ubi 0.0.3
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- checksums.yaml +7 -0
- data/.gitignore +9 -0
- data/.rspec +3 -0
- data/Gemfile +11 -0
- data/Guardfile +16 -0
- data/MIT-LICENSE +19 -0
- data/README.md +57 -0
- data/Rakefile +9 -0
- data/bin/ubi +10 -0
- data/lib/ubi.rb +37 -0
- data/lib/ubi/aranea.rb +42 -0
- data/lib/ubi/artifex.rb +28 -0
- data/lib/ubi/consultor.rb +61 -0
- data/lib/ubi/consultores/bing.rb +26 -0
- data/lib/ubi/consultores/duckduckgo.rb +26 -0
- data/lib/ubi/consultores/facebook.rb +6 -0
- data/lib/ubi/consultores/foursquare.rb +6 -0
- data/lib/ubi/consultores/google.rb +24 -0
- data/lib/ubi/consultores/linkedin.rb +0 -0
- data/lib/ubi/consultores/twitter.rb +6 -0
- data/lib/ubi/consultores/wikipedia.rb +6 -0
- data/lib/ubi/consultores/yahoo.rb +26 -0
- data/lib/ubi/datum.rb +43 -0
- data/lib/ubi/impero.rb +20 -0
- data/lib/ubi/memoria.rb +72 -0
- data/lib/ubi/memorias/address.rb +71 -0
- data/lib/ubi/memorias/document.rb +50 -0
- data/lib/ubi/memorias/email.rb +19 -0
- data/lib/ubi/memorias/phone.rb +33 -0
- data/lib/ubi/memorias/site.rb +29 -0
- data/lib/ubi/memorias/social.rb +20 -0
- data/lib/ubi/memorias/who.rb +20 -0
- data/lib/ubi/thema.rb +62 -0
- data/lib/ubi/version.rb +4 -0
- data/spec/fixtures/email.txt +5 -0
- data/spec/fixtures/mobile.txt +17 -0
- data/spec/fixtures/page.txt +21 -0
- data/spec/fixtures/phone.txt +17 -0
- data/spec/fixtures/site.txt +21 -0
- data/spec/spec_helper.rb +40 -0
- data/spec/ubi/aranea_spec.rb +19 -0
- data/spec/ubi/artifex_spec.rb +4 -0
- data/spec/ubi/memorias/address_spec.rb +56 -0
- data/spec/ubi/memorias/document_spec.rb +48 -0
- data/spec/ubi/memorias/email_spec.rb +59 -0
- data/spec/ubi/memorias/phone_spec.rb +79 -0
- data/spec/ubi/memorias/site_spec.rb +82 -0
- data/spec/ubi/thema_spec.rb +33 -0
- data/ubi.gemspec +39 -0
- metadata +232 -0
checksums.yaml
ADDED
@@ -0,0 +1,7 @@
|
|
1
|
+
---
|
2
|
+
SHA1:
|
3
|
+
metadata.gz: 330dfff07aab154b3fc714a71b10cf0f148d541d
|
4
|
+
data.tar.gz: 81ff08e9a049671a6f2c51c43641007466fa30c3
|
5
|
+
SHA512:
|
6
|
+
metadata.gz: fae80b06b1e6f5bac23357f8b4c09ef45879ba62b4e1e6e90c9e42a16490f7496cc614b27538beaed7d6529d8784fce29a79722e4a351379079cc0e4be6bf1cb
|
7
|
+
data.tar.gz: 4a4ec085b39e27a0e64addcf9c22fff3a969bdff21a72c8684bcb500451d3c4503eb1935383fd96715dd72257082f4851476b9b2c8d4a26244b6363b9264588b
|
data/.gitignore
ADDED
data/.rspec
ADDED
data/Gemfile
ADDED
data/Guardfile
ADDED
@@ -0,0 +1,16 @@
|
|
1
|
+
#
|
2
|
+
# Ubi Guard
|
3
|
+
#
|
4
|
+
guard :rubocop, all_on_start: false, keep_failed: false, cli: ['-D'] do
|
5
|
+
watch(/.+\.rb$/)
|
6
|
+
watch(/(?:.+\/)?\.rubocop\.yml$/) { |m| File.dirname(m[0]) }
|
7
|
+
end
|
8
|
+
|
9
|
+
guard :rspec, cmd: 'bundle exec rspec' do
|
10
|
+
watch(/^spec\/.+_spec\.rb$/)
|
11
|
+
watch(/^lib\/(.+)\.rb$/) { |m| "spec/#{m[1]}_spec.rb" }
|
12
|
+
watch(/^generators\/(.+)\.rb$/) { |_m| 'spec/schemaless/worker_spec' }
|
13
|
+
|
14
|
+
watch('lib/ubi.rb') { 'spec' }
|
15
|
+
watch('spec/spec_helper.rb') { 'spec' }
|
16
|
+
end
|
data/MIT-LICENSE
ADDED
@@ -0,0 +1,19 @@
|
|
1
|
+
Copyright (c) 2014-2015 Ubi Authors
|
2
|
+
|
3
|
+
Permission is hereby granted, free of charge, to any person obtaining a
|
4
|
+
copy of this software and associated documentation files (the "Software"),
|
5
|
+
to deal in the Software without restriction, including without limitation
|
6
|
+
the rights to use, copy, modify, merge, publish, distribute, sublicense,
|
7
|
+
and/or sell copies of the Software, and to permit persons to whom the
|
8
|
+
Software is furnished to do so, subject to the following conditions:
|
9
|
+
|
10
|
+
The above copyright notice and this permission notice shall be
|
11
|
+
included in all copies or substantial portions of the Software.
