metacrunch-mab2 1.0.1 → 1.1.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/.gitignore +22 -3
- data/Gemfile +17 -5
- data/Rakefile +1 -0
- data/bin/console +11 -0
- data/bin/setup +7 -0
- data/lib/metacrunch/mab2.rb +3 -7
- data/lib/metacrunch/mab2/builder.rb +52 -0
- data/lib/metacrunch/mab2/document.rb +49 -45
- data/lib/metacrunch/mab2/{aleph_mab_xml_document_factory.rb → document/aleph_mab_xml_parser.rb} +9 -16
- data/lib/metacrunch/mab2/document/controlfield.rb +8 -4
- data/lib/metacrunch/mab2/document/datafield.rb +21 -18
- data/lib/metacrunch/mab2/document/datafield_set.rb +46 -36
- data/lib/metacrunch/mab2/document/subfield.rb +13 -16
- data/lib/metacrunch/mab2/document/subfield_set.rb +39 -36
- data/lib/metacrunch/mab2/version.rb +1 -1
- data/lib/metacrunch/mab2/xml_builder.rb +48 -0
- data/metacrunch-mab2.gemspec +2 -0
- data/spec/{document → metacrunch/mab2/document}/control_field_spec.rb +2 -2
- data/spec/{document → metacrunch/mab2}/document_spec.rb +31 -6
- metadata +44 -11
- data/spec/document/aleph_mab_xml_document_factory_spec.rb +0 -37
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA1:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 60d656f23a0579c4c1e0a01c23aec38904fbe529
|
4
|
+
data.tar.gz: 84aeda9371bedc709734eade5d5ae5111d88c76b
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 74fabeef356246a5e3eb56961f4a3fabd7b89e915bed639e2f9e08b9bf3676b34c5231a451ad791044f4d21de76e2679b50ccb90e61aab9ad42c4d2c6db4415c
|
7
|
+
data.tar.gz: f263ffe68b46f4c8bf85bf66514b761452b529c924bd675367349825d6eccea8c6b4afda9cd2593c08ae64cef8671964b86e46334710b44778a134bfc9e84095
|
data/.gitignore
CHANGED
@@ -1,5 +1,24 @@
|
|
1
1
|
.DS_Store
|
2
|
-
/Gemfile.lock
|
3
|
-
/.yardoc
|
4
2
|
/doc
|
5
|
-
|
3
|
+
*.gem
|
4
|
+
*.rbc
|
5
|
+
.bundle
|
6
|
+
.config
|
7
|
+
.yardoc
|
8
|
+
Gemfile.lock
|
9
|
+
InstalledFiles
|
10
|
+
_yardoc
|
11
|
+
coverage
|
12
|
+
doc/
|
13
|
+
lib/bundler/man
|
14
|
+
pkg
|
15
|
+
rdoc
|
16
|
+
spec/reports
|
17
|
+
test/tmp
|
18
|
+
test/version_tmp
|
19
|
+
tmp
|
20
|
+
*.bundle
|
21
|
+
*.so
|
22
|
+
*.o
|
23
|
+
*.a
|
24
|
+
mkmf.log
|
data/Gemfile
CHANGED
@@ -4,10 +4,22 @@ gemspec
|
|
4
4
|
|
5
5
|
gem "metacrunch", ">= 2.1.0", github: "ubpb/metacrunch", branch: "master"
|
6
6
|
|
7
|
-
|
8
|
-
gem "
|
9
|
-
gem "
|
10
|
-
gem "
|
7
|
+
group :development do
|
8
|
+
gem "benchmark-ips"
|
9
|
+
gem "bundler"
|
10
|
+
gem "rake"
|
11
|
+
gem "rspec", ">= 3.0.0", "< 4.0.0"
|
12
|
+
gem "simplecov", ">= 0.8.0"
|
13
|
+
|
14
|
+
if !ENV["CI"]
|
15
|
+
gem "hashdiff"
|
16
|
+
gem "pry", "~> 0.9.12.6"
|
17
|
+
gem "pry-byebug", "<= 1.3.2"
|
18
|
+
gem "pry-rescue", "~> 1.4.2"
|
19
|
+
gem "pry-stack_explorer", "~> 0.4.9.1"
|
20
|
+
gem "pry-syntax-hacks", "~> 0.0.6"
|
21
|
+
end
|
22
|
+
end
|
11
23
|
|
12
24
|
# for benchmarking
|
13
|
-
gem "mabmapper", github: "ubpb/mabmapper", branch: "master"
|
25
|
+
# gem "mabmapper", github: "ubpb/mabmapper", branch: "master"
|
data/Rakefile
CHANGED
data/bin/console
ADDED
@@ -0,0 +1,11 @@
|
|
1
|
+
#!/usr/bin/env ruby
|
2
|
+
|
3
|
+
require "bundler/setup"
|
4
|
+
require "metacrunch/mab2"
