Web Scraping

Beginning Python Programming

Makzan, 2020 April.

Web Scraping

  • Downloading files in Python
  • Open web URL
  • Fetch data with API
  • Web scraping with Requests and BeautifulSoup
  • Web automation with Selenium
  • Converting Wikipedia tabular data into CSV

Code Snippet: Download file

'''Download chart from AAStock server with given stock numbers.'''

from urllib.request import urlretrieve

stock_numbers = ['0011','0005','0001','0700','3333','0002','0012']

for stock_number in stock_numbers:
    url = "http://charts.aastocks.com/servlet/Charts?fontsize=12&15MinDelay=T&lang=1&titlestyle=1&vol=1&Indicator=1&indpara1=10&indpara2=20&indpara3=50&indpara4=100&indpara5=150&subChart1=2&ref1para1=14&ref1para2=0&ref1para3=0&subChart2=3&ref2para1=12&ref2para2=26&ref2para3=9&subChart3=12&ref3para1=0&ref3para2=0&ref3para3=0&scheme=3&com=100&chartwidth=660&chartheight=855&stockid=00{}.HK&period=6&type=1&logoStyle=1".format(stock_number)
    urlretrieve(url, '{}-chart.gif'.format(stock_number))

Example: Open Web URL

import webbrowser

query = input("Please input search query to search Python doc. ")

url = "https://docs.python.org/3/search.html?q=" + query + "&check_keywords=yes&area=default"

webbrowser.open(url)

Example: Open Google map

import webbrowser
import urllib.parse

query = input("Please input search query to search near-by Macao. ")

# A map search in Macao.
url = f"https://www.google.com/maps/search/{query}/@22.1612464,113.5303786,13z"

webbrowser.open(url)

Example: SMG.gov.mo

import untangle
import datetime


def fetch():
    obj = untangle.parse('http://xml.smg.gov.mo/c_actual_brief.xml')

    temperature = obj.ActualWeatherBrief.Custom.Temperature.Value.cdata
    humidity = obj.ActualWeatherBrief.Custom.Humidity.Value.cdata

    print("現時澳門氣溫 " + temperature + " 度，濕度 " + humidity + "%。")

fetch()

Fetching XML

import untangle
import datetime

year = datetime.date.today().year

# list begins at 0, and we look for previous month.
last_month = datetime.date.today().month -1 -1

if last_month < 0:
    year = year - 1
    last_month = 11 # list beings at 0.

url = f"http://www.dicj.gov.mo/web/cn/information/DadosEstat_mensal/{year}/report_cn.xml?id=8"

data = untangle.parse(url)

month_data = data.STATISTICS.REPORT.DATA.RECORD[last_month]

month = month_data.DATA[0].cdata
net_income = month_data.DATA[1].cdata
last_net_income = month_data.DATA[2].cdata
change_rate = month_data.DATA[3].cdata
acc_net_income = month_data.DATA[4].cdata
acc_last_net_income = month_data.DATA[5].cdata
acc_change_rate = month_data.DATA[6].cdata

print(f"{month} 毛收入 {net_income} ({year-1}:{last_net_income}), {change_rate}")
print(f"{month} 累計毛收入 {acc_net_income} ({year-1}:{acc_last_net_income}), {acc_change_rate}")

Example: 博彩月計毛收入

Fetching XML

Example: Exchange rate

import json
import requests

url = "https://api.exchangeratesapi.io/latest?symbols=HKD&base=CNY"

response = requests.get(url)
data = json.loads(response.text)
print(data)

print(data['rates']['HKD'])

Fetching JSON

Web Scraping

  1. Querying web page
  2. Parse the DOM tree
  3. Get the data we want from the HTML code

Fetching news title

from bs4 import BeautifulSoup
import requests

try:
    res = requests.get("https://news.gov.mo/home/zh-hant")
except requests.exceptions.ConnectionError:
    print("Error: Invalid URL")
    exit()


soup = BeautifulSoup(res.text, "html.parser")

print("\n".join( map(lambda x: x.getText().strip(), soup.select("h5"))) )

Web Scraping

Fetching gov.mo holiday

import requests
from bs4 import BeautifulSoup

response = requests.get("https://www.gov.mo/zh-hant/public-holidays/year-2020/")
soup = BeautifulSoup(response.text, "html.parser")

tables = soup.select(".table")

for row in tables[0].select("tr"):
  if len(row.select("td")) > 0:
    is_obligatory = (row.select("td")[0].text == "*")
    if is_obligatory:
      date = row.select("td")[1].text
      name = row.select("td")[3].text
      print(f"{date}: {name}")

Web Scraping

Is today government holiday?