|
12
|
+
|
13
|
+
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
|
14
|
+
EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
|
15
|
+
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
|
16
|
+
NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE
|
17
|
+
LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
|
18
|
+
OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
|
19
|
+
WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
|
data/README.md
ADDED
@@ -0,0 +1,57 @@
|
|
1
|
+
_ _
|
2
|
+
/\ /\ | |__ (_)
|
3
|
+
/ / \ \| '_ \ | |
|
4
|
+
\ \_/ /| |_) || |
|
5
|
+
\___/ |_.__/ |_|
|
6
|
+
|
7
|
+
|
8
|
+
Ubi finds information in the subject's webpage(s).
|
9
|
+
A forager.
|
10
|
+
|
11
|
+
|
12
|
+
## Thema -> Subject/Matter
|
13
|
+
|
14
|
+
- Name
|
15
|
+
- Email
|
16
|
+
- URL
|
17
|
+
|
18
|
+
|
19
|
+
## Memoria -> Attribute/Trait/Memory
|
20
|
+
|
21
|
+
- URLs
|
22
|
+
- Emails
|
23
|
+
- Phones
|
24
|
+
- Social
|
25
|
+
- Files
|
26
|
+
- Documents
|
27
|
+
- Address
|
28
|
+
- Logo/Images
|
29
|
+
|
30
|
+
|
31
|
+
## Aranea -> Spider/Crawler
|
32
|
+
|
33
|
+
Search
|
34
|
+
API
|
35
|
+
|
36
|
+
Others
|
37
|
+
|
38
|
+
|
39
|
+
## Use
|
40
|
+
|
41
|
+
```
|
42
|
+
Ubi::Thema.new('A Company on City').urls
|
43
|
+
Ubi::Thema.new('A Company on City', 'company.com').phones
|
44
|
+
```
|
45
|
+
|
46
|
+
### Memorias Directly
|
47
|
+
|
48
|
+
```
|
49
|
+
Ubi::Memoria::Email.parse('A text with some valid@emails.com')
|
50
|
+
Ubi::Memoria::Phone.parse('A text with some +55-5555-5555')
|
51
|
+
Ubi::Memoria::Site.parse('A text with some http://urls.com')
|
52
|
+
```
|
53
|
+
|
54
|
+
##### Keep crawling
|
55
|
+
|
56
|
+
> Ubi, and the class names are latin, in case you're curious.
|
57
|
+
> Yeah, apparently I got plenty of free time.