|
5
|
+
|
6
|
+
# You can add fixtures and/or initialization code here to make experimenting
|
7
|
+
# with your gem easier. You can also use a different console, if you like.
|
8
|
+
|
9
|
+
# (If you use this, don't forget to add pry to your Gemfile!)
|
10
|
+
require "pry"
|
11
|
+
Pry.start
|
data/bin/setup
ADDED
data/lib/metacrunch/mab2.rb
CHANGED
@@ -2,15 +2,11 @@ require "metacrunch"
|
|
2
2
|
require "ox"
|
3
3
|
require "htmlentities"
|
4
4
|
|
5
|
-
begin
|
6
|
-
require "pry"
|
7
|
-
rescue LoadError ; end
|
8
|
-
|
9
5
|
module Metacrunch
|
10
6
|
module Mab2
|
11
|
-
require_relative "./mab2/
|
12
|
-
require_relative "./mab2/document"
|
7
|
+
require_relative "./mab2/builder"
|
13
8
|
require_relative "./mab2/cli"
|
14
|
-
require_relative "./mab2/
|
9
|
+
require_relative "./mab2/document"
|
10
|
+
require_relative "./mab2/version"
|
15
11
|
end
|
16
12
|
end
|
@@ -0,0 +1,52 @@
|
|
1
|
+
require_relative "./document"
|
2
|
+
|
3
|
+
module Metacrunch
|
4
|
+
module Mab2
|
5
|
+
class Builder
|
6
|
+
|
7
|
+
attr_reader :document
|
8
|
+
|
9
|
+
def self.build(&block)
|
10
|
+
builder = new
|
11
|
+
document = builder.instance_eval(&block) if block_given?
|
12
|
+
document || Document.new
|
13
|
+
end
|
14
|
+
|
15
|
+
def initialize
|
16
|
+
@document = Document.new
|
17
|
+
end
|
18
|
+
|
19
|
+
def controlfield(tag, values)
|
20
|
+
controlfield = Document::Controlfield.new(tag, values)
|
21
|
+
@document.add_controlfield(controlfield)
|
22
|
+
@document
|
23
|
+
end
|
24
|
+
|
25
|
+
def datafield(tag, ind1:nil, ind2:nil, &block)
|
26
|
+
datafield = Document::Datafield.new(tag, ind1: ind1, ind2: ind2)
|
27
|
+
@document.add_datafield(datafield)
|
28
|
+
|
29
|
+
if block_given?
|
30
|
+
subfield_builder = SubfieldBuilder.new(datafield)
|
31
|
+
subfield_builder.instance_eval(&block)
|
32
|
+
end
|
33
|
+
|
34
|
+
@document
|
35
|
+
end
|
36
|
+
|
37
|
+
|
38
|
+
class SubfieldBuilder
|
39
|
+
def initialize(datafield)
|
40
|
+
@datafield = datafield
|
41
|
+
end
|
42
|
+
|
43
|
+
def subfield(code, value)
|
44
|
+
subfield = Document::Subfield.new(code, value)
|
45
|
+
@datafield.add_subfield(subfield)
|
46
|
+
subfield
|
47
|
+
end
|
48
|
+
end
|
49
|
+
|
50
|
+
end
|
51
|
+
end
|
52
|
+
end
|
@@ -1,6 +1,7 @@
|
|
1
1
|
module Metacrunch
|
2
2
|
module Mab2
|
3
3
|
class Document
|
4
|
+
require_relative "./document/aleph_mab_xml_parser"
|
4
5
|
require_relative "./document/controlfield"
|
5
6
|
require_relative "./document/datafield"
|
6
7
|
require_relative "./document/datafield_set"
|
@@ -16,31 +17,26 @@ module Metacrunch
|
|
16
17
|
# @return [Metacrunch::Mab2::Document]
|
17
18
|
#
|
18
19
|
def self.from_aleph_mab_xml(xml)
|
19
|
-
|
20
|
+
AlephMabXmlParser.parse(xml)
|
21
|
+
end
|
22
|
+
|
23
|
+
def initialize
|
24
|
+
@controlfields = {}
|
25
|
+
@datafields = {}
|
20
26
|
end
|
21
27
|
|
22
28
|
# ------------------------------------------------------------------------------
|
23
29
|
# Control fields
|
24
30
|
# ------------------------------------------------------------------------------
|
25
31
|
|
26
|
-
#
|
27
|
-
# @return [Hash{String => Metacrunch::Mab2::Document::Controlfield}]
|
28
|
-
# @private
|
29
|
-
#
|
30
|
-
def controlfields_struct
|
31
|
-
@controlfields_struct ||= {}
|
32
|
-
end
|
33
|
-
private :controlfields_struct
|
34
|
-
|
35
32
|
#
|
36
33
|
# Returns the control field matching the given tag.
|
37
34
|
#
|
38
35
|
# @param [String] tag of the control field
|
39
|
-
# @return [Controlfield
|
40
|
-
# if the control field doesn't exists.
|
36
|
+
# @return [Controlfield] control field with the given tag.
|
41
37
|
#
|
42
38
|
def controlfield(tag)
|
43
|
-
|
39
|
+
@controlfields[tag] || Controlfield.new(tag)
|
44
40
|
end
|
45
41
|
|
46
42
|
#
|
@@ -49,44 +45,35 @@ module Metacrunch
|
|
49
45
|
# @param [Metacrunch::Mab2::Document::Controlfield] controlfield
|
50
46
|
#
|
51
47
|
def add_controlfield(controlfield)
