spidy 0.1.5 → 0.2.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/.ruby-version +1 -1
- data/Gemfile.lock +10 -1
- data/README.md +15 -0
- data/exe/spidy +4 -4
- data/lib/spidy.rb +6 -12
- data/lib/spidy/command_line.rb +94 -0
- data/lib/spidy/console.rb +4 -5
- data/lib/spidy/definition_file.rb +1 -1
- data/lib/spidy/interface.rb +12 -0
- data/lib/spidy/shell.rb +8 -84
- data/lib/spidy/version.rb +1 -1
- data/spidy.gemspec +1 -0
- metadata +19 -3
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA256:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 2ba417984708d3441c47f69574850c379e4a26dbcf810734545449173e1b886b
|
4
|
+
data.tar.gz: '0881f6202ffbcc8c1a7e153e79b16d7dec68694dbb915430cc3c0a8912e29817'
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 18ce16b8d0d5d3b536ccba74c842063c7c86d96fa9455010e58c78de20ed37e92051909a6054b77e7022c2c3072f9402fc246daf530c00287e22adbde2d598cf
|
7
|
+
data.tar.gz: e15ee2bec15d64dce143bad5eb0e59f177e32f4aeafed1626a288eaff690c028de7446e40a88bd16696b381ca08e912c959f5597eea11aeede78677ef3c9ed9e
|
data/.ruby-version
CHANGED
@@ -1 +1 @@
|
|
1
|
-
2.
|
1
|
+
2.6.5
|
data/Gemfile.lock
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
PATH
|
2
2
|
remote: .
|
3
3
|
specs:
|
4
|
-
spidy (0.1.
|
4
|
+
spidy (0.1.6)
|
5
5
|
activesupport
|
6
6
|
mechanize
|
7
7
|
pry
|
@@ -41,6 +41,7 @@ GEM
|
|
41
41
|
mime-types-data (3.2019.1009)
|
42
42
|
mini_portile2 (2.4.0)
|
43
43
|
minitest (5.14.0)
|
44
|
+
mixlib-shellout (2.4.4)
|
44
45
|
net-http-digest_auth (1.4.1)
|
45
46
|
net-http-persistent (3.1.0)
|
46
47
|
connection_pool (~> 2.2)
|
@@ -55,11 +56,18 @@ GEM
|
|
55
56
|
rspec-core (~> 3.8.0)
|
56
57
|
rspec-expectations (~> 3.8.0)
|
57
58
|
rspec-mocks (~> 3.8.0)
|
59
|
+
rspec-command (1.0.3)
|
60
|
+
mixlib-shellout (~> 2.0)
|
61
|
+
rspec (~> 3.2)
|
62
|
+
rspec-its (~> 1.2)
|
58
63
|
rspec-core (3.8.2)
|
59
64
|
rspec-support (~> 3.8.0)
|
60
65
|
rspec-expectations (3.8.4)
|
61
66
|
diff-lcs (>= 1.2.0, < 2.0)
|
62
67
|
rspec-support (~> 3.8.0)
|
68
|
+
rspec-its (1.3.0)
|
69
|
+
rspec-core (>= 3.0.0)
|
70
|
+
rspec-expectations (>= 3.0.0)
|
63
71
|
rspec-mocks (3.8.1)
|
64
72
|
diff-lcs (>= 1.2.0, < 2.0)
|
65
73
|
rspec-support (~> 3.8.0)
|
@@ -82,6 +90,7 @@ DEPENDENCIES
|
|
82
90
|
pry
|
83
91
|
rake (~> 10.0)
|
84
92
|
rspec (~> 3.0)
|
93
|
+
rspec-command
|
85
94
|
spidy!
|
86
95
|
|
87
96
|
BUNDLED WITH
|
data/README.md
CHANGED
@@ -44,6 +44,21 @@ cat urls | spidy call website.rb > website.json
|
|
44
44
|
echo 'http://example.com' | spidy each website.rb | spidy call website.rb | jq .
