preadly-bulbasaur 0.8.1 → 0.8.2
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
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA1:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 8260cd6a3e468ab1139c0096ee1cbb5086fe7ff1
|
4
|
+
data.tar.gz: 146d9e064d05db1eb3c19c7cdc7290c6027fa79a
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 61721cf5b99c6edba03b754fe98b863bd9747ab0bd279d4b3023becfc160309a591d931a2b8b814aab69884c940e026c72d8dc52022769101ef9f79f4da90b49
|
7
|
+
data.tar.gz: f8da5d762adf404102767f71c0c68ea99a3f49fbbb716612f01e95f9fda6ab8a5c982068d35d14232bdb8d033bfdcefab69d628c36a3fbdeb54dab390ab490d7
|
@@ -1,7 +1,5 @@
|
|
1
1
|
module Bulbasaur
|
2
2
|
class NormalizeImageSources
|
3
|
-
DOMAIN_REGEX = /^(?:https?:\/\/)?(?:[^@\n]+@)?(?:www\.)?([^:\/\n]+)/im
|
4
|
-
|
5
3
|
def initialize(html, target_attrs)
|
6
4
|
@html = html
|
7
5
|
@target_attrs = target_attrs
|
@@ -27,16 +25,10 @@ module Bulbasaur
|
|
27
25
|
end
|
28
26
|
|
29
27
|
def adjust(element, attr)
|
30
|
-
element.set_attribute 'src',
|
28
|
+
element.set_attribute 'src', element.xpath(attr).text
|
31
29
|
remove_target_attrs_from element
|
32
30
|
end
|
33
31
|
|
34
|
-
def lazy_load_url(element, text)
|
35
|
-
text_match = text.match(DOMAIN_REGEX).to_s
|
36
|
-
element_match = element.css('@src').text.match(DOMAIN_REGEX).to_s
|
37
|
-
element_match.empty? || text_match == element_match ? text : "#{element_match}/#{text}"
|
38
|
-
end
|
39
|
-
|
40
32
|
def remove_target_attrs_from(element)
|
41
33
|
@target_attrs.each { |attr| element.xpath("@#{attr}").remove }
|
42
34
|
end
|
@@ -1,7 +1,7 @@
|
|
1
1
|
module Bulbasaur
|
2
2
|
class NormalizeURL
|
3
3
|
def initialize(base_url, context_url)
|
4
|
-
@base_url = base_url
|
4
|
+
@base_url = clean_url(base_url)
|
5
5
|
@context_url = context_url
|
6
6
|
end
|
7
7
|
|
@@ -11,5 +11,13 @@ module Bulbasaur
|
|
11
11
|
rescue
|
12
12
|
raise ArgumentError, "Unable to normalize URL. Params: [base_url: #{@base_url}, context_url: #{@context_url}]."
|
13
13
|
end
|
14
|
+
|
15
|
+
private
|
16
|
+
def clean_url(url)
|
17
|
+
uri = URI(url)
|
18
|
+
uri.path = ''
|
19
|
+
uri.query = ''
|
20
|
+
uri.to_s
|
21
|
+
end
|
14
22
|
end
|
15
23
|
end
|
data/lib/bulbasaur/version.rb
CHANGED
@@ -9,15 +9,16 @@ RSpec.describe Bulbasaur::NormalizeImageSources do
|
|
9
9
|
describe "#call" do
|
10
10
|
|
11
11
|
let(:html) do
|
12
|
-
|
12
|
+
<<-HTML
|
13
13
|
<img src="http://somewhere.to/get/a-pixel.gif" data-lazy-src="http://somewhere.to/get/the-real-image.jpg" alt="Image" width="800" height="1200">
|
14
14
|
<img src="http://somewhere.to/get/another-pixel.gif" data-image="http://somewhere.to/get/the-other-real-image.jpg">
|
15
15
|
<img src="http://somewhere.to/get/a-third-pixel.gif" data-src="get/the-third-real-image.jpg">
|
16
16
|
<img src="otherplace.to/load/a-fourth-pixel.gif" lazy-data="otherplace.to/load/the-fourth-real-image.jpg">
|
17
|
+
<img src="http://somewhere.to/get/a-fifth-pixel.gif" data-image="http://elsewhere.to/get/the-fifth-real-image.jpg">
|
17
18
|
<img src="https://place.where/an-image/is/without-extension/" data-src="https://place.where/an-image/is/without-extension/">
|
18
19
|
<img src="https://place.where/an-image/has-two/lazy-params/" data-src="https://place.where/an-image/has-two/lazy-params/" data-image="https://place.where/an-image/has-two/lazy-params/">
|
19
20
|
<img lazy-data="https://place.where/an-image/has/no-src-tag.jpg">
|
20
|
-
|
21
|
+
HTML
|
21
22
|
end
|
22
23
|
|
23
24
|
context "When there are no target attributes" do
|
@@ -38,15 +39,16 @@ RSpec.describe Bulbasaur::NormalizeImageSources do
|
|
38
39
|
end
|
39
40
|
|
40
41
|
let(:html_parsed) do
|
41
|
-
|
42
|
+
<<-HTML
|
42
43
|
<img src="http://somewhere.to/get/the-real-image.jpg" alt="Image" width="800" height="1200">
|
43
44
|
<img src="http://somewhere.to/get/the-other-real-image.jpg">
|
44
45
|
<img src="http://somewhere.to/get/a-third-pixel.gif" data-src="get/the-third-real-image.jpg">
|
45
46
|
<img src="otherplace.to/load/a-fourth-pixel.gif" lazy-data="otherplace.to/load/the-fourth-real-image.jpg">
|
47
|
+