|
52
|
-
|
48
|
+
@controlfields[controlfield.tag] = controlfield
|
53
49
|
end
|
54
50
|
|
55
51
|
# ------------------------------------------------------------------------------
|
56
52
|
# Data fields
|
57
53
|
# ------------------------------------------------------------------------------
|
58
54
|
|
59
|
-
#
|
60
|
-
# @return [Hash{String => Metacrunch::Mab2::Document::Datafield::Set}]
|
61
|
-
# @private
|
62
|
-
#
|
63
|
-
def datafields_struct
|
64
|
-
@datafields_struct ||= {}
|
65
|
-
end
|
66
|
-
private :datafields_struct
|
67
|
-
|
68
55
|
#
|
69
56
|
# Returns the data fields matching the given tag and ind1/ind2.
|
70
57
|
#
|
71
|
-
# @param [String] tag of the data field
|
72
|
-
# @param [String, nil] ind1 filter for ind1. Can be nil to match any.
|
73
|
-
# @param [String, nil] ind2 filter for ind2. Can be nil to match any.
|
74
|
-
# @return [Metacrunch::Mab2::Document::
|
58
|
+
# @param [String, nil] tag of the data field. Can be nil to match any data field.
|
59
|
+
# @param [String, nil] ind1 filter for ind1. Can be nil to match any indicator 1.
|
60
|
+
# @param [String, nil] ind2 filter for ind2. Can be nil to match any indicator 2.
|
61
|
+
# @return [Metacrunch::Mab2::Document::DatafieldSet] data field with the given tag and ind1/ind2.
|
75
62
|
# The set is empty if a matching field with the tag and/or ind1/ind2 doesn't exists.
|
76
63
|
#
|
77
|
-
def datafields(tag, ind1: nil, ind2: nil)
|
78
|
-
|
79
|
-
|
80
|
-
|
81
|
-
if ind1 || ind2
|
82
|
-
filtered_datafields = set.select do |datafield|
|
83
|
-
(!ind1 || (ind1 == :blank ? datafield.ind1 == " " || datafield.ind1 == "-" : datafield.ind1 == ind1)) &&
|
84
|
-
(!ind2 || (ind2 == :blank ? datafield.ind2 == " " || datafield.ind2 == "-" : datafield.ind2 == ind2))
|
85
|
-
end
|
86
|
-
|
87
|
-
Datafield::Set.new(filtered_datafields)
|
64
|
+
def datafields(tag = nil, ind1: nil, ind2: nil)
|
65
|
+
if tag.nil?
|
66
|
+
DatafieldSet.new(@datafields.values.flatten(1))
|
88
67
|
else
|
89
|
-
set
|
68
|
+
set = DatafieldSet.new(@datafields[tag] || [])
|
69
|
+
return set if set.empty? || (ind1.nil? && ind2.nil?)
|
70
|
+
|
71
|
+
ind1 = map_indicator(ind1)
|
72
|
+
ind2 = map_indicator(ind2)
|
73
|
+
|
74
|
+
set.select do |_datafield|
|
75
|
+
check_indicator(ind1, _datafield.ind1) && check_indicator(ind2, _datafield.ind2)
|
76
|
+
end
|
90
77
|
end
|
91
78
|
end
|
92
79
|
|
@@ -96,10 +83,7 @@ module Metacrunch
|
|
96
83
|
# @param [Metacrunch::Mab2::Document::Datafield] datafield
|
97
84
|
#
|
98
85
|
def add_datafield(datafield)
|
99
|
-
|
100
|
-
datafield_set << datafield
|
101
|
-
|
102
|
-
datafields_struct[datafield.tag] = datafield_set
|
86
|
+
(@datafields[datafield.tag] ||= []) << datafield
|
103
87
|
end
|
104
88
|
|
105
89
|
# ------------------------------------------------------------------------------
|
@@ -107,19 +91,39 @@ module Metacrunch
|
|
107
91
|
# ------------------------------------------------------------------------------
|
108
92
|
|
109
93
|
def to_xml
|
110
|
-
builder = Builder::XmlMarkup.new(indent: 2)
|
94
|
+
builder = ::Builder::XmlMarkup.new(indent: 2)
|
111
95
|
builder.instruct!(:xml, :encoding => "UTF-8")
|
112
96
|
builder.mab_xml do
|
113
97
|
controlfields_struct.values.each do |_controlfield|
|
114
98
|
_controlfield.to_xml(builder)
|
115
99
|
end
|
116
100
|
|
117
|
-
|
101
|
+
@datafields.values.each do |_datafield_set|
|
118
102
|
_datafield_set.to_xml(builder)
|
119
103
|
end
|
120
104
|
end
|
121
105
|
end
|
122
106
|
|
107
|
+
private
|
108
|
+
|
109
|
+
def map_indicator(ind)
|
110
|
+
ind.is_a?(Array) ? ind.map { |_el| _el == :blank ? [" ", "-", nil] : _el }.flatten(1) : ind
|
111
|
+
end
|
112
|
+
|
113
|
+
def check_indicator(requested_ind, datafield_ind)
|
114
|
+
if !requested_ind
|
115
|
+
true
|
116
|
+
elsif requested_ind == :blank && (datafield_ind == " " || datafield_ind == "-" || datafield_ind.nil?)