|
45
45
|
```
|
46
46
|
|
47
|
+
### When development console
|
48
|
+
```bash
|
49
|
+
spidy console website.rb
|
50
|
+
```
|
51
|
+
|
52
|
+
### reload source code
|
53
|
+
```
|
54
|
+
pry(#<Spidy::Console>)> reload!
|
55
|
+
```
|
56
|
+
|
57
|
+
```rb
|
58
|
+
each('http://example.com') { |url| break url }
|
59
|
+
call('http://example.com') { |html| break html } # html as nokogiri object ( mechanize )
|
60
|
+
```
|
61
|
+
|
47
62
|
### When used from the ruby code
|
48
63
|
``
|
49
64
|
a = Spidy.define do
|
data/exe/spidy
CHANGED
@@ -7,17 +7,17 @@ require 'pry'
|
|
7
7
|
if ARGV[1].blank?
|
8
8
|
case ARGV[0]
|
9
9
|
when 'version' then STDOUT.puts(Spidy::VERSION)
|
10
|
-
when 'console' then Spidy.
|
10
|
+
when 'console' then Spidy.shell.interactive
|
11
11
|
else
|
12
12
|
STDOUT.puts 'usage: spidy [version console]'
|
13
13
|
end
|
14
14
|
else
|
15
15
|
case ARGV[0]
|
16
|
-
when 'console' then Spidy.
|
17
|
-
when '
|
16
|
+
when 'console' then Spidy.shell(ARGV[1]).interactive
|
17
|
+
when 'function' then Spidy.shell(ARGV[1]).function
|
18
18
|
when 'call' then Spidy.shell(ARGV[1]).call(ARGV[2])
|
19
19
|
when 'each' then Spidy.shell(ARGV[1]).each(ARGV[2])
|
20
20
|
else
|
21
|
-
STDOUT.puts 'usage: spidy [call each
|
21
|
+
STDOUT.puts 'usage: spidy [console function call each] [file]'
|
22
22
|
end
|
23
23
|
end
|
data/lib/spidy.rb
CHANGED
@@ -10,36 +10,30 @@ require 'open-uri'
|
|
10
10
|
#
|
11
11
|
module Spidy
|
12
12
|
extend ActiveSupport::Autoload
|
13
|
+
autoload :Interface
|
13
14
|
autoload :Shell
|
15
|
+
autoload :CommandLine
|
14
16
|
autoload :Console
|
15
17
|
autoload :Definition
|
16
18
|
autoload :DefinitionFile
|
17
19
|
autoload :Binder
|
18
20
|
autoload :Connector
|
19
21
|
|
20
|
-
def self.
|
21
|
-
|
22
|
-
if filepath
|
23
|
-
Pry.start(Spidy::Console.new(Spidy::DefinitionFile.open(filepath)))
|
24
|
-
else
|
25
|
-
Pry.start(Spidy::Console.new)
|
26
|
-
end
|
22
|
+
def self.shell(filepath = nil)
|
23
|
+
Spidy::Shell.new(filepath)
|
27
24
|
end
|
28
25
|
|
29
26
|
def self.open(filepath)
|
30
27
|
Spidy::DefinitionFile.open(filepath).spidy
|
31
28
|
end
|
32
29
|
|
33
|
-
def self.shell(filepath)
|
34
|
-
Spidy::Shell.new(Spidy::DefinitionFile.open(filepath))
|
35
|
-
end
|
36
|
-
|
37
30
|
def self.define(&block)
|
38
|
-
Module.new do
|
31
|
+
spidy = Module.new do
|
39
32
|
class_eval do
|
40
33
|
extend ::Spidy::Definition
|
41
34
|
module_eval(&block)
|
42
35
|
end
|
43
36
|
end
|
37
|
+
Spidy::Interface.new(spidy)
|
44
38
|
end
|
45
39
|
end
|
@@ -0,0 +1,94 @@
|
|
1
|
+
# frozen_string_literal: true
|
2
|
+
|
3
|
+
#
|
4
|
+
# spidy shell interface
|
5
|
+
#
|
6
|
+
class Spidy::CommandLine
|
7
|
+
delegate :spidy, to: :@definition_file
|
8
|
+
class_attribute :output, default: (proc { |result| STDOUT.puts(result.to_s) })
|
9
|
+
class_attribute :error_handler, default: (proc { |e, url| STDERR.puts({ url: url, message: e.message, backtrace: e.backtrace }.to_json) })
|
10
|
+
|
11
|
+
def initialize(definition_file)
|
12
|
+
@definition_file = definition_file
|
13
|
+
raise 'unloaded spidy' if definition_file.spidy.nil?
