app.py

from flask import Flask, request, jsonify
import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
from sklearn.metrics import mean_squared_error
from math import sqrt
import urllib.parse
import pickle

app = Flask(__name__)

@app.route('/')
def hello_world():
    return 'Hello, World!'

@app.route('/api/predictions', methods=['GET'])
def predict_comments():

    page_likes = float(request.args.get('page_likes'))
    page_checkins = float(request.args.get('page_checkins'))
    page_talking_about = float(request.args.get('page_talking_about'))
    page_category = float(request.args.get('page_category'))
    derived_5 = float(request.args.get('derived_5'))
    derived_6 = float(request.args.get('derived_6'))
    derived_7 = float(request.args.get('derived_7'))
    derived_8 = float(request.args.get('derived_8'))
    derived_9 = float(request.args.get('derived_9'))
    derived_10 = float(request.args.get('derived_10'))
    derived_11 = float(request.args.get('derived_11'))
    derived_12 = float(request.args.get('derived_12'))
    derived_13 = float(request.args.get('derived_13'))
    derived_14 = float(request.args.get('derived_14'))
    derived_15 = float(request.args.get('derived_15'))
    derived_16 = float(request.args.get('derived_16'))
    derived_17 = float(request.args.get('derived_17'))
    derived_18 = float(request.args.get('derived_18'))
    derived_19 = float(request.args.get('derived_19'))
    derived_20 = float(request.args.get('derived_20'))
    derived_21 = float(request.args.get('derived_21'))
    derived_22 = float(request.args.get('derived_22'))
    derived_23 = float(request.args.get('derived_23'))
    derived_24 = float(request.args.get('derived_24'))
    derived_25 = float(request.args.get('derived_25'))
    derived_26 = float(request.args.get('derived_26'))
    derived_27 = float(request.args.get('derived_27'))
    derived_28 = float(request.args.get('derived_28'))
    derived_29 = float(request.args.get('derived_29'))
    post_published_weekday_1 = float(request.args.get('post_published_weekday_1'))
    post_published_weekday_2 = float(request.args.get('post_published_weekday_2'))
    post_published_weekday_3 = float(request.args.get('post_published_weekday_3'))
    post_published_weekday_4 = float(request.args.get('post_published_weekday_4'))
    post_published_weekday_5 = float(request.args.get('post_published_weekday_5'))
    post_published_weekday_6 = float(request.args.get('post_published_weekday_6'))
    post_published_weekday_7 = float(request.args.get('post_published_weekday_7'))
    base_datetime_weekday_1 = float(request.args.get('base_datetime_weekday_1'))
    base_datetime_weekday_2 = float(request.args.get('base_datetime_weekday_2'))
    base_datetime_weekday_3 = float(request.args.get('base_datetime_weekday_3'))
    base_datetime_weekday_4 = float(request.args.get('base_datetime_weekday_4'))
    base_datetime_weekday_5 = float(request.args.get('base_datetime_weekday_5'))
    base_datetime_weekday_6 = float(request.args.get('base_datetime_weekday_6'))
    base_datetime_weekday_7 = float(request.args.get('base_datetime_weekday_7'))
    cc1 = float(request.args.get('cc1'))
    cc2 = float(request.args.get('cc2'))
    cc3 = float(request.args.get('cc3'))
    cc4 = float(request.args.get('cc4'))
    cc5 = float(request.args.get('cc5'))
    base_time = float(request.args.get('base_time'))
    post_length = float(request.args.get('post_length'))
    post_share_count = float(request.args.get('post_share_count'))
    post_promotion_status = float(request.args.get('post_promotion_status'))
    h_local = float(request.args.get('h_local'))

    data = {
        'Page Popularity/likes': [page_likes],
        'Page Checkins': [page_checkins],
        'Page talking about': [page_talking_about],
        'Page Category': [page_category],
        'Derived_5': [derived_5],
        'Derived_6': [derived_6],
        'Derived_7': [derived_7],
        'Derived_8': [derived_8],
        'Derived_9': [derived_9],
        'Derived_10': [derived_10],
        'Derived_11': [derived_11],
        'Derived_12': [derived_12],
        'Derived_13': [derived_13],
        'Derived_14': [derived_14],
        'Derived_15': [derived_15],
        'Derived_16': [derived_16],
        'Derived_17': [derived_17],
        'Derived_18': [derived_18],
        'Derived_19': [derived_19],
        'Derived_20': [derived_20],
        'Derived_21': [derived_21],
        'Derived_22': [derived_22],
        'Derived_23': [derived_23],
        'Derived_24': [derived_24],
        'Derived_25': [derived_25],
        'Derived_26': [derived_26],
        'Derived_27': [derived_27],
        'Derived_28': [derived_28],
        'Derived_29': [derived_29],
        'Post published weekday_1': [post_published_weekday_1],
        'Post published weekday_2': [post_published_weekday_2],
        'Post published weekday_3': [post_published_weekday_3],
        'Post published weekday_4': [post_published_weekday_4],
        'Post published weekday_5': [post_published_weekday_5],
        'Post published weekday_6': [post_published_weekday_6],
        'Post published weekday_7': [post_published_weekday_7],
        'Base DateTime weekday_1': [base_datetime_weekday_1],
        'Base DateTime weekday_2': [base_datetime_weekday_2],
        'Base DateTime weekday_3': [base_datetime_weekday_3],
        'Base DateTime weekday_4': [base_datetime_weekday_4],
        'Base DateTime weekday_5': [base_datetime_weekday_5],
        'Base DateTime weekday_6': [base_datetime_weekday_6],
        'Base DateTime weekday_7': [base_datetime_weekday_7],
        'CC1': [cc1],
        'CC2': [cc2],
        'CC3': [cc3],
        'CC4': [cc4],
        'CC5': [cc5],
        'Base time': [base_time],
        'Post length': [post_length],
        'Post Share Count': [post_share_count],
        'Post Promotion Status': [post_promotion_status],
        'H Local': [h_local]
    }
    df = pd.DataFrame(data)
    print(df.head())
    print("Preprocessing the data...")
    # Preprocess the data
    df = preprocessor.transform(df)

