Create utils.py

utils.py

@@ -0,0 +1,60 @@
+from io import BytesIO
+import base64
+import numpy as np
+import subprocess
+import soundfile as sf
+from speech_recognition import AudioFile, Recognizer
+
+
+def ffmpeg_read(bpayload: bytes, sampling_rate: int) -> np.array:
+    """
+    Helper function to read an audio file through ffmpeg.
+    """
+    ar = f"{sampling_rate}"
+    ac = "1"
+    format_for_conversion = "f32le"
+    ffmpeg_command = [
+        "ffmpeg",
+        "-i",
+        "pipe:0",
+        "-ac",
+        ac,
+        "-ar",
+        ar,
+        "-f",
+        format_for_conversion,
+        "-hide_banner",
+        "-loglevel",
+        "quiet",
+        "pipe:1",
+    ]
+
+    try:
+        ffmpeg_process = subprocess.Popen(ffmpeg_command, stdin=subprocess.PIPE, stdout=subprocess.PIPE)
+    except FileNotFoundError:
+        raise ValueError("ffmpeg was not found but is required to load audio files from filename")
+    output_stream = ffmpeg_process.communicate(bpayload)
+    out_bytes = output_stream[0]
+    audio = np.frombuffer(out_bytes, np.float32)
+    sf.write('temp.wav', audio, sampling_rate, subtype='PCM_16')
+    return 'temp.wav'
+
+
+def stt(audio: object, language='ja') -> str:
+    """Converts speech to text.
+    Args:
+        audio: record of user speech
+        language (str): language of text
+    Returns:
+        text (str): recognized speech of user
+    """
+    # Create a Recognizer object
+    r = Recognizer()
+    # Open the audio file
+    with AudioFile(audio) as source:
+        # Listen for the data (load audio to memory)
+        audio_data = r.record(source)
+        # Transcribe the audio using Google's speech-to-text API
+        text = r.recognize_google(audio_data, language=language)
+    return text
+