|
14
|
+
end
|
15
|
+
|
16
|
+
def each_stdin_lines(name)
|
17
|
+
STDIN.each_line do |url|
|
18
|
+
begin
|
19
|
+
spidy.each(url.strip, name: name, &output)
|
20
|
+
rescue => e
|
21
|
+
error_handler.call(e, url)
|
22
|
+
end
|
23
|
+
end
|
24
|
+
end
|
25
|
+
|
26
|
+
def call_stdin_lines(name)
|
27
|
+
STDIN.each_line do |url|
|
28
|
+
begin
|
29
|
+
spidy.call(url.strip, name: name, &output)
|
30
|
+
rescue => e
|
31
|
+
error_handler.call(e, url)
|
32
|
+
end
|
33
|
+
end
|
34
|
+
end
|
35
|
+
|
36
|
+
def call(name)
|
37
|
+
return call_stdin_lines(name) if FileTest.pipe?(STDIN)
|
38
|
+
spidy.call(name: name, &output) unless FileTest.pipe?(STDIN)
|
39
|
+
rescue => e
|
40
|
+
error_handler.call(e, nil)
|
41
|
+
end
|
42
|
+
|
43
|
+
def each(name)
|
44
|
+
return each_stdin_lines(name) if FileTest.pipe?(STDIN)
|
45
|
+
spidy.each(name: name, &output)
|
46
|
+
rescue => e
|
47
|
+
error_handler.call(e, nil)
|
48
|
+
end
|
49
|
+
|
50
|
+
def function
|
51
|
+
print <<~SHELL
|
52
|
+
function spider() {
|
53
|
+
spidy spider #{definition_file.path} $1
|
54
|
+
}
|
55
|
+
function scraper() {
|
56
|
+
spidy call #{definition_file.path} $1
|
57
|
+
}
|
58
|
+
SHELL
|
59
|
+
end
|
60
|
+
|
61
|
+
def build(name)
|
62
|
+
build_shell(name)
|
63
|
+
build_ruby(name)
|
64
|
+
end
|
65
|
+
|
66
|
+
def build_shell(name)
|
67
|
+
File.open("#{name}.sh", 'w') do |f|
|
68
|
+
f.write <<~SHELL
|
69
|
+
#!/bin/bash
|
70
|
+
eval "$(spidy $(dirname "${0}")/#{name}.rb shell)"
|
71
|
+
spider example
|
72
|
+
SHELL
|
73
|
+
end
|
74
|
+
end
|
75
|
+
|
76
|
+
def build_ruby(name)
|
77
|
+
File.open("#{name}.rb", 'w') do |f|
|
78
|
+
f.write <<~RUBY
|
79
|
+
# frozen_string_literal: true
|
80
|
+
|
81
|
+
Spidy.define do
|
82
|
+
spider(:example) do |yielder, connector|
|
83
|
+
# connector.call(url) do |resource|
|
84
|
+
# yielder.call(url or resource)
|
85
|
+
# end
|
86
|
+
end
|
87
|
+
|
88
|
+
define(:example) do
|
89
|
+
end
|
90
|
+
end
|
91
|
+
RUBY
|
92
|
+
end
|
93
|
+
end
|
94
|
+
end
|
data/lib/spidy/console.rb
CHANGED
@@ -4,11 +4,10 @@
|
|
4
4
|
# spidy console
|
5
5
|
#
|
6
6
|
class Spidy::Console
|
7
|
-
|
8
|
-
delegate :
|
9
|
-
delegate :call, :each, to: :spidy
|
7
|
+
delegate :spidy, to: :@definition_file
|
8
|
+
delegate :call, :each, :namespace, allow_nil: true, to: :spidy
|
10
9
|
|
11
|
-
def initialize(definition_file
|
10
|
+
def initialize(definition_file)
|
12
11
|
@definition_file = definition_file
|
13
12
|
end
|
14
13
|
|
@@ -17,6 +16,6 @@ class Spidy::Console
|
|
17
16
|
end
|
18
17
|
|
19
18
|
def reload!
