andrei-stoica
/
ai_sandbox
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

file response streamlining

This commit is contained in:
Andrei Stoica 2024-02-25 17:09:35 -05:00
parent ebcfa7e19e
commit b7787be635
2 changed files with 11 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

1
.gitignore vendored
View File

@ -2,3 +2,4 @@
**/__pycache__/ **/__pycache__/
**/audio **/audio
*.mp3 *.mp3
*.webm

24
speech-speech/backend/api.py
View File

@ -3,12 +3,10 @@ from fastapi import FastAPI, File, Response, Request
from fastapi.middleware.cors import CORSMiddleware from fastapi.middleware.cors import CORSMiddleware
from fastapi.responses import FileResponse from fastapi.responses import FileResponse
from pydantic import BaseModel from pydantic import BaseModel
import whisper
app = FastAPI() app = FastAPI()
openAI_clinet = OpenAI() openAI_clinet = OpenAI()
model = whisper.load_model("base")
app.add_middleware( app.add_middleware(
CORSMiddleware, CORSMiddleware,
@ -29,16 +27,14 @@ class Conversation(BaseModel):
@app.post("/get-text") @app.post("/get-text")
def stt(audio: bytes = File()): def stt(audio: bytes = File()):
with open("audio", "wb") as f: with open("audio.webm", "wb+") as f:
f.write(audio) f.write(audio)
# transcript = openAI_clinet.audio.transcriptions.create( transcript = openAI_clinet.audio.transcriptions.create(
# model="whisper-1", model="whisper-1",
# file=audio, file=f,
# response_format="text", response_format="text",
# RequestBody )
# ) data = {"len": len(audio), "user-transcript": transcript}
result = model.transcribe("audio")
data = {"len": len(audio), "user-transcript": result["text"]}
return data return data
@ -58,13 +54,13 @@ async def get_next_response(request: Request):
return {"role": role, "content": res_msg} return {"role": role, "content": res_msg}
@app.get("/speak", response_class=FileResponse) @app.get("/speak")
def tts(text: str): def tts(text: str):
res = openAI_clinet.audio.speech.create( res = openAI_clinet.audio.speech.create(
model="tts-1", model="tts-1",
voice="nova", voice="nova",
input=text, input=text,
response_format='mp3'
) )
# this works for now but I need to find a way to stream this to response # this works for now but I need to find a way to stream this to response
res.stream_to_file("tts.mp3") return Response(content=res.content, media_type="audio/mp3")
return "tts.mp3"