hawker 0.0.0.1
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/lib/hawker.rb +29 -0
- data/lib/hawker/drivers/default.rb +32 -0
- data/lib/hawker/drivers/github.rb +71 -0
- data/lib/hawker/drivers/instagram.rb +66 -0
- data/lib/hawker/drivers/twitter.rb +82 -0
- data/lib/hawker/mapper.rb +22 -0
- data/lib/hawker/request.rb +13 -0
- data/lib/hawker/version.rb +12 -0
- data/spec/helper.rb +5 -0
- metadata +114 -0
checksums.yaml
ADDED
@@ -0,0 +1,7 @@
|
|
1
|
+
---
|
2
|
+
SHA256:
|
3
|
+
metadata.gz: 51136fdddf3a7c5b7f533100dea0deb5167569946943d9719b23f6a19f1353f8
|
4
|
+
data.tar.gz: 8d7e8615ad452f0f1145c68762ca5c61cbf387ab95255f945e571322c20198e6
|
5
|
+
SHA512:
|
6
|
+
metadata.gz: fc14643a3acc9ff407475b259d078579388f429b1148505aef69ec6daefc06da36fa2fd9dd6696fd6b11ed9df4c52014f8af0b95958533fd3b19d57ddff9668b
|
7
|
+
data.tar.gz: a95638854a9b5b6c9fa4278dcc875c1bedf351b664664f7a27985ca28c30f36b0c704943f1dcfc7cb53d127799f74d0f0e7530a0db680683fcdca7f27f75aa1a
|
data/lib/hawker.rb
ADDED
@@ -0,0 +1,29 @@
|
|
1
|
+
require 'hawker/request'
|
2
|
+
require 'hawker/drivers/default'
|
3
|
+
require 'hawker/drivers/github'
|
4
|
+
require 'hawker/drivers/instagram'
|
5
|
+
require 'hawker/drivers/twitter'
|
6
|
+
require 'hawker/mapper'
|
7
|
+
|
8
|
+
module Hawker
|
9
|
+
# Returns the driver instance mapped to the given URL
|
10
|
+
#
|
11
|
+
# @return [Object]
|
12
|
+
def self.get(url)
|
13
|
+
response = Hawker::Request.call(url)
|
14
|
+
|
15
|
+
Hawker::Mapper.map(url, response)
|
16
|
+
end
|
17
|
+
|
18
|
+
# The list of available drivers
|
19
|
+
#
|
20
|
+
# @return [Array<String>]
|
21
|
+
def self.drivers
|
22
|
+
[
|
23
|
+
"GitHub",
|
24
|
+
"Twitter",
|
25
|
+
"Instagram",
|
26
|
+
"Default"
|
27
|
+
]
|
28
|
+
end
|
29
|
+
end
|
@@ -0,0 +1,32 @@
|
|
1
|
+
require 'nokogiri'
|
2
|
+
|
3
|
+
module Hawker
|
4
|
+
module Drivers
|
5
|
+
class Default
|
6
|
+
def initialize(response)
|
7
|
+
@response = response
|
8
|
+
end
|
9
|
+
|
10
|
+
# The current page title
|
11
|
+
#
|
12
|
+
# @return [String]
|
13
|
+
def page_title
|
14
|
+
node_text html.at("head title")
|
15
|
+
end
|
16
|
+
|
17
|
+
private
|
18
|
+
|
19
|
+
attr_reader :response
|
20
|
+
|
21
|
+
def html
|
22
|
+
@html ||= ::Nokogiri::HTML(response.body)
|
23
|
+
end
|
24
|
+
|
25
|
+
def node_text(node)
|
26
|
+
return if node.nil?
