relaton-nist 0.3.3 → 0.3.4
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/Gemfile.lock +7 -7
- data/README.adoc +1 -1
- data/lib/relaton_nist.rb +4 -20
- data/lib/relaton_nist/hit.rb +1 -30
- data/lib/relaton_nist/hit_collection.rb +54 -58
- data/lib/relaton_nist/nist_bibliography.rb +1 -5
- data/lib/relaton_nist/processor.rb +33 -0
- data/lib/relaton_nist/scrapper.rb +4 -6
- data/lib/relaton_nist/version.rb +1 -1
- metadata +3 -3
- data/lib/relaton/processor.rb +0 -23
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA1:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 6ad28b9e4eda9f14ac735a69842d2e4d63c7c88e
|
4
|
+
data.tar.gz: e20f4b0086adbe773a420d22d0ee30c352001149
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 8ac638788da72901abd8373f68a1c6228584b4754823251e597b20b4898a3f14de21ad870b5dc7e27deccb09f2c18e0af7dee8e8952f5395ea5f001ca824094e
|
7
|
+
data.tar.gz: 46f960c6178eebaeb1ed9a36bf9989746e59b3266691e07aa4d5ce457f2c7e6d871905d8e51ba025d3409457126bfc8c1dcbdd4924935843243b143fd012528f
|
data/Gemfile.lock
CHANGED
@@ -1,15 +1,15 @@
|
|
1
1
|
PATH
|
2
2
|
remote: .
|
3
3
|
specs:
|
4
|
-
relaton-nist (0.3.
|
4
|
+
relaton-nist (0.3.4)
|
5
5
|
relaton-bib (~> 0.3.0)
|
6
6
|
rubyzip
|
7
7
|
|
8
8
|
GEM
|
9
9
|
remote: https://rubygems.org/
|
10
10
|
specs:
|
11
|
-
addressable (2.
|
12
|
-
public_suffix (>= 2.0.2, <
|
11
|
+
addressable (2.7.0)
|
12
|
+
public_suffix (>= 2.0.2, < 5.0)
|
13
13
|
byebug (11.0.1)
|
14
14
|
coderay (1.1.2)
|
15
15
|
crack (0.4.3)
|
@@ -33,11 +33,11 @@ GEM
|
|
33
33
|
pry-byebug (3.7.0)
|
34
34
|
byebug (~> 11.0)
|
35
35
|
pry (~> 0.10)
|
36
|
-
public_suffix (
|
36
|
+
public_suffix (4.0.1)
|
37
37
|
rake (10.5.0)
|
38
|
-
relaton-bib (0.3.
|
38
|
+
relaton-bib (0.3.6)
|
39
39
|
addressable
|
40
|
-
nokogiri
|
40
|
+
nokogiri
|
41
41
|
rspec (3.8.0)
|
42
42
|
rspec-core (~> 3.8.0)
|
43
43
|
rspec-expectations (~> 3.8.0)
|
@@ -53,7 +53,7 @@ GEM
|
|
53
53
|
rspec-support (3.8.2)
|
54
54
|
ruby-debug-ide (0.7.0)
|
55
55
|
rake (>= 0.8.1)
|
56
|
-
rubyzip (1.2.
|
56
|
+
rubyzip (1.2.4)
|
57
57
|
safe_yaml (1.0.5)
|
58
58
|
simplecov (0.16.1)
|
59
59
|
docile (~> 1.1)
|
data/README.adoc
CHANGED
data/lib/relaton_nist.rb
CHANGED
@@ -1,27 +1,11 @@
|
|
1
1
|
require "relaton_nist/version"
|
2
2
|
require "relaton_nist/nist_bibliography"
|
3
3
|
|
4
|
-
if defined? Relaton
|
5
|
-
|
6
|
-
|
7
|
-
end
|
4
|
+
# if defined? Relaton
|
5
|
+
# require_relative "relaton/processor"
|
6
|
+
# Relaton::Registry.instance.register(Relaton::RelatonNist::Processor)
|
7
|
+
# end
|
8
8
|
|
9
9
|
module RelatonNist
|
10
10
|
class Error < StandardError; end
|
11
|
-
|
12
|
-
class << self
|
13
|
-
# @param date [String]
|
14
|
-
# @return [Date, NilClass]
|
15
|
-
def parse_date(sdate)
|
16
|
-
if /(?<date>\w+\s\d{4})/ =~ sdate # February 2012
|
17
|
-
Date.strptime(date, "%B %Y")
|
18
|
-
elsif /(?<date>\w+\s\d{1,2},\s\d{4})/ =~ sdate # February 11, 2012
|
19
|
-
Date.strptime(date, "%B %d, %Y")
|
20
|
-
elsif /(?<date>\d{4}-\d{2}-\d{2})/ =~ sdate # 2012-02-11
|
21
|
-
Date.parse(date)
|
22
|
-
elsif /(?<date>\d{4}-\d{2})/ =~ sdate # 2012-02
|
23
|
-
Date.strptime date, "%Y-%m"
