Example PCHome crawler script for AWS Lambda

PCHome_crawler.py

import numpy as np
import pandas as pd
import requests
import bs4
import lxml.etree as xml
import time
from sqlalchemy import create_engine
import pymysql
import pymssql

def pchome_crawler(keyword):
    query_res_arr = []
    for page_number in range(100):
        URL = 'https://ecshweb.pchome.com.tw/search/v3.3/all/results?q=' + keyword + '&page=' + str(page_number)
        res = requests.get(URL)
        try:
            query_res = res.json()
            query_res_arr.append(query_res['prods'])
            time.sleep(5)
        except:
            print(page_number)
            pass
        
    res_arr = []
    for page in query_res_arr:
        res_arr = res_arr + page

    return res_arr

def db_insert(res_arr):
    df = pd.DataFrame(res_arr)
    del df['couponActid']
    engine = create_engine("mysql+pymysql://admin:[email protected]:3306/test", encoding='utf-8', echo=True)
    with engine.begin() as connection:
        df.to_sql('pchome_result', con=connection, if_exists='append')
    return pass

def lambda_handler(event, context):
    keyword = event['pathParameters']['keyword'] or event['queryString']['keyword']
    try:
        res_arr = pchome_crawler(keyword)
        db_insert(res_arr)
    except:
        pass
    
    return {
        'statusCode': 200,
        'body': json.dumps('PCHome crawler: '+keyword+'('+len(res_arr)+') success.')
    }