|
27
|
+
|
28
|
+
node.text.strip
|
29
|
+
end
|
30
|
+
end
|
31
|
+
end
|
32
|
+
end
|
@@ -0,0 +1,71 @@
|
|
1
|
+
module Hawker
|
2
|
+
module Drivers
|
3
|
+
class Github < Hawker::Drivers::Default
|
4
|
+
|
5
|
+
# The current user full name
|
6
|
+
#
|
7
|
+
# @return [String]
|
8
|
+
def name
|
9
|
+
node_text html.at("h1.vcard-names span.vcard-fullname[itemprop='name']")
|
10
|
+
end
|
11
|
+
|
12
|
+
# The current user nickname
|
13
|
+
#
|
14
|
+
# @return [String]
|
15
|
+
def username
|
16
|
+
node_text html.at("h1.vcard-names span.vcard-username[itemprop='additionalName']")
|
17
|
+
end
|
18
|
+
|
19
|
+
# The current user company
|
20
|
+
#
|
21
|
+
# @return [String]
|
22
|
+
def company
|
23
|
+
node_text html.at("li[itemprop='worksFor'] div")
|
24
|
+
end
|
25
|
+
|
26
|
+
# The current user location
|
27
|
+
#
|
28
|
+
# @return [String]
|
29
|
+
def location
|
30
|
+
node_text html.at("li[itemprop='homeLocation'] span")
|
31
|
+
end
|
32
|
+
|
33
|
+
# The current user website
|
34
|
+
#
|
35
|
+
# @return [String]
|
36
|
+
def website
|
37
|
+
node = html.at("li[itemprop='url'] a")
|
38
|
+
node[:href] unless node.nil?
|
39
|
+
end
|
40
|
+
|
41
|
+
# The count of repositiories owned by the current user
|
42
|
+
#
|
43
|
+
# @return [Integer]
|
44
|
+
def repositories
|
45
|
+
node_text(html.at("a[title='Repositories'] span.Counter")).to_i
|
46
|
+
end
|
47
|
+
|
48
|
+
# The count of stars that the current user gave
|
49
|
+
#
|
50
|
+
# @return [Integer]
|
51
|
+
def stars
|
52
|
+
node_text(html.at("a[title='Stars'] span.Counter")).to_i
|
53
|
+
end
|
54
|
+
|
55
|
+
# The count of users following the current user
|
56
|
+
#
|
57
|
+
# @return [Integer]
|
58
|
+
def followers
|
59
|
+
node_text(html.at("a[title='Followers'] span.Counter")).to_i
|
60
|
+
end
|
61
|
+
|
62
|
+
# The count of users that the current user is following
|
63
|
+
#
|
64
|
+
# @return [Integer]
|
65
|
+
def following
|
66
|
+
node_text(html.at("a[title='Following'] span.Counter")).to_i
|
67
|
+
end
|
68
|
+
|
69
|
+
end
|
70
|
+
end
|
71
|
+
end
|
@@ -0,0 +1,66 @@
|
|
1
|
+
require 'json'
|
2
|
+
|
3
|
+
module Hawker
|
4
|
+
module Drivers
|
5
|
+
class Instagram < Hawker::Drivers::Default
|
6
|
+
|
7
|
+
# The current user biography
|
8
|
+
#
|
9
|
+
# @return [String]
|
10
|
+
def biography
|
11
|
+
json["entry_data"]["ProfilePage"].first["graphql"]["user"]["biography"]
|
12
|
+
end
|
13
|
+
|
14
|
+
# The current user full name
|
15
|
+
#
|
16
|
+
# @return [String]
|
17
|
+
def full_name
|
18
|
+
json["entry_data"]["ProfilePage"].first["graphql"]["user"]["full_name"]
|
19
|
+
end
|
20
|
+
|
21
|
+
# The current user followers count
|
22
|
+
#
|
23
|
+
# @return [Integer]
|
24
|
+
def followers
|
25
|
+