|
117
|
+
true
|
118
|
+
elsif requested_ind == datafield_ind
|
119
|
+
true
|
120
|
+
elsif requested_ind.is_a?(Array) && requested_ind.include?(datafield_ind)
|
121
|
+
true
|
122
|
+
else
|
123
|
+
false
|
124
|
+
end
|
125
|
+
end
|
126
|
+
|
123
127
|
end
|
124
128
|
end
|
125
129
|
end
|
data/lib/metacrunch/mab2/{aleph_mab_xml_document_factory.rb → document/aleph_mab_xml_parser.rb}
RENAMED
@@ -1,22 +1,15 @@
|
|
1
1
|
require "htmlentities"
|
2
2
|
require "ox"
|
3
|
-
require_relative "
|
3
|
+
require_relative "../document"
|
4
4
|
|
5
5
|
module Metacrunch
|
6
6
|
module Mab2
|
7
|
-
class
|
8
|
-
|
9
|
-
|
10
|
-
|
11
|
-
|
12
|
-
|
13
|
-
def to_document
|
14
|
-
@document ||= Parser.new.parse(@aleph_mab_xml)
|
15
|
-
end
|
16
|
-
|
17
|
-
private
|
7
|
+
class Document
|
8
|
+
class AlephMabXmlParser < Ox::Sax
|
9
|
+
def self.parse(aleph_mab_xml)
|
10
|
+
new.parse(aleph_mab_xml)
|
11
|
+
end
|
18
12
|
|
19
|
-
class Parser < Ox::Sax
|
20
13
|
def parse(io_or_string)
|
21
14
|
# initialize state machine
|
22
15
|
@in_controlfield = @in_datafield = @in_subfield = false
|
@@ -37,13 +30,13 @@ module Metacrunch
|
|
37
30
|
def start_element(name)
|
38
31
|
if name == :subfield
|
39
32
|
@in_subfield = true
|
40
|
-
@subfield = Document::
|
33
|
+
@subfield = Metacrunch::Mab2::Document::Subfield.new
|
41
34
|
elsif name == :datafield
|
42
35
|
@in_datafield = true
|
43
|
-
@datafield = Document::Datafield.new
|
36
|
+
@datafield = Metacrunch::Mab2::Document::Datafield.new
|
44
37
|
elsif name == :controlfield
|
45
38
|
@in_controlfield = true
|
46
|
-
@controlfield = Document::Controlfield.new
|
39
|
+
@controlfield = Metacrunch::Mab2::Document::Controlfield.new
|
47
40
|
end
|
48
41
|
end
|
49
42
|
|
@@ -60,7 +60,7 @@ module Metacrunch
|
|
60
60
|
# ------------------------------------------------------------------------------
|
61
61
|
|
62
62
|
def to_xml(builder)
|
63
|
-
builder.controlfield(
|
63
|
+
builder.controlfield(values2string, tag: tag)
|
64
64
|
end
|
65
65
|
|
66
66
|
private
|
@@ -69,11 +69,11 @@ module Metacrunch
|
|
69
69
|
string.chars.to_a.map{ |e| (e=="|") ? nil : e }
|
70
70
|
end
|
71
71
|
|
72
|
-
def array2values(
|
73
|
-
|
72
|
+
def array2values(array)
|
73
|
+
array.map do |v|
|
74
74
|
if v.present?
|
75
75
|
v = v.to_s
|
76
|
-
raise ArgumentError, "invalid value of controlfield #{tag}: #{
|
76
|
+
raise ArgumentError, "invalid value of controlfield #{tag}: #{array}" if v.length > 1
|
77
77
|
v
|
78
78
|
else
|
79
79
|
nil
|
@@ -81,6 +81,10 @@ module Metacrunch
|
|
81
81
|
end
|
82
82
|
end
|
83
83
|
|
84
|
+
def values2string
|
85
|
+
values.map{|c| c.blank? ? "|" : c}.join
|
86
|
+
end
|
87
|
+
|
84
88
|
end
|
85
89
|
end
|
86
90
|
end
|
@@ -10,42 +10,45 @@ module Metacrunch
|
|
10
10
|
@tag = tag
|
11
11
|
@ind1 = ind1
|
12
12
|
@ind2 = ind2
|
13
|
+
@subfields = {}
|
14
|
+
end
|
15
|
+
|
16
|
+
def value
|
17
|
+
subfields.value
|
13
18
|
end
|
14
19
|
|
15
20
|
# ------------------------------------------------------------------------------
|
16
21
|
# Sub fields
|
17
22
|
# ------------------------------------------------------------------------------
|
18
23
|
|
19
|
-
#
|
20
|
-
# @return [Hash{String => Metacrunch::Mab2::Document::Datafield::Subfield::Set}]
|
21
|
-
# @private
|
22
|
-
#
|
23
|
-
def subfields_struct
|
24
|
-
@subfields_struct ||= {}
|
25
|
-
end
|
26
|
-
private :subfields_struct
|
27
|
-
|
28
24
|
#
|
29
25
|
# Returns the sub field matching the given code.
|
30
26
|
#
|
31
27
|
# @param [String] code of the sub field
|
32
|
-
# @return [Metacrunch::Mab2::Document::
|
28
|
+
# @return [Metacrunch::Mab2::Document::SubfieldSet] sub field with the given code. The set
|
33
29
|
# is empty if the sub field doesn't exists.
|
34
30
|
#
|
35
|
-
def subfields(code)
|
36
|
-
|
31
|
+
def subfields(code = nil)
|
32
|
+
result = Metacrunch::Mab2::Document::SubfieldSet.new
|
33
|
+
|
34
|
+
if code.nil?
|
35
|
+
result.concat(@subfields.values.flatten(1))
|
36
|
+
elsif _subfields = @subfields[code]
|
37
|
+
result.concat(_subfields)
|
38
|
+
elsif (codes = code).is_a?(Array)
|
39
|
+
result.concat(codes.map { |_code| @subfields[_code] }.compact.flatten(1))