|
20
|
-
@definition_file
|
19
|
+
@definition_file.eval_definition
|
21
20
|
end
|
22
21
|
end
|
data/lib/spidy/shell.rb
CHANGED
@@ -1,96 +1,20 @@
|
|
1
1
|
# frozen_string_literal: true
|
2
2
|
|
3
|
-
require 'pry'
|
4
|
-
|
5
3
|
#
|
6
|
-
# spidy
|
4
|
+
# spidy Shell
|
7
5
|
#
|
8
6
|
class Spidy::Shell
|
9
|
-
|
10
|
-
|
11
|
-
class_attribute :error_handler, default: (proc { |e, url| STDERR.puts({ url: url, message: e.message, backtrace: e.backtrace }.to_json) })
|
12
|
-
delegate :spidy, to: :definition_file
|
13
|
-
|
14
|
-
def initialize(definition_file)
|
15
|
-
@definition_file = definition_file
|
16
|
-
end
|
17
|
-
|
18
|
-
def each_stdin_lines(name)
|
19
|
-
STDIN.each_line do |url|
|
20
|
-
begin
|
21
|
-
spidy.each(url.strip, name: name, &output)
|
22
|
-
rescue => e
|
23
|
-
error_handler.call(e, url)
|
24
|
-
end
|
25
|
-
end
|
26
|
-
end
|
27
|
-
|
28
|
-
def call_stdin_lines(name)
|
29
|
-
STDIN.each_line do |url|
|
30
|
-
begin
|
31
|
-
spidy.call(url.strip, name: name, &output)
|
32
|
-
rescue => e
|
33
|
-
error_handler.call(e, url)
|
34
|
-
end
|
35
|
-
end
|
36
|
-
end
|
37
|
-
|
38
|
-
def call(name)
|
39
|
-
return call_stdin_lines(name) if FileTest.pipe?(STDIN)
|
40
|
-
spidy.call(name: name, &output) unless FileTest.pipe?(STDIN)
|
41
|
-
rescue => e
|
42
|
-
error_handler.call(e, nil)
|
7
|
+
def initialize(path)
|
8
|
+
@definition_file = Spidy::DefinitionFile.open(path)
|
43
9
|
end
|
44
10
|
|
45
|
-
def
|
46
|
-
|
47
|
-
spidy.each(name: name, &output)
|
48
|
-
rescue => e
|
49
|
-
error_handler.call(e, nil)
|
11
|
+
def interactive
|
12
|
+
Pry.start(Spidy::Console.new(@definition_file))
|
50
13
|
end
|
51
14
|
|
52
|
-
def
|
53
|
-
|
54
|
-
function spider() {
|
55
|
-
spidy spider #{definition_file.path} $1
|
56
|
-
}
|
57
|
-
function scraper() {
|
58
|
-
spidy call #{definition_file.path} $1
|
59
|
-
}
|
60
|
-
SHELL
|
15
|
+
def command_line
|
16
|
+
Spidy::CommandLine.new(@definition_file)
|
61
17
|
end
|
62
18
|
|
63
|
-
|
64
|
-
build_shell(name)
|
65
|
-
build_ruby(name)
|
66
|
-
end
|
67
|
-
|
68
|
-
def build_shell(name)
|
69
|
-
File.open("#{name}.sh", 'w') do |f|
|
70
|
-
f.write <<~SHELL
|
71
|
-
#!/bin/bash
|
72
|
-
eval "$(spidy $(dirname "${0}")/#{name}.rb shell)"
|
73
|
-
spider example
|
74
|
-
SHELL
|
75
|
-
end
|
76
|
-
end
|
77
|
-
|
78
|
-
def build_ruby(name)
|
79
|
-
File.open("#{name}.rb", 'w') do |f|
|
80
|
-
f.write <<~RUBY
|
81
|
-
# frozen_string_literal: true
|
82
|
-
|
83
|
-
Spidy.define do
|
84
|
-
spider(:example) do |yielder, connector|
|
85
|
-
# connector.call(url) do |resource|
|
86
|
-
# yielder.call(url or resource)
|
87
|
-
