feat: script

speech-emotion-recognition-dataset.py

@@ -1,37 +1,35 @@
+from pathlib import Path
+
 import datasets
+import numpy as np
+import pandas as pd
 import PIL.Image
 import PIL.ImageOps
-import numpy as np
 
 _CITATION = """\
 @InProceedings{huggingface:dataset,
-title = {generated-usa-passeports-dataset},
+title = {speech-emotion-recognition-dataset},
 author = {TrainingDataPro},
 year = {2023}
 }
 """
 
 _DESCRIPTION = """\
-Data generation in machine learning involves creating or manipulating data
-to train and evaluate machine learning models. The purpose of data generation
-is to provide diverse and representative examples that cover a wide range of
-scenarios, ensuring the model's robustness and generalization.
-Data augmentation techniques involve applying various transformations to
-existing data samples to create new ones. These transformations include:
-random rotations, translations, scaling, flips, and more. Augmentation helps
-in increasing the dataset size, introducing natural variations, and improving
-model performance by making it more invariant to specific transformations.
-The dataset contains **GENERATED** USA passports, which are replicas of
-official passports but with randomly generated details, such as name, date of
-birth etc. The primary intention of generating these fake passports is to
-demonstrate the structure and content of a typical passport document and to
-train the neural network to identify this type of document.
-Generated passports can assist in conducting research without accessing or
-compromising real user data that is often sensitive and subject to privacy
-regulations. Synthetic data generation allows researchers to develop and
-refine models using simulated passport data without risking privacy leaks.
+The audio dataset consists of a collection of texts spoken with four distinct
+emotions. These texts are spoken in English and represent four different
+emotional states: **euphoria, joy, sadness and surprise**.
+Each audio clip captures the tone, intonation, and nuances of speech as
+individuals convey their emotions through their voice.
+The dataset includes a diverse range of speakers, ensuring variability in age,
+gender, and cultural backgrounds*, allowing for a more comprehensive
+representation of the emotional spectrum.
+The dataset is labeled and organized based on the emotion expressed in each
+audio sample, making it a valuable resource for emotion recognition and
+analysis. Researchers and developers can utilize this dataset to train and
+evaluate machine learning models and algorithms, aiming to accurately
+recognize and classify emotions in speech.
 """
-_NAME = 'generated-usa-passeports-dataset'
+_NAME = 'speech-emotion-recognition-dataset'
 
 _HOMEPAGE = f"https://huggingface.co/datasets/TrainingDataPro/{_NAME}"
 
@@ -40,108 +38,77 @@ _LICENSE = "cc-by-nc-nd-4.0"
 _DATA = f"https://huggingface.co/datasets/TrainingDataPro/{_NAME}/resolve/main/data/"
 
 
-def exif_transpose(img):
-    if not img:
-        return img
-
-    exif_orientation_tag = 274
-
-    # Check for EXIF data (only present on some files)
-    if hasattr(img, "_getexif") and isinstance(
-            img._getexif(), dict) and exif_orientation_tag in img._getexif():
-        exif_data = img._getexif()
-        orientation = exif_data[exif_orientation_tag]
-
-        # Handle EXIF Orientation
-        if orientation == 1:
-            # Normal image - nothing to do!
-            pass
-        elif orientation == 2:
-            # Mirrored left to right
-            img = img.transpose(PIL.Image.FLIP_LEFT_RIGHT)
-        elif orientation == 3:
-            # Rotated 180 degrees
-            img = img.rotate(180)
-        elif orientation == 4:
-            # Mirrored top to bottom
-            img = img.rotate(180).transpose(PIL.Image.FLIP_LEFT_RIGHT)
-        elif orientation == 5:
-            # Mirrored along top-left diagonal
-            img = img.rotate(-90,
-                             expand=True).transpose(PIL.Image.FLIP_LEFT_RIGHT)
-        elif orientation == 6:
-            # Rotated 90 degrees
-            img = img.rotate(-90, expand=True)
-        elif orientation == 7:
-            # Mirrored along top-right diagonal
-            img = img.rotate(90,
-                             expand=True).transpose(PIL.Image.FLIP_LEFT_RIGHT)
-        elif orientation == 8:
-            # Rotated 270 degrees
-            img = img.rotate(90, expand=True)
-
-    return img
-
-
-def load_image_file(file, mode='RGB'):
-    # Load the image with PIL
-    img = PIL.Image.open(file)
-
-    if hasattr(PIL.ImageOps, 'exif_transpose'):
-        # Very recent versions of PIL can do exit transpose internally
-        img = PIL.ImageOps.exif_transpose(img)
-    else:
-        # Otherwise, do the exif transpose ourselves
-        img = exif_transpose(img)
-
-    img = img.convert(mode)
-
-    return np.array(img)
-
-
-class GeneratedUsaPasseportsDataset(datasets.GeneratorBasedBuilder):
+class SpeechEmotionRecognitionDataset(datasets.GeneratorBasedBuilder):
 