|
40
|
+
end
|
41
|
+
|
42
|
+
result
|
37
43
|
end
|
38
44
|
|
39
45
|
#
|
40
46
|
# Adds a new sub field.
|
41
47
|
#
|
42
|
-
# @param [Metacrunch::Mab2::Document::
|
48
|
+
# @param [Metacrunch::Mab2::Document::Subfield] subfield
|
43
49
|
#
|
44
50
|
def add_subfield(subfield)
|
45
|
-
|
46
|
-
subfield_set << subfield
|
47
|
-
|
48
|
-
subfields_struct[subfield.code] = subfield_set
|
51
|
+
(@subfields[subfield.code] ||= []) << subfield
|
49
52
|
end
|
50
53
|
|
51
54
|
# ------------------------------------------------------------------------------
|
@@ -54,7 +57,7 @@ module Metacrunch
|
|
54
57
|
|
55
58
|
def to_xml(builder)
|
56
59
|
builder.datafield(tag: tag, ind1: ind1, ind2: ind2) do
|
57
|
-
|
60
|
+
@subfields.values.each do |_subfield_set|
|
58
61
|
_subfield_set.to_xml(builder)
|
59
62
|
end
|
60
63
|
end
|
@@ -1,56 +1,66 @@
|
|
1
|
+
require "self_enumerable"
|
2
|
+
|
1
3
|
module Metacrunch
|
2
4
|
module Mab2
|
3
5
|
class Document
|
4
|
-
class
|
5
|
-
|
6
|
-
include Enumerable
|
6
|
+
class DatafieldSet
|
7
|
+
include SelfEnumerable
|
7
8
|
|
8
|
-
|
9
|
-
|
10
|
-
|
9
|
+
def initialize(datafields = [])
|
10
|
+
@datafields = datafields
|
11
|
+
end
|
11
12
|
|
12
|
-
|
13
|
-
|
14
|
-
|
13
|
+
def each
|
14
|
+
block_given? ? @datafields.each { |_datafield| yield _datafield } : to_enum
|
15
|
+
end
|
15
16
|
|
16
|
-
|
17
|
-
|
18
|
-
|
17
|
+
def <<(datafield)
|
18
|
+
@datafields << datafield
|
19
|
+
end
|
19
20
|
|
20
|
-
|
21
|
-
|
22
|
-
|
21
|
+
def concat(datafield_set)
|
22
|
+
@datafields.concat(datafield_set.to_a)
|
23
|
+
self
|
24
|
+
end
|
23
25
|
|
24
|
-
|
25
|
-
|
26
|
-
|
26
|
+
def value
|
27
|
+
@datafields.find { |_datafield| _datafield.value }.try(:value)
|
28
|
+
end
|
29
|
+
alias_method :first_value, :value
|
27
30
|
|
28
|
-
|
29
|
-
|
30
|
-
|
31
|
+
def to_a
|
32
|
+
@datafields
|
33
|
+
end
|
34
|
+
|
35
|
+
def empty?
|
36
|
+
@datafields.empty?
|
37
|
+
end
|
31
38
|
|
32
|
-
|
33
|
-
|
34
|
-
|
39
|
+
def present?
|
40
|
+
!empty?
|
41
|
+
end
|
35
42
|
|
36
|
-
|
37
|
-
|
38
|
-
|
43
|
+
# @return [Metacrunch::Mab2::Document::SubfieldSet]
|
44
|
+
def subfields(code = nil)
|
45
|
+
result = Metacrunch::Mab2::Document::SubfieldSet.new
|
39
46
|
|
40
|
-
|
47
|
+
@datafields.each do |_datafield|
|
48
|
+
result.concat(_datafield.subfields(code))
|
41
49
|
end
|
42
50
|
|
43
|
-
|
44
|
-
|
45
|
-
# ------------------------------------------------------------------------------
|
51
|
+
result
|
52
|
+
end
|
46
53
|
|
47
|
-
|
48
|
-
|
49
|
-
|
50
|
-
end
|
51
|
-
end
|
54
|
+
# ------------------------------------------------------------------------------
|
55
|
+
# Serialization
|
56
|
+
# ------------------------------------------------------------------------------
|
52
57
|
|
58
|
+
def to_xml(builder)
|
59
|
+
self.each do |_datafield|
|
60
|
+
_datafield.to_xml(builder)
|
61
|
+
end
|
53
62
|
end
|
63
|
+
|
54
64
|
end
|
55
65
|
end
|
56
66
|
end
|
@@ -1,28 +1,25 @@
|
|
1
1
|
module Metacrunch
|
2
2
|
module Mab2
|
3
3
|
class Document
|
4
|
-
class
|
5
|
-
class Subfield
|
4
|
+
class Subfield
|
6
5
|
|
7
|
-
|
8
|
-
|
6
|
+
attr_accessor :code
|
7
|
+
attr_accessor :value
|
9
8
|
|
10
|
-
|
11
|
-
|
12
|
-
|
13
|
-
|
14
|
-
|
15
|
-
# ------------------------------------------------------------------------------
|
16
|
-
# Serialization
|
17
|
-
# ------------------------------------------------------------------------------
|
9
|
+
def initialize(code = nil, value = nil)
|
10
|
+
@code = code
|
11
|
+
@value = value
|
12
|
+
end
|
18
13
|
|
19
|
-
|
20
|
-
|
21
|
-
|
14
|
+
# ------------------------------------------------------------------------------
|
15
|
+
# Serialization
|
16
|
+
# ------------------------------------------------------------------------------
|
22
17
|
|
18
|
+
def to_xml(builder)
|
19
|
+
builder.subfield(value, code: code)
|
23
20
|
end
|
21
|
+
|
24
22
|
end
|
25
23
|
end
|
26
24
|
end
|
27
25
|
end
|
28
|
-
|
@@ -1,55 +1,58 @@
|
|
1
|
+
require "self_enumerable"
|
2
|
+
|
1
3
|
module Metacrunch
|
2
4
|
module Mab2
|
3
5
|
class Document
|
4
|
-
class
|
5
|
-
|
6
|
-
class Set
|
7
|
-
include Enumerable
|
6
|
+
class SubfieldSet
|
7
|
+
include SelfEnumerable
|
8
8
|
|
9
|
-
|
10
|
-
|
11
|
-
|
9
|
+
def initialize(subfields = [])
|
10
|
+
@subfields = subfields
|
11
|
+
end
|
12
12
|
|
13
|
-
|
14
|
-
|
15
|
-
|
13
|
+
def each
|
14
|
+
block_given? ? @subfields.each { |_subfield| yield _subfield } : to_enum
|
15
|
+
end
|
16
16
|
|
17
|
-
|
18
|
-
|
19
|
-
|
17
|
+
def <<(subfield)
|
18
|
+
@subfields << subfield
|
19
|
+
end
|
20
20
|
|
21
|
-
|
22
|
-
|
23
|
-
|
21
|
+
def concat(subfield_set)
|
22
|
+
@subfields.concat(subfield_set.to_a)
|
23
|
+
end
|
24
24
|
|
25
|
-
|
26
|
-
|
27
|
-
|
25
|
+
def to_a
|
26
|
+
@subfields
|
27
|
+
end
|
28
28
|
|
29
|
-
|
30
|
-
|
31
|
-
|
29
|
+
def empty?
