Пример #1
0
def main(argv=None):
	if argv is None:
	   argv = sys.argv

	opts, extraparams = getopt.gnu_getopt(argv[1:], "hvedo", ["help", 'export', 'drop', 'order'])

	verbose = False
	export = False
	drop = False
	order = False    

	for o, a in opts:
		if o == "-v":
			verbose = True
		elif o in ("-h", "--help"):
			print(__doc__)
			sys.exit(0)
		elif o in ("-e", "--export"):
			export = True
		elif o in ("-d", "--drop"):
			drop = True
		elif o in ("-o", "--order"):
			order = True
		else:
			assert False, "UnhandledOption"

    #begin
	google = Google(settings.proxy)
	di = SmartDict(size=4)
	km = KeywordManager(di, google)
	    
	if order:
		keywords = km.order_keywords()
		print('in keygrabber: '+str(len(keywords)))
		km.create_orderedkeys_collection(keywords)
		return 0
    
	if export:
		km.export_keywords()
		return 0
    
	if drop:
		km.drop_database()
		return 0

	km.not_so_simple_search(base='come')

	return 0
# -*- coding: utf-8 -*-

import sys
sys.path.append("lib")

from article_manager import Article
from article_manager import ArticleManager
from keyword_manager import KeywordManager
from google_news import GoogleNews

from ironnews_utility import IronnewsUtility

print "Content-Type: text/plain"
print ""

KeywordManager.initialize()

keyword = KeywordManager.get()
print keyword.encode("utf-8")

articles = GoogleNews.search(keyword, 30)
for article in articles:
  url   = article["url"]
  title = article["title"]
  print "---"
  print url
  print title.encode("utf-8")
  if IronnewsUtility.reject(url):
    print "reject!"
    continue
Пример #3
0
print "Content-Type: text/html; charset=utf-8"
print ""

print "<html>"
print " <head>"
print "  <title>ironnews-crawler status</title>"
print " </head>"
print " <body>"

print "  <h1>ironnews-crawler status</h1>"

print "  <h2>keywords</h2>"
print "  <table border='1'>"

for keyword in KeywordManager.all():
  print "<tr>"
  print " <td>" + (keyword.updated_at + datetime.timedelta(hours = 9)).strftime("%Y-%m-%d %H:%M") + "</td>"
  print " <td>" + keyword.name.encode("utf-8") + "</td>" # FIXME: HTML escape
  print "</tr>"

print "  </table>"

print "  <h2>latest articles</h2>"

category_stats = ArticleManager.category_stats()
state_stats    = ArticleManager.state_stats()

print "<table border='1'>"
print " <tr>"
print "  <th colspan='3'>カテゴリ</th>"
Пример #4
0
def main(argv=None):
    if argv is None:
        argv = sys.argv

    opts, extraparams = getopt.gnu_getopt(argv[1:], "hvedo",
                                          ["help", 'export', 'drop', 'order'])

    verbose = False
    export = False
    drop = False
    order = False

    for o, a in opts:
        if o == "-v":
            verbose = True
        elif o in ("-h", "--help"):
            print(__doc__)
            sys.exit(0)
        elif o in ("-e", "--export"):
            export = True
        elif o in ("-d", "--drop"):
            drop = True
        elif o in ("-o", "--order"):
            order = True
        else:
            assert False, "UnhandledOption"

#begin
    google = Google(settings.proxy)
    di = SmartDict(size=4)
    km = KeywordManager(di, google)

    if order:
        keywords = km.order_keywords()
        print('in keygrabber: ' + str(len(keywords)))
        km.create_orderedkeys_collection(keywords)
        return 0

    if export:
        km.export_keywords()
        return 0

    if drop:
        km.drop_database()
        return 0

    km.not_so_simple_search(base='come')

    return 0