     def _info(self):
-        return datasets.DatasetInfo(
-            description=_DESCRIPTION,
-            features=datasets.Features({
-                'original': datasets.Image(),
-                'us_pass_augmentated_1': datasets.Image(),
-                'us_pass_augmentated_2': datasets.Image(),
-                'us_pass_augmentated_3': datasets.Image()
-            }),
-            supervised_keys=None,
-            homepage=_HOMEPAGE,
-            citation=_CITATION,
-            license=_LICENSE)
+        return datasets.DatasetInfo(description=_DESCRIPTION,
+                                    features=datasets.Features({
+                                        'set_id': datasets.Value('string'),
+                                        'euphoric': datasets.Audio(),
+                                        'joyfully': datasets.Audio(),
+                                        'sad': datasets.Audio(),
+                                        'surprised': datasets.Audio(),
+                                        'text': datasets.Value('string'),
+                                        'gender': datasets.Value('string'),
+                                        'age': datasets.Value('int8'),
+                                        'country': datasets.Value('string')
+                                    }),
+                                    supervised_keys=None,
+                                    homepage=_HOMEPAGE,
+                                    citation=_CITATION,
+                                    license=_LICENSE)
 
     def _split_generators(self, dl_manager):
-        original = dl_manager.download_and_extract(f"{_DATA}original.zip")
-        augmentation = dl_manager.download_and_extract(
-            f"{_DATA}augmentation.zip")
+        audio = dl_manager.download_and_extract(f"{_DATA}audio.zip")
         annotations = dl_manager.download(f"{_DATA}{_NAME}.csv")
-        original = dl_manager.iter_files(original)
-        augmentation = dl_manager.iter_files(augmentation)
+        # audio = dl_manager.iter_files(audio)
         return [
             datasets.SplitGenerator(name=datasets.Split.TRAIN,
                                     gen_kwargs={
-                                        "original": original,
-                                        'augmentation': augmentation,
+                                        "audio": audio,
                                         'annotations': annotations
                                     }),
         ]
 
-    def _generate_examples(self, original, augmentation, annotations):
-        original = list(original)
-        augmentation = list(augmentation)
-        augmentation = [
-            augmentation[i:i + 3] for i in range(0, len(augmentation), 3)
-        ]
+    def _generate_examples(self, audio, annotations):
+        annotations_df = pd.read_csv(annotations, sep=';')
+        audio = list(audio)
+
+        for idx, sub_dir in enumerate(audio):
+            sub_dir = Path(sub_dir)
+            set_id = sub_dir.name
+
+            for audio_file in sub_dir.iterdir():
+                if audio_file.name.startswith('euphoric'):
+                    euphoric = audio_file
+                elif audio_file.name.startswith('joyfully'):
+                    joyfully = audio_file
+                elif audio_file.name.startswith('sad'):
+                    sad = audio_file
+                elif audio_file.name.startswith('surprised'):
+                    surprised = audio_file
 
-        for idx, (org, aug) in enumerate(zip(original, augmentation)):
             yield idx, {
-                'original': load_image_file(org),
-                'us_pass_augmentated_1': load_image_file(aug[0]),
-                'us_pass_augmentated_2': load_image_file(aug[1]),
-                'us_pass_augmentated_3': load_image_file(aug[2])
+                'set_id':
+                    set_id,
+                'euphoric':
+                    euphoric,
+                'joyfully':
+                    joyfully,
+                'sad':
+                    sad,
+                'surprised':
+                    surprised,
+                'text':
+                    annotations_df.loc[annotations_df['set_id'] == set_id]
+                    ['text'].values[0],
+                'gender':
+                    annotations_df.loc[annotations_df['set_id'] == set_id]
+                    ['gender'].values[0],
+                'age':
+                    annotations_df.loc[annotations_df['set_id'] == set_id]
+                    ['age'].values[0],
+                'country':
+                    annotations_df.loc[annotations_df['set_id'] == set_id]
+                    ['country'].values[0]
             }