Upload 5 files

backend/app.py

@@ -0,0 +1,29 @@
+from fastapi import FastAPI, UploadFile
+from fastapi.responses import FileResponse
+from backend.whisper_utils import transcribe_audio
+from backend.gtts_utils import generate_speech
+from backend.llm_utils import get_llm_response
+import os
+
+app = FastAPI()
+
+@app.post("/transcribe/")
+async def transcribe(file: UploadFile):
+    file_path = f"audio/{file.filename}"
+    with open(file_path, "wb") as audio:
+        audio.write(await file.read())
+
+    text = transcribe_audio(file_path)
+    os.remove(file_path)  # Cleanup audio file
+    return {"transcription": text}
+
+@app.post("/response/")
+async def get_response(input_text: str):
+    llm_response = get_llm_response(input_text)
+    audio_path = generate_speech(llm_response)
+    return {"response": llm_response, "audio_url": audio_path}
+
+@app.get("/audio/{file_name}")
+async def serve_audio(file_name: str):
+    file_path = f"audio/{file_name}"
+    return FileResponse(file_path)

backend/gtts_utils.py

@@ -0,0 +1,10 @@
+from gtts import gTTS
+import os
+import uuid
+
+def generate_speech(text: str) -> str:
+    file_name = f"{uuid.uuid4()}.mp3"
+    file_path = f"audio/{file_name}"
+    tts = gTTS(text)
+    tts.save(file_path)
+    return file_path

backend/llm_utils.py

@@ -0,0 +1,8 @@
+from transformers import pipeline
+
+# Load the Hugging Face LLM
+llm = pipeline("text-generation", model="gpt2", max_length=100)
+
+def get_llm_response(prompt: str) -> str:
+    response = llm(prompt)
+    return response[0]["generated_text"]

backend/requirements.txt

@@ -0,0 +1,5 @@
+fastapi
+uvicorn
+openai-whisper
+transformers
+gtts

backend/whisper_utils.py

@@ -0,0 +1,7 @@
+import whisper
+
+model = whisper.load_model("base")
+
+def transcribe_audio(file_path: str) -> str:
+    result = model.transcribe(file_path)
+    return result["text"]