json["entry_data"]["ProfilePage"].first["graphql"]["user"]["edge_followed_by"]["count"]
|
26
|
+
end
|
27
|
+
|
28
|
+
# The number of accounts that the current user follows
|
29
|
+
#
|
30
|
+
# @return [Integer]
|
31
|
+
def following
|
32
|
+
json["entry_data"]["ProfilePage"].first["graphql"]["user"]["edge_follow"]["count"]
|
33
|
+
end
|
34
|
+
|
35
|
+
# The current user external URL that is present in the biography
|
36
|
+
#
|
37
|
+
# @return [String]
|
38
|
+
def external_url
|
39
|
+
json["entry_data"]["ProfilePage"].first["graphql"]["user"]["external_url"]
|
40
|
+
end
|
41
|
+
|
42
|
+
# The current user profile picture URL
|
43
|
+
#
|
44
|
+
# @return [String]
|
45
|
+
def profile_pic_url
|
46
|
+
json["entry_data"]["ProfilePage"].first["graphql"]["user"]["profile_pic_url"]
|
47
|
+
end
|
48
|
+
|
49
|
+
# The current user Instagram username
|
50
|
+
#
|
51
|
+
# @return [String]
|
52
|
+
def username
|
53
|
+
json["entry_data"]["ProfilePage"].first["graphql"]["user"]["username"]
|
54
|
+
end
|
55
|
+
|
56
|
+
private
|
57
|
+
|
58
|
+
def json
|
59
|
+
@json ||= begin
|
60
|
+
raw_json = html.to_s.match(/window\._sharedData =.*(?=;)/)
|
61
|
+
JSON.parse(raw_json.to_s.gsub("window._sharedData = ", ""))
|
62
|
+
end
|
63
|
+
end
|
64
|
+
end
|
65
|
+
end
|
66
|
+
end
|
@@ -0,0 +1,82 @@
|
|
1
|
+
module Hawker
|
2
|
+
module Drivers
|
3
|
+
class Twitter < Hawker::Drivers::Default
|
4
|
+
|
5
|
+
# The current user name
|
6
|
+
#
|
7
|
+
# @return [String]
|
8
|
+
def name
|
9
|
+
node_text(html.at("h1.ProfileHeaderCard-name a"))
|
10
|
+
end
|
11
|
+
|
12
|
+
# The current user biography note
|
13
|
+
#
|
14
|
+
# @return [String]
|
15
|
+
def bio
|
16
|
+
node_text(html.at("p.ProfileHeaderCard-bio"))
|
17
|
+
end
|
18
|
+
|
19
|
+
# The current user location
|
20
|
+
#
|
21
|
+
# @return [String]
|
22
|
+
def location
|
23
|
+
node_text(html.at("span.ProfileHeaderCard-locationText"))
|
24
|
+
end
|
25
|
+
|
26
|
+
# The current user website
|
27
|
+
#
|
28
|
+
# @return [String]
|
29
|
+
def website
|
30
|
+
node = html.at("span.ProfileHeaderCard-urlText a")
|
31
|
+
node[:title] if !node.nil?
|
32
|
+
end
|
33
|
+
|
34
|
+
# The current user account creation date
|
35
|
+
#
|
36
|
+
# @return [String]
|
37
|
+
def joined
|
38
|
+
html.at("span.ProfileHeaderCard-joinDateText")[:title]
|
39
|
+
end
|
40
|
+
|
41
|
+
# The current user tweets count
|
42
|
+
#
|
43
|
+
# @return [Integer]
|
44
|
+
def tweets
|
45
|
+
node = html.at("a[data-nav='tweets'] span.ProfileNav-value")
|
46
|
+
node.nil? ? 0 : node["data-count"].to_i
|
47
|
+
end
|
48
|
+
|
49
|
+
# The current user followers count
|
50
|
+
#
|
51
|
+
# @return [Integer]
|
52
|
+
def followers
|
53
|
+
node = html.at("li.ProfileNav-item--followers span.ProfileNav-value")
|
54
|
+
node.nil? ? 0 : node["data-count"].to_i