|
data/Rakefile
ADDED
data/bin/ubi
ADDED
data/lib/ubi.rb
ADDED
@@ -0,0 +1,37 @@
|
|
1
|
+
require 'pry'
|
2
|
+
require 'thor'
|
3
|
+
require 'net/http'
|
4
|
+
require 'net/https'
|
5
|
+
require 'uri'
|
6
|
+
require 'open-uri'
|
7
|
+
require 'json'
|
8
|
+
require 'phonelib'
|
9
|
+
require 'nokogiri'
|
10
|
+
require 'active_model'
|
11
|
+
require 'active_support'
|
12
|
+
require 'active_support/core_ext/hash'
|
13
|
+
require 'ubi/version'
|
14
|
+
|
15
|
+
# Ubiquous getters
|
16
|
+
module Ubi
|
17
|
+
cattr_accessor(:memorias) { [] }
|
18
|
+
cattr_accessor(:araneas) { [] }
|
19
|
+
|
20
|
+
# Phonie.configuration.default_country_code = '55'
|
21
|
+
# Phonie.configuration.n1_length = 4
|
22
|
+
end
|
23
|
+
|
24
|
+
require 'ubi/aranea'
|
25
|
+
require 'ubi/memoria'
|
26
|
+
require 'ubi/consultor'
|
27
|
+
|
28
|
+
%w( memorias consultores ).each do |ns|
|
29
|
+
Dir["#{File.dirname(__FILE__)}/ubi/#{ns}/**.rb"].sort.each do |lib|
|
30
|
+
require lib
|
31
|
+
end
|
32
|
+
end
|
33
|
+
|
34
|
+
# Loads after all
|
35
|
+
require 'ubi/artifex'
|
36
|
+
require 'ubi/datum'
|
37
|
+
require 'ubi/thema'
|
data/lib/ubi/aranea.rb
ADDED
@@ -0,0 +1,42 @@
|
|
1
|
+
require 'polipus'
|
2
|
+
|
3
|
+
module Ubi
|
4
|
+
# Base for araneas (spiders)
|
5
|
+
class Aranea
|
6
|
+
OPTIONS = {
|
7
|
+
workers: 3,
|
8
|
+
user_agent: "Ubi v#{Ubi::VERSION}",
|
9
|
+
depth_limit: 1
|
10
|
+
|
11
|
+
# storage: MemoryStore
|
12
|
+
}
|
13
|
+
attr_accessor :thema, :url, :datum
|
14
|
+
|
15
|
+
def initialize(thema, url, opts = {})
|
16
|
+
@thema = thema
|
17
|
+
@url = url
|
18
|
+
@opts = opts
|
19
|
+
end
|
20
|
+
|
21
|
+
delegate :name, to: :thema
|
22
|
+
|
23
|
+
def crawl!
|
24
|
+
Polipus.crawler(name, url, OPTIONS.merge(@opts)) do |crawler|
|
25
|
+
# In-place page processing
|
26
|
+
crawler.on_page_downloaded do |page|
|
27
|
+
# A nokogiri object
|
28
|
+
puts "'#{page.doc.css('title').text}' (#{page.url})"
|
29
|
+
end
|
30
|
+
end
|
31
|
+
end
|
32
|
+
|
33
|
+
def parser(chunk)
|
34
|
+
Nokogiri::HTML(chunk)
|
35
|
+
end
|
36
|
+
|
37
|
+
def datum
|
38
|
+
crawl! unless @datum
|
39
|
+
@datum
|
40
|
+
end
|
41
|
+
end
|
42
|
+
end
|
data/lib/ubi/artifex.rb
ADDED
@@ -0,0 +1,28 @@
|
|
1
|
+
module Ubi
|
2
|
+
# Artifex: I work!
|
3
|
+
class Artifex
|
4
|
+
include Consultor
|
5
|
+
attr_accessor :thema
|
6
|
+
|
7
|
+
def initialize(query)
|
8
|
+
@thema = Thema.new(query)
|
9
|
+
start_with_search
|
10
|
+
social_search
|
11
|
+
other_search
|
12
|
+
end
|
13
|
+
|
14
|
+
def start_with_search
|
15
|
+
[Google, Yahoo, Bing, DuckDuckGo].each do |s|
|
16
|
+
@thema.try_consultor(s)
|
17
|
+
end
|
18
|
+
end
|
19
|
+
|
20
|
+
def social_search
|
21
|
+
end
|
22
|
+
|
23
|
+
def other_search
|
24
|
+
end
|
25
|
+
|
26
|
+
delegate :spec, to: :thema
|
27
|
+
end
|
28
|
+
end
|
@@ -0,0 +1,61 @@
|
|
1
|
+
module Ubi
|
2
|
+
module Consultor
|
3
|
+
# Base for araneas (spiders)
|
4
|
+
class Base
|
5
|
+
HEADERS = { 'User-Agent' => "Ubi v#{Ubi::VERSION}" }
|
6
|
+
|
7
|
+
def initialize(thema)
|
8
|
+
@thema = thema
|
9
|
+
end
|
10
|
+
|
11
|
+
def query_url
|
12
|
+
self.class.url + query.to_query
|
13
|
+
end
|
14
|
+
|
15
|
+
def links
|
16
|
+
'//a'
|
17
|
+
end
|
18
|
+
|
19
|
+
def parser(chunk)
|
20
|
+
Nokogiri::HTML(chunk)
|
21
|
+
end
|
22
|
+
|
23
|
+
def datum
|
24
|
+
@datum ||= Ubi::Datum.new(parser(request), words, links)
