H2O_FastApi.py

import pandas as pd
import io
import h2o
import streamlit as st

import mlflow
import mlflow.h2o
from mlflow.tracking import MlflowClient
from mlflow.entities import ViewType

h2o.init()
client = MlflowClient()

@st.cache()

class NotANumber(Exception):
    def __init__(self, message = "Values entered are not Numerical"):
        self.message = message
        super().__init__(self.message)

def predict(data):
    # Load best model
    exper_list = [ex.experiment_id for ex in client.list_experiments()]
    runs_list = mlflow.search_runs(experiment = exper_list, 
                                run_view_type = ViewType.ALL)

    run_id = runs_list.loc[runs_list['metrics.log_loss'].idxmin()]['run_id']
    exp_id = runs_list.loc[runs_list['metrics.log_loss'].idxmin()]['experiment_id']

    print(f'Best Model: Run {run_id} in Experiment {exp_id}')
    bst_mdl = mlflow.h2o.load_model(f'mlruns/{exp_id}/{run_id}/artifacts/best_model/model.h2o/')
    prediction = bst_mdl.predict(data).tolist()[0]
    return prediction   

def validate_input(dict_request):
    for _, val in dict_request.items():
        try:
            val = float(val)
        except Exception as e:
            raise NotANumber

    return True

def form_response(dict_request):
    try:
        if validate_input(dict_request):
            data = dict_request.values()
            data = [list(map(float, data))]
            response = predict(data)
            return response
    except:
        print('Not a Number')


def main():       
    # front end elements of the web page 
    html_temp = """ 
    <div style ="background-color:yellow;padding:13px"> 
    <h1 style ="color:black;text-align:center;">Churn Prediction ML App</h1> 
    </div> 
    """
      
    # display the front end aspect
    st.markdown(html_temp, unsafe_allow_html = True) 
      
    # following lines create boxes in which user can enter data required to make prediction 
    vmail_msg = st.slider('Number vmail messages',min_value = 1, max_value = 30)
    tot_day_calls = st.slider('Total day calls',min_value = 1, max_value = 30)
    tot_eve_min = st.slider('Total eve minutes',min_value = 1, max_value = 30) 
    tot_eve_chr = st.slider('Total eve charge',min_value = 1, max_value = 30)
    tot_int_min = st.slider('Total Intl minutes',min_value = 1, max_value = 30)
    cust_sev_calls = st.slider('Customer service calls',min_value = 1, max_value = 30)
    dict_request = {'vmail_msg': vmail_msg, 
                'tot_day_calls': tot_day_calls, 
                'tot_eve_min': tot_eve_min, 
                'tot_eve_chr': tot_eve_chr, 
                'tot_int_min': tot_int_min, 
                'cust_sev_calls': cust_sev_calls}
    result =""
      
    # when 'Predict' is clicked, make the prediction and store it 
    if st.button("Predict"): 
        result =  form_response(dict_request)
        st.success('Prediction is {}'.format(result))
     
if __name__=='__main__': 
    main()