<img src="http://elsewhere.to/get/the-fifth-real-image.jpg">
|
46
48
|
<img src="https://place.where/an-image/is/without-extension/" data-src="https://place.where/an-image/is/without-extension/">
|
47
49
|
<img src="https://place.where/an-image/has-two/lazy-params/" data-src="https://place.where/an-image/has-two/lazy-params/">
|
48
50
|
<img lazy-data="https://place.where/an-image/has/no-src-tag.jpg">
|
49
|
-
|
51
|
+
HTML
|
50
52
|
end
|
51
53
|
|
52
54
|
it "Returns the HTML code with the specified image tags adjusted" do
|
@@ -61,15 +63,16 @@ RSpec.describe Bulbasaur::NormalizeImageSources do
|
|
61
63
|
end
|
62
64
|
|
63
65
|
let(:html_parsed) do
|
64
|
-
|
66
|
+
<<-HTML
|
65
67
|
<img src="http://somewhere.to/get/the-real-image.jpg" alt="Image" width="800" height="1200">
|
66
68
|
<img src="http://somewhere.to/get/the-other-real-image.jpg">
|
67
|
-
<img src="
|
69
|
+
<img src="get/the-third-real-image.jpg">
|
68
70
|
<img src="otherplace.to/load/the-fourth-real-image.jpg">
|
71
|
+
<img src="http://elsewhere.to/get/the-fifth-real-image.jpg">
|
69
72
|
<img src="https://place.where/an-image/is/without-extension/">
|
70
73
|
<img src="https://place.where/an-image/has-two/lazy-params/">
|
71
74
|
<img src="https://place.where/an-image/has/no-src-tag.jpg">
|
72
|
-
|
75
|
+
HTML
|
73
76
|
end
|
74
77
|
|
75
78
|
it "Returns the HTML code with the involved image tags fixed with domain and path" do
|
@@ -5,7 +5,7 @@ RSpec.describe Bulbasaur::NormalizeURL do
|
|
5
5
|
subject do
|
6
6
|
described_class.new(base_url, context_url).call
|
7
7
|
end
|
8
|
-
|
8
|
+
|
9
9
|
let(:base_url) do
|
10
10
|
"http://pread.ly"
|
11
11
|
end
|
@@ -15,16 +15,16 @@ RSpec.describe Bulbasaur::NormalizeURL do
|
|
15
15
|
end
|
16
16
|
|
17
17
|
describe "#call" do
|
18
|
-
|
18
|
+
|
19
19
|
context "When use url normalized url: http://www.test.com/hello.jpg" do
|
20
|
-
|
20
|
+
|
21
21
|
it "Does return url normalized: http://www.test.com/hello.jpg" do
|
22
22
|
expect(subject).to eq "http://www.test.com/hello.jpg"
|
23
23
|
end
|
24
24
|
end
|
25
25
|
|
26
26
|
context "When use url unnormalized url: test.jpg" do
|
27
|
-
|
27
|
+
|
28
28
|
let(:context_url) do
|
29
29
|
"test.jpg"
|
30
30
|
end
|
@@ -35,7 +35,7 @@ RSpec.describe Bulbasaur::NormalizeURL do
|
|
35
35
|
end
|
36
36
|
|
37
37
|
context "When use url https normalized: https://www.test.com/hello.jpg" do
|
38
|
-
|
38
|
+
|
39
39
|
let(:context_url) do
|
40
40
|
"https://www.test.com/hello.jpg"
|
41
41
|
end
|
@@ -54,7 +54,7 @@ RSpec.describe Bulbasaur::NormalizeURL do
|
|
54
54
|
end
|
55
55
|
|
56
56
|
context "When use url not normalized with slash on base: hello.jpg" do
|
57
|
-
|
57
|
+
|
58
58
|
let(:base_url) do
|
59
59
|
"https://www.test.com/"
|
60
60
|
end
|
@@ -68,8 +68,36 @@ RSpec.describe Bulbasaur::NormalizeURL do
|
|
68
68
|
end
|
69
69
|
end
|
70
70
|
|
71
|
+
context "When base_url contains path" do
|
72
|
+
let!(:base_url) do
|
73
|
+
"http://pread.ly/other-path/"
|
74
|
+
end
|
75
|
+
|
76
|
+
let!(:context_url) do
|
77
|
+
"test.jpg"
|
78
|
+
end
|
79
|
+
|
80
|
+
it "Does return url normalized: http://pread.ly/test.jpg" do
|
81
|
+
expect(subject).to eq "http://pread.ly/test.jpg"
|
82
|
+
end
|
83
|
+
end
|
84
|
+
|
85
|
+
context "When base_url contains query string" do
|
86
|
+
let!(:base_url) do
|
87
|
+
"http://pread.ly/?page=1"
|
88
|
+
end
|
89
|
+
|
90
|
+
let!(:context_url) do
|
91
|
+
"test.jpg"
|
92
|
+
end
|
93
|
+
|
94
|
+
it "Does return url normalized: http://pread.ly/test.jpg" do
|
95
|
+
expect(subject).to eq "http://pread.ly/test.jpg"
|
96
|
+
end
|
97
|
+
end
|
98
|
+
|
71
99
|
context "When base url not valid" do
|
72
|
-
|
100
|
+
|
73
101
|
let(:base_url) do
|
74
102
|
"test/httml"
|
75
103
|
end
|
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: preadly-bulbasaur
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.8.
|
4
|
+
version: 0.8.2
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Magno Costa
|
8
8
|
autorequire:
|
9
9
|
bindir: bin
|
10
10
|
cert_chain: []
|
11
|
-
date: 2015-
|
11
|
+
date: 2015-09-15 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
14
|
name: bundler
|