|
25
|
+
end
|
26
|
+
|
27
|
+
#
|
28
|
+
# Make an HTTP(S) request to a geocoding API and
|
29
|
+
# return the response object.
|
30
|
+
#
|
31
|
+
def request(opts = {})
|
32
|
+
timeout(10) do
|
33
|
+
uri = URI.parse(query_url)
|
34
|
+
puts "#{self} working on `#{@thema}` (#{query_url}) #{opts}"
|
35
|
+
uri.open(HEADERS).read
|
36
|
+
end
|
37
|
+
end
|
38
|
+
|
39
|
+
class << self
|
40
|
+
def inherited(base)
|
41
|
+
puts "Using aranea #{base}"
|
42
|
+
Ubi.araneas << base
|
43
|
+
end
|
44
|
+
|
45
|
+
#
|
46
|
+
# Human-readable name of the aranea
|
47
|
+
#
|
48
|
+
def name
|
49
|
+
fail "Not implemented by #{self}"
|
50
|
+
end
|
51
|
+
|
52
|
+
#
|
53
|
+
# Url to query
|
54
|
+
#
|
55
|
+
def url
|
56
|
+
fail "Not implemented by #{self}"
|
57
|
+
end
|
58
|
+
end
|
59
|
+
end
|
60
|
+
end
|
61
|
+
end
|
@@ -0,0 +1,26 @@
|
|
1
|
+
module Ubi
|
2
|
+
module Consultor
|
3
|
+
# URL finder consultor
|
4
|
+
class Bing < Base
|
5
|
+
#
|
6
|
+
# query
|
7
|
+
def query
|
8
|
+
{ q: @thema.name, pc: 'ubi' }
|
9
|
+
end
|
10
|
+
|
11
|
+
def links
|
12
|
+
'//a[not (contains(@href,"bing"))]'
|
13
|
+
end
|
14
|
+
|
15
|
+
def words
|
16
|
+
'//li[@class="g"]'
|
17
|
+
end
|
18
|
+
|
19
|
+
class << self
|
20
|
+
def url
|
21
|
+
'https://bing.com/search?'
|
22
|
+
end
|
23
|
+
end
|
24
|
+
end
|
25
|
+
end
|
26
|
+
end
|
@@ -0,0 +1,26 @@
|
|
1
|
+
module Ubi
|
2
|
+
module Consultor
|
3
|
+
# URL finder consultor
|
4
|
+
class DuckDuckGo < Base
|
5
|
+
#
|
6
|
+
# query
|
7
|
+
def query
|
8
|
+
{ q: @thema.name, t: 'ubi' }
|
9
|
+
end
|
10
|
+
|
11
|
+
def links
|
12
|
+
'//a[not (contains(@href,"duckduckgo"))]'
|
13
|
+
end
|
14
|
+
|
15
|
+
def words
|
16
|
+
'//li[@class="g"]'
|
17
|
+
end
|
18
|
+
|
19
|
+
class << self
|
20
|
+
def url
|
21
|
+
'https://duckduckgo.com/?'
|
22
|
+
end
|
23
|
+
end
|
24
|
+
end
|
25
|
+
end
|
26
|
+
end
|
@@ -0,0 +1,24 @@
|
|
1
|
+
module Ubi
|
2
|
+
module Consultor
|
3
|
+
# URL finder consultor
|
4
|
+
class Google < Base # HTML
|
5
|
+
def query
|
6
|
+
{ q: @thema.name }
|
7
|
+
end
|
8
|
+
|
9
|
+
def links
|
10
|
+
'//a[not (contains(@href,"google"))]'
|
11
|
+
end
|
12
|
+
|
13
|
+
def words
|
14
|
+
'//div[@class="res"]'
|
15
|
+
end
|
16
|
+
|
17
|
+
class << self
|
18
|
+
def url
|
19
|
+
'https://www.google.com.br/search?'
|
20
|
+
end
|
21
|
+
end
|
22
|
+
end
|
23
|
+
end
|
24
|
+
end
|
File without changes
|
@@ -0,0 +1,26 @@
|
|
1
|
+
module Ubi
|
2
|
+
module Consultor
|
3
|
+
# URL finder consultor
|
4
|
+
class Yahoo < Base
|
5
|
+
#
|
6
|
+
# query
|
7
|
+
def query
|
8
|
+
{ p: @thema.name, hspart: 'ubi' }
|
9
|
+
end
|
10
|
+
|
11
|
+
def links
|
12
|
+
'//a[not (contains(@href,"yahoo"))]'
|
13
|
+
end
|
14
|
+
|
15
|
+
def words
|
16
|
+
'//li[@class="g"]'
|
17
|
+
end
|
18
|
+
|
19
|
+
class << self
|
20
|
+
def url
|
21
|
+
'https://search.yahoo.com/yhs/search?'
|
22
|
+
end
|
23
|
+
end
|
24
|
+
end
|
25
|
+
end
|
26
|
+
end
|