api.py

import pickle
import mlflow
from fastapi import FastAPI
from pydantic import BaseModel
from mlflow import MlflowClient
from dotenv import load_dotenv
import os
import pandas as pd
import xgboost as xgb

load_dotenv(override=True)  # Carga las variables del archivo .env

mlflow.set_tracking_uri("databricks")
client = MlflowClient()

EXPERIMENT_NAME = "/Users/isabel.valladolid@iteso.mx/nyc-taxi-experiments"

run_ = mlflow.search_runs(order_by=['metrics.rmse ASC'],
                          output_format="list",
                          experiment_names=[EXPERIMENT_NAME]
                          )[0]

run_id = run_.info.run_id

run_uri = f"runs:/{run_id}/preprocessor"

client.download_artifacts(
    run_id=run_id,
    path='preprocessor',
    dst_path='.'
)

with open("preprocessor/preprocessor.b", "rb") as f_in:
    dv = pickle.load(f_in)

model_name = "workspace.default.nyc-taxi-model"
alias = "champion"

model_uri = f"models:/{model_name}@{alias}"

champion_model = mlflow.pyfunc.load_model(
    model_uri=model_uri
)

def preprocess(input_data):

    input_dict = {
        'PU_DO': input_data.PULocationID + "_" + input_data.DOLocationID,
        'trip_distance': input_data.trip_distance,
    }
    X = dv.transform([input_dict])

    # Names depend on sklearn version
    try:
        cols = dv.get_feature_names_out()
    except AttributeError:
        cols = dv.get_feature_names()

    # 
    X_df = pd.DataFrame(X.toarray(), columns=cols)

    return X_df

def predict(input_data):

    X_val = preprocess(input_data)

    return champion_model.predict(X_val)

app = FastAPI()

class InputData(BaseModel):
    PULocationID: str
    DOLocationID: str
    trip_distance: float


@app.post("/api/v1/predict")
def predict_endpoint(input_data: InputData):
    result = predict(input_data)[0]
    return {"prediction": float(result)}