MCPcopy
hub / github.com/geekcomputers/Python / SearchResults

Function SearchResults

googleweb.py:11–66  ·  view source on GitHub ↗
()

Source from the content-addressed store, hash-verified

9
10
11def SearchResults():
12 lis = []
13 f = open("Input", "r")
14 header = {
15 "User-Agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_11_6) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/61.0.3163.100 Safari/537.36"
16 }
17 StabUrl = "https://www.google.com/search?rlz=1C1CHBD_enIN872IN872&sxsrf=ALeKk03OHYAnSxX60oUwmblKn36Hyi8MhA%3A1600278715451&ei=u1BiX9ibG7qU4-EP_qGPgA8&q="
18 midUrl = "&oq="
19 EndUrl = "&gs_lcp=CgZwc3ktYWIQAzoECAAQR1C11AxYtdQMYJXcDGgAcAF4AIABpQKIAaUCkgEDMi0xmAEAoAECoAEBqgEHZ3dzLXdpesgBCMABAQ&sclient=psy-ab&ved=0ahUKEwiY5YDjnu7rAhU6yjgGHf7QA_AQ4dUDCA0&uact=5"
20 for i in f:
21 singleLink = []
22 singleRatio = []
23 singleWrite = []
24 singleWrite.append(i.strip("\n"))
25 checkString = i.replace("+", "")
26 searchString = i.replace("+", "%2B")
27 searchString = searchString.replace(" ", "+")
28 searchString = StabUrl + searchString + midUrl + searchString + EndUrl
29 r = requests.get(searchString, headers=header)
30 soup = bs4.BeautifulSoup(r.text, features="html.parser")
31 elements = soup.select(".r a")
32 for g in elements:
33 lis.append(g.get("href"))
34 for k in lis:
35 sentence = ""
36 if (k[0] != "#") and k[0] != "/":
37 checker = k[8:16]
38 if checker != "webcache":
39 rr = requests.get(k, headers=header, verify=False)
40 soupInside = bs4.BeautifulSoup(rr.text, features="html.parser")
41 elementInside = soupInside.select("body")
42 for line in elementInside:
43 sentence = sentence + line.text
44 ratio = fuzz.token_set_ratio(sentence, checkString)
45 if ratio > 80:
46 singleLink.append(k)
47 singleRatio.append(ratio)
48 if len(singleLink) >= 4:
49 singleLink = np.array(singleLink)
50 singleRatio = np.array(singleRatio)
51 inds = singleRatio.argsort()
52 sortedLink = singleLink[inds]
53 sortedFinalList = list(sortedLink[::-1])
54 sortedFinalList = sortedFinalList[:4]
55 FinalResult.append(singleWrite + sortedFinalList)
56 elif (len(singleLink) < 4) and len(singleLink) > 0:
57 singleLink = np.array(singleLink)
58 singleRatio = np.array(singleRatio)
59 inds = singleRatio.argsort()
60 sortedLink = singleLink[inds]
61 sortedFinalList = list(sortedLink[::-1])
62 sortedFinalList = sortedFinalList + (4 - len(sortedFinalList)) * [[" "]]
63 FinalResult.append(singleWrite + sortedFinalList)
64 else:
65 sortedFinalList = [[" "]] * 4
66 FinalResult.append(singleWrite + sortedFinalList)
67
68

Callers 1

googleweb.pyFile · 0.85

Calls 2

appendMethod · 0.45
getMethod · 0.45

Tested by

no test coverage detected