first pass at turn based conversation

vocode/turn_based/transcriber/base_transcriber.py

@@ -0,0 +1,6 @@
+from pydub import AudioSegment
+
+
+class BaseTranscriber:
+    def transcribe(self, audio_segment: AudioSegment) -> str:
+        raise NotImplementedError

vocode/turn_based/transcriber/whisper_transcriber.py

@@ -0,0 +1,21 @@
+from pydub import AudioSegment
+import io
+import os
+from dotenv import load_dotenv
+import openai
+
+from vocode.turn_based.transcriber.base_transcriber import BaseTranscriber
+
+load_dotenv()
+
+openai.api_key = os.getenv("OPENAI_API_KEY")
+
+
+class WhisperTranscriber(BaseTranscriber):
+    def transcribe(self, audio_segment: AudioSegment) -> str:
+        in_memory_wav = io.BytesIO()
+        audio_segment.export(in_memory_wav, format="wav")
+        in_memory_wav.seek(0)
+        in_memory_wav.name = "whisper.wav"
+        transcript = openai.Audio.transcribe("whisper-1", in_memory_wav)
+        return transcript.text