Update audio_to_text.py

audio_to_text.py

@@ -1,26 +1,45 @@
 from transformers import WhisperProcessor, WhisperForConditionalGeneration
 import torchaudio
 import torch
+import os
+from pydub import AudioSegment
+
+
+# Get the directory of the current file
+current_dir = os.path.dirname(os.path.abspath(__file__))
+# Construct the absolute path to the 'ffmpeg/bin' directory
+ffmpeg_bin_path = os.path.join(current_dir, 'ffmpeg', 'bin')
+# Add this path to the PATH environment variable
+os.environ["PATH"] += os.pathsep + ffmpeg_bin_path
+# Ensure ffmpeg is in PATH
+AudioSegment.converter = os.path.join(ffmpeg_bin_path, 'ffmpeg.exe')
 
 
 # load model and processor
-processor = WhisperProcessor.from_pretrained("openai/whisper-tiny")
-model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-tiny")
+processor = WhisperProcessor.from_pretrained("openai/whisper-small")
+model = WhisperForConditionalGeneration.from_pretrained("openai/whisper-small")
 model.config.forced_decoder_ids = None
 
 
-def audio_to_text(audio_data,sample_rate):
-    # Convert raw audio frame (numpy array) to tensor and resample it to 16 kHz
-    waveform = torch.tensor(audio_data, dtype=torch.float32).unsqueeze(0)
-    # Check if the sample rate is 16 kHz; if not, resample it
-    if sample_rate != 16000:
+def audio_to_text(webm_file_path):
+    wav_file = "recorded_audio.wav"
+    absolute_path = os.path.abspath(webm_file_path)
+
+    # Load and convert audio
+    # Check if the file exists
+    if os.path.exists(webm_file_path):
+        wav_audio = AudioSegment.from_file(absolute_path, format="webm")
+        wav_audio.export(wav_file, format="wav")
+        # Load the audio and resample it
+        waveform, sample_rate = torchaudio.load('recorded_audio.wav')
         resampler = torchaudio.transforms.Resample(orig_freq=sample_rate, new_freq=16000)
         waveform = resampler(waveform)
-    waveform = waveform.squeeze().numpy()
-    input_features = processor(waveform, sampling_rate=16000, return_tensors="pt").input_features
-
-    # generate token ids
-    predicted_ids = model.generate(input_features)
-    # decode token ids to text
-    transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True)
-    return transcription
+        waveform = waveform.squeeze().numpy()
+        input_features = processor(waveform, sampling_rate=16000, return_tensors="pt").input_features
+        # generate token ids
+        predicted_ids = model.generate(input_features)
+        # decode token ids to text
+        transcription = processor.batch_decode(predicted_ids, skip_special_tokens=True)
+        return transcription
+    else:
+        return None