Run FastAPI inside Jupyter

fastapi_inside_jupyter.md

How to Run FastAPI inside Jupyter

• Ensure you have these installed and accessible from the notebook pyngrok, nest_asyncio, fastapi, uvicorn and other libs you want

%pip install pyngrok nest_asyncio fastapi uvicorn loguru

• Create a FastAPI app

from fastapi import FastAPI
from pydantic import BaseModel
from loguru import logger

app = FastAPI()

class UserRequestIn(BaseModel):
    text: str
    
@app.post("/test")
def index(request: UserRequestIn):
    logger.debug(request)
    return {"ok": True}

• Start ngrok tunnel

from pyngrok import ngrok

ngrok_tunnel = ngrok.connect(8000)

ngrok_tunnel

• Patch Event Loop and start server

import nest_asyncio
import uvicorn

nest_asyncio.apply()
uvicorn.run(app, port=8000)

• Click on the ngrok url printed above, and visit /docs endpoint and you will have the familiar Swagger page and call your api's from there

it works! thank you !
can you explain the rule of each raw, such as :

ngrok_tunnel = ngrok.connect(8000)
ngrok_tunnel - most important to understand !

nest_asyncio.apply()
uvicorn.run(app, port=8000)

can one specify all the imports at the top or it has to be written as you did?

thanks in advance!

i think ngrok can start earlier than fastapi
you can do all the imports at the top as well, dont see an issue with that