回答編集履歴
1
edit
test
CHANGED
@@ -11,3 +11,141 @@
|
|
11
11
|
v = products[a]
|
12
12
|
|
13
13
|
```
|
14
|
+
|
15
|
+
|
16
|
+
|
17
|
+
---
|
18
|
+
|
19
|
+
|
20
|
+
|
21
|
+
Selenium.webdriverにバグがあってJANコードがあってもたまに取得にミスるみたいですね。
|
22
|
+
|
23
|
+
|
24
|
+
|
25
|
+
```python
|
26
|
+
|
27
|
+
#!/usr/bin/env python
|
28
|
+
|
29
|
+
# -*- coding: utf-8 -*-
|
30
|
+
|
31
|
+
from selenium import webdriver
|
32
|
+
|
33
|
+
from selenium.common.exceptions import NoSuchElementException
|
34
|
+
|
35
|
+
from selenium.webdriver.common.keys import Keys
|
36
|
+
|
37
|
+
import argparse
|
38
|
+
|
39
|
+
|
40
|
+
|
41
|
+
DELAY_SLEEP = 1
|
42
|
+
|
43
|
+
JANJAN_URL = "https://antlion.xsrv.jp/"
|
44
|
+
|
45
|
+
|
46
|
+
|
47
|
+
def get_input(fmimic=True):
|
48
|
+
|
49
|
+
if fmimic:
|
50
|
+
|
51
|
+
class MimicArgs:
|
52
|
+
|
53
|
+
def __init__(self):
|
54
|
+
|
55
|
+
self.jan = [4976219091275, 4873116953, 4873116950000]
|
56
|
+
|
57
|
+
args = MimicArgs()
|
58
|
+
|
59
|
+
else:
|
60
|
+
|
61
|
+
parser = argparse.ArgumentParser(description='Get asin from JANJAN_URL')
|
62
|
+
|
63
|
+
parser.add_argument('--jan', nargs='+')
|
64
|
+
|
65
|
+
args = parser.parse_args()
|
66
|
+
|
67
|
+
return args
|
68
|
+
|
69
|
+
|
70
|
+
|
71
|
+
def main():
|
72
|
+
|
73
|
+
args = get_input()
|
74
|
+
|
75
|
+
|
76
|
+
|
77
|
+
driver = webdriver.Chrome()
|
78
|
+
|
79
|
+
|
80
|
+
|
81
|
+
products = {}
|
82
|
+
|
83
|
+
for jan in args.jan:
|
84
|
+
|
85
|
+
try:
|
86
|
+
|
87
|
+
driver.get(JANJAN_URL)
|
88
|
+
|
89
|
+
jan_input_box = driver.find_element_by_xpath('//*[@id="content"]/div[1]/div/form/input[2]')
|
90
|
+
|
91
|
+
jan_input_box.send_keys(jan)
|
92
|
+
|
93
|
+
jan_input_button = driver.find_element_by_xpath('//*[@id="content"]/div[1]/div/form/input[3]')
|
94
|
+
|
95
|
+
jan_input_button.click()
|
96
|
+
|
97
|
+
|
98
|
+
|
99
|
+
content = driver.find_elements_by_xpath('//*[@id="content"]/div[2]/div[5]/ul/li')
|
100
|
+
|
101
|
+
if len(content) == 0:
|
102
|
+
|
103
|
+
raise NoSuchElementException
|
104
|
+
|
105
|
+
asin = None
|
106
|
+
|
107
|
+
for e in content:
|
108
|
+
|
109
|
+
desc = e.get_attribute('textContent')
|
110
|
+
|
111
|
+
if desc.startswith("ASIN"):
|
112
|
+
|
113
|
+
asin = desc.split()[-1]
|
114
|
+
|
115
|
+
if asin is None:
|
116
|
+
|
117
|
+
print('ASIN not found')
|
118
|
+
|
119
|
+
title = driver.find_elements_by_xpath('//*[@id="content"]/div[2]/div[3]/h3/a')[0].text
|
120
|
+
|
121
|
+
except NoSuchElementException:
|
122
|
+
|
123
|
+
print("not available")
|
124
|
+
|
125
|
+
print(jan)
|
126
|
+
|
127
|
+
continue
|
128
|
+
|
129
|
+
products[title] = {}
|
130
|
+
|
131
|
+
v = products[title]
|
132
|
+
|
133
|
+
v['ASIN'] = asin
|
134
|
+
|
135
|
+
v['JAN'] = jan
|
136
|
+
|
137
|
+
|
138
|
+
|
139
|
+
for k, v in products.items():
|
140
|
+
|
141
|
+
print(k, v)
|
142
|
+
|
143
|
+
driver.close()
|
144
|
+
|
145
|
+
|
146
|
+
|
147
|
+
if __name__ == "__main__":
|
148
|
+
|
149
|
+
main()
|
150
|
+
|
151
|
+
```
|