|
55
|
+
end
|
56
|
+
|
57
|
+
# The count of accounts that the current user follows
|
58
|
+
#
|
59
|
+
# @return [Integer]
|
60
|
+
def following
|
61
|
+
node = html.at("li.ProfileNav-item--following span.ProfileNav-value")
|
62
|
+
node.nil? ? 0 : node["data-count"].to_i
|
63
|
+
end
|
64
|
+
|
65
|
+
# The current user likes count
|
66
|
+
#
|
67
|
+
# @return [Integer]
|
68
|
+
def likes
|
69
|
+
node = html.at("li.ProfileNav-item.ProfileNav-item--favorites span.ProfileNav-value")
|
70
|
+
node.nil? ? 0 : node["data-count"].to_i
|
71
|
+
end
|
72
|
+
|
73
|
+
# The current user lists count
|
74
|
+
#
|
75
|
+
# @return [Integer]
|
76
|
+
def lists
|
77
|
+
node_text(html.at("li.ProfileNav-item.ProfileNav-item--lists span.ProfileNav-value")).to_i
|
78
|
+
end
|
79
|
+
|
80
|
+
end
|
81
|
+
end
|
82
|
+
end
|
@@ -0,0 +1,22 @@
|
|
1
|
+
require 'uri'
|
2
|
+
|
3
|
+
module Hawker
|
4
|
+
class Mapper
|
5
|
+
MAPPINGS = {
|
6
|
+
/(http|https):\/\/(www.|)github.com\/[a-zA-Z0-9_]*/ => Hawker::Drivers::Github,
|
7
|
+
/(http|https):\/\/(www.|)instagram.com\/[a-zA-Z0-9_]*/ => Hawker::Drivers::Instagram,
|
8
|
+
/(http|https):\/\/(www.|)twitter.com\/[a-zA-Z0-9_]*/ => Hawker::Drivers::Twitter,
|
9
|
+
"default" => Hawker::Drivers::Default
|
10
|
+
}.freeze
|
11
|
+
|
12
|
+
# Maps the given URL to the driver class
|
13
|
+
#
|
14
|
+
# @return [Object]
|
15
|
+
def self.map(url, response)
|
16
|
+
driver_mappings = MAPPINGS.find { |key, value| url.match(key) }
|
17
|
+
driver = driver_mappings.nil? ? MAPPINGS["default"] : driver_mappings.last
|
18
|
+
|
19
|
+
driver.new(response)
|
20
|
+
end
|
21
|
+
end
|
22
|
+
end
|
@@ -0,0 +1,13 @@
|
|
1
|
+
require 'rest-client'
|
2
|
+
|
3
|
+
module Hawker
|
4
|
+
class Request
|
5
|
+
|
6
|
+
# Performs the request to the given URL in order to get the page html
|
7
|
+
#
|
8
|
+
# @return [RestClient::Response]
|
9
|
+
def self.call(url)
|
10
|
+
::RestClient.get(url, user_agent: 'Mozilla/5.0 Mac OS')
|
11
|
+
end
|
12
|
+
end
|
13
|
+
end
|
data/spec/helper.rb
ADDED
metadata
ADDED
@@ -0,0 +1,114 @@
|
|
1
|
+
--- !ruby/object:Gem::Specification
|
2
|
+
name: hawker
|
3
|
+
version: !ruby/object:Gem::Version
|
4
|
+
version: 0.0.0.1
|
5
|
+
platform: ruby
|
6
|
+
authors:
|
7
|
+
- Paweł Dąbrowski
|
8
|
+
autorequire:
|
9
|
+
bindir: bin
|
10
|
+
cert_chain: []
|
11
|
+
date: 2018-04-20 00:00:00.000000000 Z
|
12
|
+
dependencies:
|
13
|
+
- !ruby/object:Gem::Dependency
|
14
|
+
name: rest-client
|
15
|
+
requirement: !ruby/object:Gem::Requirement
|
16
|
+
requirements:
|
17
|
+
- - "~>"
|
18
|
+
- !ruby/object:Gem::Version
|
19
|
+
version: '2.0'
|
20
|
+
- - ">="
|
21
|
+
- !ruby/object:Gem::Version
|
22
|
+
version: 2.0.2
|
23
|
+
type: :runtime
|
24
|
+
prerelease: false
|
25
|
+
