ting11222001

import requests
from bs4 import BeautifulSoup as bs

#NBA網址
res = requests.get('https://www.ptt.cc/bbs/NBA/index.html')

#用lxml’s HTML parser直接解析html的tag
soup = bs(res.text,'lxml')

#建議看完就把它comment掉省記憶體

ting11222001

#載入這次需要的所有套件和function
from selenium import webdriver
from selenium.webdriver.common.keys import Keys
import time

#安裝chrome WebDriver瀏覽器驅動程式
#https://chromedriver.chromium.org/downloads (注意先確認自己的chrome版本)
#下載到自己指定的存放位置，並解壓縮
chrome_driver_path = '填入自己的下載檔的檔案位置'

ting11222001

#https://technews.tw/
#爬取內容請以當日網頁為準
#第一階段
import requests
from bs4 import BeautifulSoup as bs
import json

res = requests.get('https://technews.tw/')
soup = bs(res.text,'lxml')

ting11222001

#載入套件
#API: https://www.dcard.tw/service/api/v2/forums
import pandas as pd
import requests
r = requests.get('https://www.dcard.tw/service/api/v2/forums')
response = r.text 

import json
data = json.loads(response)
#是線上api所以可以直接以json型態下載

ting11222001

#載入selenium套件
from selenium import webdriver
from selenium.webdriver.common.keys import Keys
import time

#我的chromedriver路徑
chrome_driver_path = '/Users/ting11222001/Downloads/chromedriver'
#第一個頁面的url
url = 'https://gogakuru.com/english/phrase/genre/180_%E5%88%9D%E7%B4%9A%E3%83%AC%E3%83%99%E3%83%AB.html?layoutPhrase=1&orderPhrase=1&condMovie=0&flow=enSearchGenre&condGenre=180&perPage=50'

ting11222001

# !pip  install  PyMySQL 
import pymysql 

#連結自己架在Virtual Box (Centos 7) 的mariadb
conn = pymysql.connect(host = '192.168.56.123', 
                       user = 'tiffany', 
                       passwd = "admin123", 
                       db = 'demo_2020_pyscrap')

#創一個cursor物件，是在資料庫操作的游標

ting11222001

# import libraries
import numpy as np
import matplotlib.pyplot as plt 

import pandas as pd  
import seaborn as sns 

%matplotlib inline

# import dataset from: https://scikit-learn.org/stable/modules/generated/sklearn.datasets.load_boston.html

ting11222001

%matplotlib notebook
import numpy as np
import matplotlib.pyplot as plt
from skimage.feature import match_template
from PIL import Image

ting11222001

# convert image into an array and ndim: 3
ImagenTotal = np.asarray(Image.open('./redmangos.jpg'))

# print(ImagenTotal.ndim) 
#output: 3

# print(ImagenTotal.shape) 
#output: (480, 640, 3)

# print(len(ImagenTotal)) 

ting11222001

# Interactive selection of points
# store all clicked coordinates here 
puntosinteres = []

# Create a new figure: width, height in inches
fig = plt.figure(figsize=(5, 4))

# position figure to 1st row, 1st column, index = 1
ax = fig.add_subplot(111)