commit

speech_io.py

@@ -1,18 +1,37 @@
+import os
+from tempfile import NamedTemporaryFile
+from typing import Optional
+from openai import OpenAI
+
+client = OpenAI(api_key=os.getenv("OPENAI_API_KEY"))
+
+
+def transcribe_audio(file_path: str, language: Optional[str] = None) -> str:
+    print(">>> Transkribiere Audio via OpenAI Audio API …")
+
+    with open(file_path, "rb") as f:
+        resp = client.audio.transcriptions.create(
+            model="gpt-4o-mini-transcribe",
+            file=f,
+            language=language
+        )
+
+    return resp.text
+
+
 def synthesize_speech(text: str, voice: str = "alloy") -> str:
     print(">>> Synthesizing speech via OpenAI TTS …")
 
-    # OpenAI SDK 2.x on HF Spaces returns HttpxBinaryResponseContent
+    # OpenAI SDK 2.x returns HttpxBinaryResponseContent
     response = client.audio.speech.create(
         model="gpt-4o-mini-tts",
         voice=voice,
         input=text
     )
 
-    # 🔥 THIS is the correct way:
-    # Get raw audio bytes directly from the response object
+    # Correct extraction method
     audio_bytes = response.read()
 
-    # Save file
     tmp = NamedTemporaryFile(delete=False, suffix=".mp3")
     tmp.write(audio_bytes)
     tmp.close()