# end
|
88
|
-
end
|
89
|
-
|
90
|
-
define(:example) do
|
91
|
-
end
|
92
|
-
end
|
93
|
-
RUBY
|
94
|
-
end
|
95
|
-
end
|
19
|
+
delegate :function, :each, :call, to: :command_line
|
96
20
|
end
|
data/lib/spidy/version.rb
CHANGED
data/spidy.gemspec
CHANGED
@@ -29,6 +29,7 @@ Gem::Specification.new do |spec|
|
|
29
29
|
spec.add_development_dependency 'rake', '~> 10.0'
|
30
30
|
spec.add_development_dependency 'rspec', '~> 3.0'
|
31
31
|
spec.add_development_dependency 'ffaker'
|
32
|
+
spec.add_development_dependency 'rspec-command'
|
32
33
|
|
33
34
|
spec.add_runtime_dependency 'activesupport'
|
34
35
|
spec.add_runtime_dependency 'mechanize'
|
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: spidy
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.
|
4
|
+
version: 0.2.0
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- aileron
|
8
8
|
autorequire:
|
9
9
|
bindir: exe
|
10
10
|
cert_chain: []
|
11
|
-
date: 2020-02-
|
11
|
+
date: 2020-02-10 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
14
|
name: bundler
|
@@ -80,6 +80,20 @@ dependencies:
|
|
80
80
|
- - ">="
|
81
81
|
- !ruby/object:Gem::Version
|
82
82
|
version: '0'
|
83
|
+
- !ruby/object:Gem::Dependency
|
84
|
+
name: rspec-command
|
85
|
+
requirement: !ruby/object:Gem::Requirement
|
86
|
+
requirements:
|
87
|
+
- - ">="
|
88
|
+
- !ruby/object:Gem::Version
|
89
|
+
version: '0'
|
90
|
+
type: :development
|
91
|
+
prerelease: false
|
92
|
+
version_requirements: !ruby/object:Gem::Requirement
|
93
|
+
requirements:
|
94
|
+
- - ">="
|
95
|
+
- !ruby/object:Gem::Version
|
96
|
+
version: '0'
|
83
97
|
- !ruby/object:Gem::Dependency
|
84
98
|
name: activesupport
|
85
99
|
requirement: !ruby/object:Gem::Requirement
|
@@ -151,6 +165,7 @@ files:
|
|
151
165
|
- lib/spidy/binder/html.rb
|
152
166
|
- lib/spidy/binder/json.rb
|
153
167
|
- lib/spidy/binder/xml.rb
|
168
|
+
- lib/spidy/command_line.rb
|
154
169
|
- lib/spidy/connector.rb
|
155
170
|
- lib/spidy/connector/direct.rb
|
156
171
|
- lib/spidy/connector/html.rb
|
@@ -159,6 +174,7 @@ files:
|
|
159
174
|
- lib/spidy/console.rb
|
160
175
|
- lib/spidy/definition.rb
|
161
176
|
- lib/spidy/definition_file.rb
|
177
|
+
- lib/spidy/interface.rb
|
162
178
|
- lib/spidy/shell.rb
|
163
179
|
- lib/spidy/spider.rb
|
164
180
|
- lib/spidy/version.rb
|
@@ -183,7 +199,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
|
|
183
199
|
- !ruby/object:Gem::Version
|
184
200
|
version: '0'
|
185
201
|
requirements: []
|
186
|
-
rubygems_version: 3.
|
202
|
+
rubygems_version: 3.0.3
|
187
203
|
signing_key:
|
188
204
|
specification_version: 4
|
189
205
|
summary: web spider dsl
|