    print("Making predictions...")
    prediction = model.predict(df)[0]

    response = {'prediction': prediction}
    print(f"Prediction: {prediction}")
    return jsonify(response)

def generate_prediction_url(row, base_url="http://127.0.0.1:5000/api/predictions"):

    # Replace column names by the endpoint parameter names

    mapping = {
        'Page Popularity/likes': 'page_likes',
        'Page Checkins': 'page_checkins',
        'Page talking about': 'page_talking_about',
        'Page Category': 'page_category',
        'Derived_5': 'derived_5',
        'Derived_6': 'derived_6',
        'Derived_7': 'derived_7',
        'Derived_8': 'derived_8',
        'Derived_9': 'derived_9',
        'Derived_10': 'derived_10',
        'Derived_11': 'derived_11',
        'Derived_12': 'derived_12',
        'Derived_13': 'derived_13',
        'Derived_14': 'derived_14',
        'Derived_15': 'derived_15',
        'Derived_16': 'derived_16',
        'Derived_17': 'derived_17',
        'Derived_18': 'derived_18',
        'Derived_19': 'derived_19',
        'Derived_20': 'derived_20',
        'Derived_21': 'derived_21',
        'Derived_22': 'derived_22',
        'Derived_23': 'derived_23',
        'Derived_24': 'derived_24',
        'Derived_25': 'derived_25',
        'Derived_26': 'derived_26',
        'Derived_27': 'derived_27',
        'Derived_28': 'derived_28',
        'Derived_29': 'derived_29',
        'Post published weekday_1': 'post_published_weekday_1',
        'Post published weekday_2': 'post_published_weekday_2',
        'Post published weekday_3': 'post_published_weekday_3',
        'Post published weekday_4': 'post_published_weekday_4',
        'Post published weekday_5': 'post_published_weekday_5',
        'Post published weekday_6': 'post_published_weekday_6',
        'Post published weekday_7': 'post_published_weekday_7',
        'Base DateTime weekday_1': 'base_datetime_weekday_1',
        'Base DateTime weekday_2': 'base_datetime_weekday_2',
        'Base DateTime weekday_3': 'base_datetime_weekday_3',
        'Base DateTime weekday_4': 'base_datetime_weekday_4',
        'Base DateTime weekday_5': 'base_datetime_weekday_5',
        'Base DateTime weekday_6': 'base_datetime_weekday_6',
        'Base DateTime weekday_7': 'base_datetime_weekday_7',
        'CC1': 'cc1',
        'CC2': 'cc2',
        'CC3': 'cc3',
        'CC4': 'cc4',
        'CC5': 'cc5',
        'Base time': 'base_time',
        'Post length': 'post_length',
        'Post Share Count': 'post_share_count',
        'Post Promotion Status': 'post_promotion_status',
        'H Local': 'h_local'
    }

    row = {mapping[k]: row[k] for k in row.keys()}

    params = urllib.parse.urlencode(row)



    return f"{base_url}?{params}"




if __name__ == '__main__':

    # Load the model and preprocessor
    with open('model.pkl', 'rb') as f:
        model = pickle.load(f)

    with open('preprocessor.pkl', 'rb') as f:
        preprocessor = pickle.load(f)

    
    # Load the dataset
    df = pd.read_csv('Dataset/Testing/TestSet/Test_Case_5.csv')

    # Define column names
    columns = [
        'Page Popularity/likes', 'Page Checkins', 'Page talking about', 'Page Category',
        'Derived_5', 'Derived_6', 'Derived_7', 'Derived_8', 'Derived_9', 'Derived_10',
        'Derived_11', 'Derived_12', 'Derived_13', 'Derived_14', 'Derived_15', 'Derived_16',
        'Derived_17', 'Derived_18', 'Derived_19', 'Derived_20', 'Derived_21', 'Derived_22',
        'Derived_23', 'Derived_24', 'Derived_25', 'Derived_26', 'Derived_27', 'Derived_28','Derived_29',
        'CC1', 'CC2', 'CC3', 'CC4', 'CC5', 'Base time', 'Post length', 'Post Share Count',
        'Post Promotion Status', 'H Local', 'Post published weekday_1', 'Post published weekday_2',
        'Post published weekday_3', 'Post published weekday_4', 'Post published weekday_5',
        'Post published weekday_6', 'Post published weekday_7', 'Base DateTime weekday_1',
        'Base DateTime weekday_2', 'Base DateTime weekday_3', 'Base DateTime weekday_4',
        'Base DateTime weekday_5', 'Base DateTime weekday_6', 'Base DateTime weekday_7',
        'Target Variable'
    ]

    # Rename columns
    df.columns = columns

    for row in df.sample(10).to_dict(orient="records"):
        target = row["Target Variable"]
        del row["Target Variable"]
        print(generate_prediction_url(row))
        print(f"Actual: {target}")
        print()

    app.run(debug=True)