vctk dataset

enhancer/data/vctk.py

@@ -1,15 +1,18 @@
 
-from genericpath import isdir
+import glob
-import librosa
+import math
 import os
+from scipy.io import wavfile
 from torch.utils.data import IterableDataset
-import torch
+
+from enhancer.utils.random import create_unique_rng
+from enhancer.utils.io import Audio
 
 
 class Vctk(IterableDataset):
     """Dataset object for Voice Bank Corpus (VCTK) Dataset"""
 
-    def __init__(self,clean_path,noisy_path,sample_length=1,num_samples=None):
+    def __init__(self,clean_path,noisy_path,duration=1,sampling_rate=16000,num_samples=None):
         
         if not os.path.isdir(clean_path):
             raise ValueError(f"{clean_path} is not a valid directory")
@@ -17,22 +20,54 @@ class Vctk(IterableDataset):
         if not os.path.isdir(noisy_path):
             raise ValueError(f"{clean_path} is not a valid directory")
 
+        self.sampling_rate = sampling_rate
         self.clean_path = clean_path
         self.noisy_path = noisy_path
+        self.wav_samples =[file.split('/')[-1] for file in glob.glob(os.path.join(clean_path,"*.wav"))]
 
         if num_samples is None:
-            self.num_samples = len([file for file in os.listdir(clean_path) if file.endswith(".wav")])
+            self.num_samples = len(self.wav_samples)
         else:
             self.num_samples = num_samples
 
-        self.sample_length = max(0.1,sample_length)
+        self.duration = max(1.0,duration)
+        self.audio = Audio(self.sampling_rate,mono=True,return_tensor=True)
+        self.files_duration = self.get_files_duration()
+
+    def get_file_duration(self):
+
+        files_duration = {}
+        for file in self.clean_path:
+            wavfile = wavfile.read(os.path.join(self.clean_path,file),rate=self.sampling_rate)
+            files_duration.update({file:math.ceil(wavfile/self.sampling_rate)})
+
+        return files_duration
+
 
     def __iter__(self):
 
+        rng = create_unique_rng(12) ##pass epoch number here
 
+        while True:
 
+            file_name = rng.choices(self.wav_samples,k=1)
+            file_duration = self.files_duration.get(file_name)
+            start_time = rng.randint(0,math.ceil(file_duration- self.duration))
+            data = self.prepare_segment(file_name,start_time)
+            yield data
 
-        pass
+    def prepare_segment(self,file_name:str, start_time:int):
+
+        clean_segment = self.audio(os.path.join(self.clean_path,file_name),
+                                    offset=start_time,duration=self.duration)
+        noisy_segment = self.audio(os.path.join(self.noisy_path,file_name),
+                                    offset=start_time,duration=self.duration)
+
+        return {"clean": clean_segment,"noisy":noisy_segment}
         
     def __len__(self):
-        pass
+
+        return math.ceil(sum(self.files_duration.values())/self.duration)
+
+
+