43 lines
1.3 KiB
Plaintext
43 lines
1.3 KiB
Plaintext
|
#!/usr/bin/env python
|
||
|
|
|
||
|
|
import subprocess
|
||
|
|
from optparse import OptionParser
|
||
|
|
from urllib import quote
|
||
|
|
import sys
|
||
|
|
|
||
|
|
__VERSION__ = 0.1
|
||
|
|
SEARCH_URL ='''\
|
||
|
|
http://www.google.com/search?sclient=psy&hl=en&site=&source=hp&q=foo&btnG=Google+Search\
|
||
|
|
'''
|
||
|
|
SCRAPY_PARAMS = ['--set', 'FEED_URI=data.json', '--set', 'FEED_FORMAT=json']
|
||
|
|
|
||
|
|
|
||
|
|
if __name__ == '__main__':
|
||
|
|
usage = 'usage: %prog query [OPTION...]'
|
||
|
|
op = OptionParser(usage, version='%%prog %s' % __VERSION__)
|
||
|
|
opts, args = op.parse_args()
|
||
|
|
#try:
|
||
|
|
if len(args) != 1:
|
||
|
|
print 'Bad arguments: '
|
||
|
|
op.print_usage()
|
||
|
|
else:
|
||
|
|
scrapy = 'scrapy'
|
||
|
|
scrapy_cmd = 'crawl'
|
||
|
|
query = args[0]
|
||
|
|
url_query = quote(query)
|
||
|
|
search = SEARCH_URL.replace('foo', url_query)
|
||
|
|
print url_query
|
||
|
|
print 'Launching scrapy with paramters <%s> <%s>' % (scrapy,
|
||
|
|
search)
|
||
|
|
p = subprocess.Popen([
|
||
|
|
scrapy, scrapy_cmd,
|
||
|
|
search, '--set',
|
||
|
|
'FEED_URI=data.json',
|
||
|
|
'--set',
|
||
|
|
'FEED_FORMAT=JSON'
|
||
|
|
],
|
||
|
|
stdout=subprocess.PIPE,
|
||
|
|
stderr=subprocess.PIPE)
|
||
|
|
output = p.communicate()[0]
|
||
|
|
print output
|