|
30
|
+
@subfields.empty?
|
31
|
+
end
|
32
32
|
|
33
|
-
|
34
|
-
|
35
|
-
|
33
|
+
def present?
|
34
|
+
!empty?
|
35
|
+
end
|
36
36
|
|
37
|
-
|
38
|
-
|
39
|
-
|
37
|
+
def value
|
38
|
+
values.find{ |v| v.present? }
|
39
|
+
end
|
40
|
+
alias_method :first_value, :value
|
40
41
|
|
41
|
-
|
42
|
-
|
43
|
-
|
42
|
+
def values
|
43
|
+
@subfields.map{ |subfield| subfield.value }
|
44
|
+
end
|
44
45
|
|
45
|
-
|
46
|
-
|
47
|
-
|
48
|
-
end
|
49
|
-
end
|
46
|
+
# ------------------------------------------------------------------------------
|
47
|
+
# Serialization
|
48
|
+
# ------------------------------------------------------------------------------
|
50
49
|
|
50
|
+
def to_xml(builder)
|
51
|
+
self.each do |_subfield|
|
52
|
+
_subfield.to_xml(builder)
|
51
53
|
end
|
52
54
|
end
|
55
|
+
|
53
56
|
end
|
54
57
|
end
|
55
58
|
end
|
@@ -0,0 +1,48 @@
|
|
1
|
+
require "nokogiri"
|
2
|
+
require_relative "../mab2"
|
3
|
+
|
4
|
+
class Metacrunch::Mab2::XmlBuilder
|
5
|
+
def initialize(identifier = nil, &block)
|
6
|
+
@builder =
|
7
|
+
Nokogiri::XML::Builder.new do
|
8
|
+
send(
|
9
|
+
:"OAI-PMH",
|
10
|
+
"xmlns" => "http://www.openarchives.org/OAI/2.0/",
|
11
|
+
"xmlns:xsi" => "http://www.w3.org/2001/XMLSchema-instance",
|
12
|
+
"xsi:schemaLocation" => "http://www.openarchives.org/OAI/2.0/ http://www.openarchives.org/OAI/2.0/OAI-PMH.xsd"
|
13
|
+
) do
|
14
|
+
ListRecords do
|
15
|
+
record do
|
16
|
+
if identifier
|
17
|
+
header do |_xml|
|
18
|
+
_xml.identifier identifier
|
19
|
+
end
|
20
|
+
end
|
21
|
+
|
22
|
+
metadata do
|
23
|
+
record("xmlns" => "http://www.ddb.de/professionell/mabxml/mabxml-1.xsd") do
|
24
|
+
define_singleton_method(:controlfield) do |_tag, _text|
|
25
|
+
send(:method_missing, :controlfield, _text, tag: _tag)
|
26
|
+
end
|
27
|
+
|
28
|
+
define_singleton_method(:datafield) do |_tag, _attributes = {}, &_block|
|
29
|
+
send(:method_missing, :datafield, {tag: _tag}.merge(_attributes), &_block)
|
30
|
+
end
|
31
|
+
|
32
|
+
define_singleton_method(:subfield) do |_code, _text|
|
33
|
+
send(:method_missing, :subfield, _text, code: _code)
|
34
|
+
end
|
35
|
+
|
36
|
+
instance_eval(&block)
|
37
|
+
end
|
38
|
+
end
|
39
|
+
end
|
40
|
+
end
|
41
|
+
end
|
42
|
+
end
|
43
|
+
end
|
44
|
+
|
45
|
+
def to_xml
|
46
|
+
@builder.to_xml
|
47
|
+
end
|
48
|
+
end
|
data/metacrunch-mab2.gemspec
CHANGED
@@ -1,6 +1,6 @@
|
|
1
1
|
describe Metacrunch::Mab2::Document::Controlfield do
|
2
2
|
|
3
|
-
describe "
|
3
|
+
describe "#initialize" do
|
4
4
|
it "creates a new control field when using String values" do
|
5
5
|
field = Metacrunch::Mab2::Document::Controlfield.new("050", "a||b")
|
6
6
|
expect(field.tag).to eq("050")
|
@@ -14,7 +14,7 @@ describe Metacrunch::Mab2::Document::Controlfield do
|
|
14
14
|
end
|
15
15
|
end
|
16
16
|
|
17
|
-
describe "
|
17
|
+
describe "#values=" do
|
18
18
|
let(:field) { Metacrunch::Mab2::Document::Controlfield.new("050", nil) }
|
19
19
|
|
20
20
|
it "sets field values using String" do
|
@@ -1,16 +1,33 @@
|
|
1
1
|
describe Metacrunch::Mab2::Document do
|
2
2
|
|
3
|
-
describe "
|
3
|
+
describe ".from_aleph_mab_xml" do
|
4
4
|
let(:xml) { default_test_xml }
|
5
5
|
let(:document) { Metacrunch::Mab2::Document.from_aleph_mab_xml(xml) }
|
6
|
+
|
6
7
|
subject { document }
|
7
8
|
|
8
|
-
it "should
|
9
|
+
it "should return a Mab2::Document" do
|
9
10
|
expect(subject).to be_instance_of(Metacrunch::Mab2::Document)
|
10
11
|
end
|
12
|
+
|
13
|
+
it "should find 2 data fields with tag 070" do
|
14
|
+