import requests
from bs4 import BeautifulSoup
import datetime

# Get today's year, month and day
today = datetime.date.today()
year = today.year
month = today.month
day = today.day
today_weekday = today.weekday()
today_date = f"{month}月{day}日"


# Fetch gov.mo
url = f"https://www.gov.mo/zh-hant/public-holidays/year-{year}/"
response = requests.get(url)
soup = BeautifulSoup(response.text, "html.parser")

tables = soup.select(".table")

holidays = {}

for table in tables:
  for row in table.select("tr"):
    if len(row.select("td")) > 0:    
      date = row.select("td")[1].text
      weekday = row.select("td")[2].text
      name = row.select("td")[3].text
      holidays[date] = name


# Query holidays
print(today_date)
if today_date in holidays:
  holiday = holidays[today_date]
  print(f"今天是公眾假期：{holiday}")
elif today_weekday == 0:
  print("今天是星期日，但不是公眾假期。")
elif today_weekday == 6:
  print("今天是星期六，但不是公眾假期。")  
else:
  print("今天不是公眾假期。")

Web Scraping

Is today obligatory holiday?

import requests
from bs4 import BeautifulSoup
import datetime

# Get today's year, month and day
today = datetime.date.today()
year = today.year
month = today.month
day = today.day
today_weekday = today.weekday()
today_date = f"{month}月{day}日"


# Fetch gov.mo
url = f"https://www.gov.mo/zh-hant/public-holidays/year-{year}/"
response = requests.get(url)
soup = BeautifulSoup(response.text, "html.parser")

tables = soup.select(".table")

holidays = {}

for table in tables:
  for row in table.select("tr"):
    if len(row.select("td")) > 0:    
      is_obligatory = (row.select("td")[0].text == "*")
      date = row.select("td")[1].text
      weekday = row.select("td")[2].text
      name = row.select("td")[3].text
      holidays[date] = {
        'date': date,
        'weekday': weekday,
        'name': name,
        'is_obligatory': is_obligatory,
      }


# Query holidays
print(today_date)
if today_date in holidays:
  holiday = holidays[today_date]
  if holiday['is_obligatory']:
    print(f"今天是強制公眾假期：{holiday['name']}")
  else:
    print(f"今天是公眾假期：{holiday['name']}")
elif today_weekday == 0:
  print("今天是星期日，但不是公眾假期。")
elif today_weekday == 6:
  print("今天是星期六，但不是公眾假期。")  
else:
  print("今天不是公眾假期。")

Web Scraping

Web Automation with Selenium

  • We use Selenium when:
    • When Requests and BeautifulSoup does not work.
    • When page requires JavaScript to render the data.
  • Pros:
    • It launches real browser and automate browser.
    • Better compatibility .
  • Cons:
    • Slow because it launches real browser.

Selenium Cheat Sheet

  • https://codoid.com/selenium-webdriver-python-cheat-sheet/

Downloading Browser Driver

Take screenshot

'''Capture the screenshot of a website via Headless Firefox.'''

from selenium import webdriver
from selenium.webdriver.firefox.options import Options

options = Options()
options.add_argument('-headless')

browser = webdriver.Firefox(options=options)
browser.maximize_window()
browser.get('https://makclass.com/')
browser.save_screenshot('makclass.png')
browser.quit()

Web Automation with Selenium

Filling in forms

from selenium import webdriver
from selenium.common.exceptions import NoSuchElementException
from selenium.webdriver.firefox.options import Options
from selenium.webdriver.support.ui import Select
import time

url = 'https://niioa.immigration.gov.tw/NIA_OnlineApply_inter/visafreeApply/visafreeApplyForm.action'

visitor = {
    'chinese_name': '陳大文',
    'last_name': 'Chan',
    'first_name': 'Tai Man',
    'passport_no': 'MA0012345',
    'gender': 'm',
    'birthday': '1986/10/20',
}

def fill_input(selector, value):
    '''find the input with given sselector and fill in the value.'''
    return browser.find_element_by_css_selector(selector).send_keys(value)

options = Options()
# options.add_argument('-headless')

browser = webdriver.Firefox(options=options)
browser.maximize_window()

browser.get(url)

radio = browser.find_element_by_css_selector('#isHKMOVisaN')
radio.click()

browser.execute_script('$.unblockUI();')

time.sleep(0.5)

select = Select(browser.find_element_by_css_selector('#tsk01Type'))
# print(select.select_by_value.__doc__)
select.select_by_value('1')

fill_input('#chineseName', visitor['chinese_name'])
fill_input('#englishName1', visitor['last_name'])
fill_input('#englishName2', visitor['first_name'])
fill_input('#passportNo', visitor['passport_no'])
fill_input('#birthDateStr', visitor['birthday'])

