AWS Website Monitor

aws_website_monitor.py

import json
import requests
import boto3
import os

bucketname='webcrawlertempdata'
urltocheck='https://www.google.com'
s3bucketsrn='arn:aws:sns:eu-central-1:783377464316:WebsiteTracker'


def fileExists(filename):
    client = boto3.client('s3')
    results = client.list_objects(Bucket=bucketname, Prefix=filename)
    return 'Contents' in results

def getTempFile(filename):
    s3 = boto3.resource('s3')    
    obj = s3.Object(bucketname, filename)
    print(obj.get())
    return obj.get()['Body'].read().decode()

def saveTempFile(content, filename):
    s3 = boto3.resource('s3')    
    object = s3.Object(
        bucket_name=bucketname, 
        key=filename
    )
    object.put(Body=content)
    print("temp file saved to: " + filename)

def getWebpage(url):
    headers = {'User-Agent': 'Mozilla/5.0 (Macintosh; Intel Mac OS X 10_10_1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/39.0.2171.95 Safari/537.36',
    'Pragma': 'no-cache', 'Cache-Control': 'no-cache', 'Connection':'close'}
    response = requests.get(url, headers=headers, timeout=1)
    return response.text

    

def sendNotification(changed, result):
    client = boto3.client('sns')
    snsArn = s3bucketsrn
    message = "Website Changed " + str(changed)
    print("Sending notification " + str(changed))

    response = client.publish(
        TargetArn = snsArn,
        Message = message ,
        Subject='Website Tracker'
    )




def lambda_handler(event, context):
    url = urltocheck
    filename = "website.txt"
    print("Starting checking "+ url)

    webpageContent = getWebpage(url)
    
    if not fileExists(filename):
        saveTempFile(webpageContent,filename)
        return {
            'statusCode': 200,
            'body': json.dumps('First run, skipping...')
        }
        
    oldContent = getTempFile(filename)
    result = False
    if (oldContent == webpageContent):
        print("nothing changed...")
    else:
        result = True
        print("Webpage changed!")
        saveTempFile(webpageContent,filename)


    print("Ergebnis: ", result)
    if result:
        sendNotification(result,webpageContent)
        return {
            'statusCode': 200,
            'body': json.dumps('Webpage has changed!')
        }
    else:
        return {
            'statusCode': 200,
            'body': json.dumps('Webpage has not changed!')
        }