1 import matplotlib.pyplot as plt
 2 import re
 3 import requests
 4 from requests import RequestException
 5 import json
 6 
 7 def get_one_page(url):   #獲取網頁
 8     try:
 9         response=requests.get(url)
10         if response.status_code==200:
11             return response.text
12         return
 
 None
13     except RequestException:
14         return None
15 
16 def parse_html(html): #獲取需要的部分
17     pattern=re.compile('<tr.*?<td>(.*?)</td>.*?<td>(.*?)<span.*?</td>.*?<td>(.*?)</td>.*?<td>.*?<img src="(.*?)">.*?</td></tr>',re.S)
18     gifts=re.findall(pattern,html)
 
19     print(gifts)
20     for gift in gifts:
21         yield {   #生成器
22             'itemTitle':gift[0].strip('\n'),
23             'description':gift[1].strip('\n'),
24             'cost':gift[2].strip('\n'),
25             'image':gift[3]
26         }
27 
28 def write_to_file(content): #寫入檔案
29     with open('
 
result.txt','a',encoding='utf-8') as f:  #a表示追加
30         f.write(json.dumps(content,ensure_ascii=False)+'\n')
31 
32 def download_img(): #下載圖片
33     with open('result.txt','r') as f:
34         lines=f.readlines()
35         for line in lines:
36             url='http://www.pythonscraping.com'+json.loads(line).get("image")[2:]
37             r=requests.get(url)
38             with open(url[-8:],'wb') as f1:
39                 f1.write(r.content)
40 
41 def get(item_title): #獲取指定圖片的相關資訊
42     with open('result.txt','r') as f:
43         lines=f.readlines()
44         for line in lines:
45             if json.loads(line).get("itemTitle")==item_title:
46                 return json.loads(line)
47 
48 def drawGift(item_title): #作圖
49     r=get(item_title)
50     img=plt.imread(r.get("image")[-8:])
51     plt.imshow(img)
52     plt.axis('off')
53     plt.title(r.get("itemTitle")+'\n'+r.get("description")+''+r.get("cost"))
54     plt.show()
55 
56 def main():
57     url='http://www.pythonscraping.com/pages/page3.html'
58     html=get_one_page(url)
59     for gift in parse_html(html):
60         write_to_file(gift)
61     download_img()
62 
63 if __name__ == '__main__':
64     item_title="Dead Parrot"
65     main()
66     drawGift(item_title)