# ... and more fields to fill in ...

if visitor['gender'] == 'm':
    browser.find_element_by_css_selector('#gender0').click()
else:
    browser.find_element_by_css_selector('#gender1').click()

Web Automation with Selenium

Example: aastock (v1)

'''Fetch current stock from aastock.'''

from selenium import webdriver
from selenium.webdriver.firefox.options import Options
import time

stock_number = '0011'

options = Options()
options.add_argument('-headless')

browser = webdriver.Firefox(options=options)
browser.maximize_window()

browser.get('http://www.aastocks.com/')
element = browser.find_element_by_css_selector('#txtHKQuote')
element.send_keys(stock_number)
browser.execute_script("shhkquote($('#txtHKQuote').val(), 'quote', mainPageMarket)")

time.sleep(3)

element = browser.find_element_by_css_selector('.font28')
print(element.text)


browser.quit()

Web Automation with Selenium

Example: aastock (v2)

'''Fetch current stock from aastock.'''

from selenium import webdriver
from selenium.common.exceptions import NoSuchElementException
from selenium.webdriver.firefox.options import Options
import time

stock_number = '0011'

options = Options()
options.add_argument('-headless')

browser = webdriver.Firefox(options=options)
browser.maximize_window()

browser.get('http://www.aastocks.com/tc/stocks/aboutus/companyinfo.aspx')
element = browser.find_element_by_css_selector('#txtHKQuote')
element.send_keys(stock_number)
browser.execute_script("shhkquote($('#txtHKQuote').val(), 'quote', mainPageMarket)")

trial_time = 0
while True:
    if trial_time > 5:
        break
    to_sleep = 0.1
    time.sleep(to_sleep)
    trial_time += to_sleep
    try:
        element = browser.find_element_by_css_selector('.font28')
        print(element.text)
        break
    except NoSuchElementException as e:
        pass



browser.quit()

Web Automation with Selenium

Example: aastock (v3)

'''Fetch stock price of given list of stock number from aastock.'''

from selenium import webdriver
from selenium.common.exceptions import NoSuchElementException
from selenium.webdriver.firefox.options import Options
import time

stock_numbers = '0011,0005,0700,1810'

options = Options()
# options.add_argument('-headless')

browser = webdriver.Firefox(options=options)
browser.maximize_window()

browser.get('http://www.aastocks.com/tc/stocks/aboutus/companyinfo.aspx')

for stock_number in stock_numbers.split(','):
    #print(stock_number)
    element = browser.find_element_by_css_selector('#txtHKQuote')
    element.clear()
    element.send_keys(stock_number)
    browser.execute_script("shhkquote($('#txtHKQuote').val(), 'quote', mainPageMarket)")

    trial_time = 0
    while True:
        if trial_time > 10:
            break
        to_sleep = 0.1
        time.sleep(to_sleep)
        trial_time += to_sleep
        try:
            element = browser.find_element_by_css_selector('.font28')
            print(stock_number + ": " + element.text)
            break
        except NoSuchElementException as e:
            pass



browser.quit()

Web Automation with Selenium

Example: aastock (v4)

'''Fetch current stock from aastock.'''

from selenium import webdriver
from selenium.common.exceptions import NoSuchElementException
from selenium.webdriver.firefox.options import Options
from selenium.webdriver.common.keys import Keys
import time

stock_numbers_input = input("Please enter a list of number separated by comma, in 4 digit: ")

stock_numbers = stock_numbers_input.split(",")

print(stock_numbers)

options = Options()
options.add_argument('-headless')

browser = webdriver.Firefox(options=options)
browser.maximize_window()

for stock_number in stock_numbers:

    browser.get('http://www.aastocks.com/tc/mobile/MwFeedback.aspx')
    element = browser.find_element_by_css_selector('#symbol')
    element.send_keys(stock_number)
    element.send_keys(Keys.RETURN);

    trial_time = 0
    while True:
        if trial_time > 5:
            break
        to_sleep = 0.1
        time.sleep(to_sleep)
        trial_time += to_sleep
        try:
            element = browser.find_element_by_css_selector('.text_last')
            print(stock_number)
            print(element.text)
            break
        except NoSuchElementException as e:
            pass



browser.quit()

input("Press enter to exit")

Web Automation with Selenium

Convert Wiki Tables to CSV

https://wikitable2csv.ggor.de/

Summary

Web Scraping

  • Downloading files in Python
  • Open web URL
  • Fetch data with API
  • Web scraping with Requests and BeautifulSoup
  • Web automation with Selenium
  • Converting Wikipedia tabular data into CSV

Python: Web Scraping

By makzan

Made with Slides.com

Python: Web Scraping

  • 1,207

More from makzan