expect(subject.datafields("070").count).to be(2)
|
15
|
+
end
|
16
|
+
|
17
|
+
it "should find 1 data field with tag 100" do
|
18
|
+
expect(subject.datafields("100").first).not_to be_nil
|
19
|
+
end
|
20
|
+
|
21
|
+
it "sub field p of data field 100 is not nil" do
|
22
|
+
expect(subject.datafields("100").first.subfields("p").first.value).not_to be_nil
|
23
|
+
end
|
24
|
+
|
25
|
+
it "decodes html entities" do
|
26
|
+
expect(subject.datafields("331").subfields("a").values.first).to eq("<<Das>> Linux für Studenten")
|
27
|
+
end
|
11
28
|
end
|
12
29
|
|
13
|
-
describe "
|
30
|
+
describe "#add_controlfield" do
|
14
31
|
let(:document) { empty_document }
|
15
32
|
|
16
33
|
it "should add control fields" do
|
@@ -19,7 +36,7 @@ describe Metacrunch::Mab2::Document do
|
|
19
36
|
end
|
20
37
|
end
|
21
38
|
|
22
|
-
describe "
|
39
|
+
describe "#controlfield" do
|
23
40
|
let(:document) { default_test_document }
|
24
41
|
subject { document.controlfield("050") }
|
25
42
|
|
@@ -30,7 +47,7 @@ describe Metacrunch::Mab2::Document do
|
|
30
47
|
end
|
31
48
|
end
|
32
49
|
|
33
|
-
describe "
|
50
|
+
describe "#datafields" do
|
34
51
|
let(:document) { default_test_document }
|
35
52
|
|
36
53
|
context "given tag=100" do
|
@@ -72,6 +89,14 @@ describe Metacrunch::Mab2::Document do
|
|
72
89
|
expect(subject.count).to eq(2)
|
73
90
|
end
|
74
91
|
end
|
92
|
+
|
93
|
+
context "given tag=100 and ind1=[:blank, 'a']" do
|
94
|
+
subject { document.datafields("100", ind1: [:blank, "a"]) }
|
95
|
+
|
96
|
+
it "returns only datafield with tag=100 and ind1=(:blank || 'a')" do
|
97
|
+
expect(subject.count).to eq(3)
|
98
|
+
end
|
99
|
+
end
|
75
100
|
end
|
76
101
|
|
77
102
|
private
|
@@ -121,7 +146,7 @@ private
|
|
121
146
|
end
|
122
147
|
|
123
148
|
def create_subfield(code, value)
|
124
|
-
Metacrunch::Mab2::Document::
|
149
|
+
Metacrunch::Mab2::Document::Subfield.new(code, value)
|
125
150
|
end
|
126
151
|
|
127
152
|
end
|
metadata
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: metacrunch-mab2
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 1.0
|
4
|
+
version: 1.1.0
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- René Sprotte
|
@@ -10,7 +10,7 @@ authors:
|
|
10
10
|
autorequire:
|
11
11
|
bindir: bin
|
12
12
|
cert_chain: []
|
13
|
-
date: 2015-
|
13
|
+
date: 2015-09-25 00:00:00.000000000 Z
|
14
14
|
dependencies:
|
15
15
|
- !ruby/object:Gem::Dependency
|
16
16
|
name: metacrunch
|
@@ -40,6 +40,20 @@ dependencies:
|
|
40
40
|
- - "~>"
|
41
41
|
- !ruby/object:Gem::Version
|
42
42
|
version: '4.3'
|
43
|
+
- !ruby/object:Gem::Dependency
|
44
|
+
name: nokogiri
|
45
|
+
requirement: !ruby/object:Gem::Requirement
|
46
|
+
requirements:
|
47
|
+
- - "~>"
|
48
|
+
- !ruby/object:Gem::Version
|
49
|
+
version: '1.6'
|
50
|
+
type: :runtime
|
51
|
+
prerelease: false
|
52
|
+
version_requirements: !ruby/object:Gem::Requirement
|
53
|
+
requirements:
|
54
|
+
- - "~>"
|
55
|
+
- !ruby/object:Gem::Version
|
56
|
+
version: '1.6'
|
43
57
|
- !ruby/object:Gem::Dependency
|
44
58
|
name: ox
|
45
59
|
requirement: !ruby/object:Gem::Requirement
|
@@ -54,9 +68,25 @@ dependencies:
|
|
54
68
|
- - "~>"
|
55
69
|
- !ruby/object:Gem::Version
|
56
70
|
version: '2.1'
|
71
|
+
- !ruby/object:Gem::Dependency
|
72
|
+
name: self_enumerable
|
73
|
+
requirement: !ruby/object:Gem::Requirement
|
74
|
+
requirements:
|
75
|
+
- - ">="
|
76
|
+
- !ruby/object:Gem::Version
|
77
|
+
version: '0'
|
78
|
+
type: :runtime
|
79
|
+
prerelease: false
|
80
|
+
version_requirements: !ruby/object:Gem::Requirement
|
81
|
+
