main.py

from fastapi import FastAPI, Body, Query
from routers import items
import models.classifier as clf
from joblib import load
from models.iris import Iris
from Levenshtein import distance

app = FastAPI()

app.include_router(items.router)

products = [
    {"id": 1, "name": "Iphone"},
    {"id": 2, "name": "Samsung"},
    {"id": 3, "name": "Xiaomi"},
    # ...
]


@app.on_event('startup')
async def load_model():
    clf.model = load('models/iris_dt_v1.joblib')


@app.post('/predict', tags=["predictions"])
async def get_prediction(iris: Iris):
    data = dict(iris)['data']
    prediction = clf.model.predict(data).tolist()

    return {"prediction": prediction}


@app.post('/similarity', tags=["similarity"])
async def get_similarity(q: str = Query(..., min_length=1, max_length=100)):
    similar_products = []

    for product in products:
        dist = distance(q, product["name"])
        if dist <= 5:
            similar_products.append(
                {"id": product["id"], "name": product["name"], "distance": dist})

    return similar_products


@app.get("/")
async def root():
    return {"message": "Hello DataScience Applications!"}