how do I make API request after deploying uvicorn via fastapi in mylaptop

Question

I am new to FastAPI/model deployment, I managed to created a model which can return a result using the FastAPI docs path in my local server/laptop.

My question is, how can i access this model in jupyter notebook?

Trying to relate into other API request such as Foursquare API. How do I wrap a params dictionary

Tried the following however it is not working

import requests

url = 'http://127.0.0.1:8000/predict'
params = { 'title' : ['Product Operations Software Engineer (DevOps / SRE)']}

result = requests.post(url, params)

<Response [422]>
result.text
'{"detail":[{"loc":["body"],"msg":"value is not a valid dict","type":"type_error.dict"}]}'

Input

Output

My Code

from  fastapi import FastAPI
import numpy as np
import dill
import pandas as pd
import uvicorn
import re
from pydantic import BaseModel

app = FastAPI()

class JobTitle(BaseModel):
    title: list


#load model
with open("logisticmodel", "rb") as to_load:
    logistic_model = dill.load(to_load)

@app.get("/")
def home():
    return 'welcome'

@app.post("/predict")
def predict(request: JobTitle):
    
    X = pd.Series(request.title)
    result = logistic_model.predict(X)
    return {'Type' : list(result) }

Is your server on another device? Then, AFAIK, you can't use `http://127.0.0.1` to refer to it. You'll have to use either external IP, or local network IP (if connected to the same network). — TheNightHawk, Nov 28 '21 at 07:04
Does this answer your question? [How to post JSON data to FastAPI and retrieve the JSON data inside the endpoint?](https://stackoverflow.com/questions/70975344/how-to-post-json-data-to-fastapi-and-retrieve-the-json-data-inside-the-endpoint) — Chris, Oct 23 '22 at 08:19
Related answers can also be found [here](https://stackoverflow.com/a/70636163/17865804) and [here](https://stackoverflow.com/a/73761724/17865804). — Chris, Oct 23 '22 at 08:20

score 1 · Accepted Answer · answered Nov 30 '21 at 08:21

Your code here indicates that your data should be put inside the HTTP request body

@app.post("/predict")
def predict(request: JobTitle):
    ...

But when you issue the HTTP request with the requests module,
you didn't put anything to your request body.
You should use json argument so the data can be put to the right place:

result = requests.post(
    url,
    json={'title': ['Product Operations Software Engineer (DevOps / SRE)']}
)

You can check this for sending an HTTP request with request body using requests module,
and this for building an API with a request body.

score 0 · Answer 2 · answered Nov 28 '21 at 08:17

0

You need to use .dict() to get the dictionary of BaseModel instance. In your case, request.dict()

Also, Define your model like this,

class JobTitle(BaseModel):
    title: List[str]

answered Nov 28 '21 at 08:17

Mohammad Ashraful Islam

1,879
15
29

score 0 · Answer 3 · answered Nov 29 '21 at 15:00

0

This works:

>>> result = requests.post(url, json=params)
>>> result
<Response [200]>

answered Nov 29 '21 at 15:00

Shod

801
3
12
33

how do I make API request after deploying uvicorn via fastapi in mylaptop

3 Answers3