requirements:
|
82
|
+
- - ">="
|
83
|
+
- !ruby/object:Gem::Version
|
84
|
+
version: '0'
|
57
85
|
description: MAB2 tools for metacrunch
|
58
86
|
email: r.sprotte@ub.uni-paderborn.de
|
59
|
-
executables:
|
87
|
+
executables:
|
88
|
+
- console
|
89
|
+
- setup
|
60
90
|
extensions: []
|
61
91
|
extra_rdoc_files: []
|
62
92
|
files:
|
@@ -69,22 +99,25 @@ files:
|
|
69
99
|
- Readme.md
|
70
100
|
- benchmarks/mab2_document_parsing.rb
|
71
101
|
- benchmarks/mab2_document_query.rb
|
102
|
+
- bin/console
|
103
|
+
- bin/setup
|
72
104
|
- lib/metacrunch/mab2.rb
|
73
|
-
- lib/metacrunch/mab2/
|
105
|
+
- lib/metacrunch/mab2/builder.rb
|
74
106
|
- lib/metacrunch/mab2/cli.rb
|
75
107
|
- lib/metacrunch/mab2/document.rb
|
108
|
+
- lib/metacrunch/mab2/document/aleph_mab_xml_parser.rb
|
76
109
|
- lib/metacrunch/mab2/document/controlfield.rb
|
77
110
|
- lib/metacrunch/mab2/document/datafield.rb
|
78
111
|
- lib/metacrunch/mab2/document/datafield_set.rb
|
79
112
|
- lib/metacrunch/mab2/document/subfield.rb
|
80
113
|
- lib/metacrunch/mab2/document/subfield_set.rb
|
81
114
|
- lib/metacrunch/mab2/version.rb
|
115
|
+
- lib/metacrunch/mab2/xml_builder.rb
|
82
116
|
- lib/metacrunch_plugin.rb
|
83
117
|
- metacrunch-mab2.gemspec
|
84
118
|
- spec/assets/aleph_mab_xml/file1.xml
|
85
|
-
- spec/document/
|
86
|
-
- spec/
|
87
|
-
- spec/document/document_spec.rb
|
119
|
+
- spec/metacrunch/mab2/document/control_field_spec.rb
|
120
|
+
- spec/metacrunch/mab2/document_spec.rb
|
88
121
|
- spec/spec_helper.rb
|
89
122
|
homepage: http://github.com/ubpb/metacrunch-mab2
|
90
123
|
licenses:
|
@@ -106,13 +139,13 @@ required_rubygems_version: !ruby/object:Gem::Requirement
|
|
106
139
|
version: '0'
|
107
140
|
requirements: []
|
108
141
|
rubyforge_project:
|
109
|
-
rubygems_version: 2.4.
|
142
|
+
rubygems_version: 2.4.8
|
110
143
|
signing_key:
|
111
144
|
specification_version: 4
|
112
145
|
summary: MAB2 tools for metacrunch
|
113
146
|
test_files:
|
114
147
|
- spec/assets/aleph_mab_xml/file1.xml
|
115
|
-
- spec/document/
|
116
|
-
- spec/
|
117
|
-
- spec/document/document_spec.rb
|
148
|
+
- spec/metacrunch/mab2/document/control_field_spec.rb
|
149
|
+
- spec/metacrunch/mab2/document_spec.rb
|
118
150
|
- spec/spec_helper.rb
|
151
|
+
has_rdoc:
|
@@ -1,37 +0,0 @@
|
|
1
|
-
describe Metacrunch::Mab2::AlephMabXmlDocumentFactory do
|
2
|
-
|
3
|
-
context "with default test xml file" do
|
4
|
-
let(:factory) { Metacrunch::Mab2::AlephMabXmlDocumentFactory.new(default_test_xml) }
|
5
|
-
|
6
|
-
describe ".to_document" do
|
7
|
-
subject { factory.to_document }
|
8
|
-
|
9
|
-
it "should return a Mab2::Document" do
|
10
|
-
expect(subject).to be_instance_of(Metacrunch::Mab2::Document)
|
11
|
-
end
|
12
|
-
|
13
|
-
it "should find 2 data fields with tag 070" do
|
14
|
-
expect(subject.datafields("070").count).to be(2)
|
15
|
-
end
|
16
|
-
|
17
|
-
it "should find 1 data field with tag 100" do
|
18
|
-
expect(subject.datafields("100").first).not_to be_nil
|
19
|
-
end
|
20
|
-
|
21
|
-
it "sub field p of data field 100 is not nil" do
|
22
|
-
expect(subject.datafields("100").first.subfields("p").first.value).not_to be_nil
|
23
|
-
end
|
24
|
-
|
25
|
-
it "decodes html entities" do
|
26
|
-
expect(subject.datafields("331").subfields("a").values.first).to eq("<<Das>> Linux für Studenten")
|
27
|
-
end
|
28
|
-
end
|
29
|
-
end
|
30
|
-
|
31
|
-
private
|
32
|
-
|
33
|
-
def default_test_xml
|
34
|
-
File.read(File.join(RSpec.root, "assets", "aleph_mab_xml", "file1.xml"))
|
35
|
-
end
|
36
|
-
|
37
|
-
end
|