predictnow 2.1.0

A restful client library, designed to access predictnow restful API.

predictnow-api

TO BEGIN ANY WORK WITH PREDICTNOW.AI CLIENT, WE START BY IMPORTING AND CREATING A CLASS INSTANCE

from predictnow.pdapi import PredictNowClient
import pandas as pd

api_key = "KeyProvidedToEachOfOurSubscriber"   
api_host = "http://%VMIP%"  

# Initial variables
username = "user1"  
email = "[e-mail]"
client = PredictNowClient(api_host,api_key)

YOU WILL NEED TO EDIT THIS INPUT DATASET FILE PATH, LABELNAME AND MODELNAME!

file_path = 'my_amazing_features.xlsx'  
labelname = 'futreturn' #might need to change this name accordingly 
modelname = 'model1' # 
import os

NOW YOUR PREDICTNOW.AI CLIENT HAS BEEN SETUP.

# For classification problem
params = {"timeseries": "yes", "weights": "no", "prob_calib": "no", "eda": "no", "type": "classification", "feature_selection": "shap", "analysis": "small", "boost": "gbdt", "mode": "train", "testsize": "1"}

# For regression problems
params = {"timeseries": "yes", "weights": "no", "prob_calib": "no", "eda": "no", "type": "regression", "feature_selection": "shap", "analysis": "small", "boost": "gbdt", "mode": "train", "testsize": "1"}

print("THE PARAMS", params)

LET'S CREATE THE MODEL BY SENDING THE PARAMETERS TO PREDICTNOW.AI

response = client.create_model
                   (
                    username=username, # only letters, numbers, or underscores
                    model_name=modelname,
                    params=params,
                   )

print(response)

LET'S LOAD UP THE FILE TO PANDAS IN THE LOCAL ENVIRONMENT

from pandas import read_csv  # If you have the Excel file, replace read_csv with read_excel
from pandas import read_excel
df = read_excel(file_path)  # Same here
df.name = "testdataframe"  # Optional, but recommended

print(df)

START TRAINING MODEL

NOTE: THIS MAY TAKE UP TO several minutes

response = client.train
                    (
                        model_name=modelname,
                        input_df=df,
                        label=labelname,
                        username=username,
                        email=email,
                        return_output=False
                    )

print("THE CLIENT HAS SENT THE DATASET TO THE SERVER AND TRIGGERED THE TRAINING MODEL TASK")
print(response)

CHECK THE STATUS OF THE MODEL

status = client.getstatus(
                            username=username,
                            train_id=response["train_id"]
                         )

print("Current status:")
print(status)

NOW WE WILL DOWNLOAD FILES

if status["state"] == "COMPLETED":

    response = client.getresult(
        model_name=modelname,
        username=username,
    )

    import pandas as pd
    predicted_prob_cv = pd.read_json(response.predicted_prob_cv)
    print("predicted_prob_cv")
    print(predicted_prob_cv)

    predicted_prob_test = pd.read_json(response.predicted_prob_test)
    print("predicted_prob_test")
    print(predicted_prob_test)


    predicted_targets_cv = pd.read_json(response.predicted_targets_cv)
    print("predicted_targets_cv")
    print(predicted_targets_cv)

START PREDICTING USING THE TRAINED MODEL

if status["state"] == "COMPLETED":

    df = read_excel("example_input_live_latest.xlsx")
    df.name = "myfirstpredictname"  # optional, but recommended

    # Predict demo
    response = client.predict(
        model_name=modelname,
        input_df=df,
        username=username,
        eda=params["eda"],
        prob_calib=params["prob_calib"]
    )