+        },
+      ),
+      datasets.SplitGenerator(
+        name=datasets.Split.TEST,
+        gen_kwargs={
+          "local_extracted_archive": local_extracted_archive.get("test"),
+          "files": dl_manager.iter_archive(archive_path["test"]),
+        },
+      ),
+    ]
+
+  def _generate_examples(self, files, local_extracted_archive):
+    """Generate examples from a LibriSpeech archive_path."""
+    audio_data = {}
+    transcripts = []
+    for path, f in files:
+      if path.endswith(".wav"):
+        id_ = path.split("/")[-1]
+        audio_data[id_] = f.read()
+      elif path.endswith("TRANS.txt"):
+        for line in f:
+          if line and (b'.wav' in line):
+            line = line.decode("utf-8").strip()
+            id_, speaker_id, transcript = line.split("\t")
+            audio_file = id_
+            audio_file = (
+              os.path.join(local_extracted_archive, audio_file)
+              if local_extracted_archive
+              else audio_file
+            )
+            transcripts.append(
+              {
+                "id": id_,
+                "speaker_id": speaker_id,
+                "file": audio_file,
+                "text": transcript,
+              }
+            )
+    if audio_data:
+      for key, transcript in enumerate(transcripts):
+        if transcript["id"] in audio_data:
+          audio = {"path": transcript["file"], "bytes": audio_data[transcript["id"]]}
+          yield key, {"audio": audio, **transcript}
+      audio_data = {}
+      transcripts = []
+
+