|
a/OSSEval/OSSEval/utils.py |
|
b/OSSEval/OSSEval/utils.py |
| 1 |
from json import JSONEncoder
|
1 |
from json import JSONEncoder
|
|
|
2 |
from lxml import html
|
|
|
3 |
import urllib2
|
|
|
4 |
from apiclient.discovery import build
|
| 2 |
|
5 |
|
| 3 |
class TrivialJSONEncoder(JSONEncoder):
|
6 |
class TrivialJSONEncoder(JSONEncoder):
|
| 4 |
def default(self, o):
|
7 |
def default(self, o):
|
| 5 |
return o.__dict__
|
8 |
return o.__dict__
|
| 6 |
|
9 |
|
|
... |
|
... |
| 27 |
try:
|
30 |
try:
|
| 28 |
return int(xmldoc.attributes[tag].firstChild.data)
|
31 |
return int(xmldoc.attributes[tag].firstChild.data)
|
| 29 |
except:
|
32 |
except:
|
| 30 |
return -1
|
33 |
return -1
|
| 31 |
|
34 |
|
|
|
35 |
class SearchEngine():
|
|
|
36 |
@staticmethod
|
|
|
37 |
def search__engine_name():
|
|
|
38 |
return "Google"
|
|
|
39 |
|
|
|
40 |
@staticmethod
|
|
|
41 |
def search_url(search_text, site_search=""):
|
|
|
42 |
if site_search == "":
|
|
|
43 |
return "https://www.google.com/?#q=" + search_text
|
|
|
44 |
else:
|
|
|
45 |
return "https://www.google.com/?#q=" + search_text + "+site%3A" + site_search
|
|
|
46 |
|
|
|
47 |
@staticmethod
|
|
|
48 |
def readable_query(search_text, site_search=""):
|
|
|
49 |
return search_text.replace("%3A", ":").replace("%20", " ").replace("%2B", "+") + ("" if site_search=="" else " site:" + site_search)
|
|
|
50 |
|
|
|
51 |
@staticmethod
|
|
|
52 |
def search_count(search_text, site_search=""):
|
|
|
53 |
'''
|
|
|
54 |
https://developers.google.com/custom-search/json-api/v1/reference/cse/list
|
|
|
55 |
exactTerms string Identifies a phrase that all documents in the search results must contain.
|
|
|
56 |
excludeTerms string Identifies a word or phrase that should not appear in any documents in the search results.
|
|
|
57 |
siteSearch string Specifies all search results should be pages from a given site.
|
|
|
58 |
siteSearchFilter string Controls whether to include or exclude results from the site named in the siteSearch parameter.
|
|
|
59 |
Acceptable values are: "e": exclude "i": include
|
|
|
60 |
'''
|
|
|
61 |
service = build("customsearch", "v1", developerKey="AIzaSyCAova46cAkHga_SZWTBqROdjoz1KcTlw8")
|
|
|
62 |
print search_text + " " + site_search
|
|
|
63 |
if site_search == "":
|
|
|
64 |
res = service.cse().list(q=search_text, cx='017576662512468239146:omuauf_lfve',).execute()
|
|
|
65 |
else:
|
|
|
66 |
res = service.cse().list(q=search_text, siteSearch=site_search, cx='017576662512468239146:omuauf_lfve',).execute()
|
|
|
67 |
return int(res['searchInformation']['totalResults'])
|