version_requirements: !ruby/object:Gem::Requirement
|
26
|
+
requirements:
|
27
|
+
- - "~>"
|
28
|
+
- !ruby/object:Gem::Version
|
29
|
+
version: '2.0'
|
30
|
+
- - ">="
|
31
|
+
- !ruby/object:Gem::Version
|
32
|
+
version: 2.0.2
|
33
|
+
- !ruby/object:Gem::Dependency
|
34
|
+
name: nokogiri
|
35
|
+
requirement: !ruby/object:Gem::Requirement
|
36
|
+
requirements:
|
37
|
+
- - "~>"
|
38
|
+
- !ruby/object:Gem::Version
|
39
|
+
version: '1.8'
|
40
|
+
- - ">="
|
41
|
+
- !ruby/object:Gem::Version
|
42
|
+
version: 1.8.2
|
43
|
+
type: :runtime
|
44
|
+
prerelease: false
|
45
|
+
version_requirements: !ruby/object:Gem::Requirement
|
46
|
+
requirements:
|
47
|
+
- - "~>"
|
48
|
+
- !ruby/object:Gem::Version
|
49
|
+
version: '1.8'
|
50
|
+
- - ">="
|
51
|
+
- !ruby/object:Gem::Version
|
52
|
+
version: 1.8.2
|
53
|
+
- !ruby/object:Gem::Dependency
|
54
|
+
name: rspec
|
55
|
+
requirement: !ruby/object:Gem::Requirement
|
56
|
+
requirements:
|
57
|
+
- - "~>"
|
58
|
+
- !ruby/object:Gem::Version
|
59
|
+
version: '3.7'
|
60
|
+
- - ">="
|
61
|
+
- !ruby/object:Gem::Version
|
62
|
+
version: 3.7.0
|
63
|
+
type: :development
|
64
|
+
prerelease: false
|
65
|
+
version_requirements: !ruby/object:Gem::Requirement
|
66
|
+
requirements:
|
67
|
+
- - "~>"
|
68
|
+
- !ruby/object:Gem::Version
|
69
|
+
version: '3.7'
|
70
|
+
- - ">="
|
71
|
+
- !ruby/object:Gem::Version
|
72
|
+
version: 3.7.0
|
73
|
+
description: Scrap the most popular websites without the knowledge about the page
|
74
|
+
structure or API access
|
75
|
+
email: dziamber@gmail.com
|
76
|
+
executables: []
|
77
|
+
extensions: []
|
78
|
+
extra_rdoc_files: []
|
79
|
+
files:
|
80
|
+
- lib/hawker.rb
|
81
|
+
- lib/hawker/drivers/default.rb
|
82
|
+
- lib/hawker/drivers/github.rb
|
83
|
+
- lib/hawker/drivers/instagram.rb
|
84
|
+
- lib/hawker/drivers/twitter.rb
|
85
|
+
- lib/hawker/mapper.rb
|
86
|
+
- lib/hawker/request.rb
|
87
|
+
- lib/hawker/version.rb
|
88
|
+
- spec/helper.rb
|
89
|
+
homepage: http://github.com/rubyhero/hawker
|
90
|
+
licenses:
|
91
|
+
- MIT
|
92
|
+
metadata: {}
|
93
|
+
post_install_message:
|
94
|
+
rdoc_options: []
|
95
|
+
require_paths:
|
96
|
+
- lib
|
97
|
+
required_ruby_version: !ruby/object:Gem::Requirement
|
98
|
+
requirements:
|
99
|
+
- - ">="
|
100
|
+
- !ruby/object:Gem::Version
|
101
|
+
version: '0'
|
102
|
+
required_rubygems_version: !ruby/object:Gem::Requirement
|
103
|
+
requirements:
|
104
|
+
- - ">="
|
105
|
+
- !ruby/object:Gem::Version
|
106
|
+
version: '0'
|
107
|
+
requirements: []
|
108
|
+
rubyforge_project:
|
109
|
+
rubygems_version: 2.7.3
|
110
|
+
signing_key:
|
111
|
+
specification_version: 4
|
112
|
+
summary: Scrap the most popular websites without the knowledge about the page structure
|
113
|
+
or API access
|
114
|
+
test_files: []
|