|
24
|
-
end
|
25
|
-
end
|
26
|
-
end
|
27
11
|
end
|
data/lib/relaton_nist/hit.rb
CHANGED
@@ -2,45 +2,16 @@
|
|
2
2
|
|
3
3
|
module RelatonNist
|
4
4
|
# Hit.
|
5
|
-
class Hit
|
5
|
+
class Hit < RelatonBib::Hit
|
6
6
|
# @return [RelatonNist::HitCollection]
|
7
7
|
attr_reader :hit_collection
|
8
8
|
|
9
|
-
# @return [Array<Hash>]
|
10
|
-
attr_reader :hit
|
11
|
-
|
12
|
-
# @param hit [Hash]
|
13
|
-
# @param hit_collection [RelatonNist:HitCollection]
|
14
|
-
def initialize(hit, hit_collection = nil)
|
15
|
-
@hit = hit
|
16
|
-
@hit_collection = hit_collection
|
17
|
-
end
|
18
|
-
|
19
9
|
# Parse page.
|
20
10
|
# @return [RelatonNist::NistBliographicItem]
|
21
11
|
def fetch
|
22
12
|
@fetch ||= Scrapper.parse_page @hit
|
23
13
|
end
|
24
14
|
|
25
|
-
# @return [String]
|
26
|
-
def to_s
|
27
|
-
inspect
|
28
|
-
end
|
29
|
-
|
30
|
-
# @return [String]
|
31
|
-
def inspect
|
32
|
-
"<#{self.class}:#{format('%#.14x', object_id << 1)} "\
|
33
|
-
"@text=\"#{@hit_collection&.text}\" "\
|
34
|
-
"@fetched=\"#{!@fetch.nil?}\" "\
|
35
|
-
"@fullIdentifier=\"#{@fetch&.shortref(nil)}\" "\
|
36
|
-
"@title=\"#{@hit[:code]}\">"
|
37
|
-
end
|
38
|
-
|
39
|
-
# @return [String]
|
40
|
-
def to_xml(**opts)
|
41
|
-
fetch.to_xml **opts
|
42
|
-
end
|
43
|
-
|
44
15
|
# @return [Iteger]
|
45
16
|
def sort_value
|
46
17
|
@sort_value ||= begin
|
@@ -8,19 +8,10 @@ require "open-uri"
|
|
8
8
|
|
9
9
|
module RelatonNist
|
10
10
|
# Page of hit collection.
|
11
|
-
class HitCollection <
|
11
|
+
class HitCollection < RelatonBib::HitCollection
|
12
12
|
DOMAIN = "https://csrc.nist.gov"
|
13
13
|
DATAFILE = File.expand_path "data/pubs-export.zip", __dir__
|
14
14
|
|
15
|
-
# @return [TrueClass, FalseClass]
|
16
|
-
attr_reader :fetched
|
17
|
-
|
18
|
-
# @return [String]
|
19
|
-
attr_reader :text
|
20
|
-
|
21
|
-
# @return [String]
|
22
|
-
attr_reader :year
|
23
|
-
|
24
15
|
# rubocop:disable Metrics/AbcSize, Metrics/MethodLength
|
25
16
|
|
26
17
|
# @param ref_nbr [String]
|
@@ -46,28 +37,6 @@ module RelatonNist
|
|
46
37
|
end
|
47
38
|
# rubocop:enable Metrics/AbcSize, Metrics/MethodLength
|
48
39
|
|
49
|
-
# @return [Iecbib::HitCollection]
|
50
|
-
def fetch
|
51
|
-
workers = RelatonBib::WorkersPool.new 4
|
52
|
-
workers.worker(&:fetch)
|
53
|
-
each do |hit|
|
54
|
-
workers << hit
|
55
|
-
end
|
56
|
-
workers.end
|
57
|
-
workers.result
|
58
|
-
@fetched = true
|
59
|
-
self
|
60
|
-
end
|
61
|
-
|
62
|
-
def to_s
|
63
|
-
inspect
|
64
|
-
end
|
65
|
-
|
66
|
-
# @return [String]
|
67
|
-
def inspect
|
68
|
-
"<#{self.class}:#{format('%#.14x', object_id << 1)} @fetched=#{@fetched}>"
|
69
|
-
end
|
70
|
-
|
71
40
|
private
|
72
41
|
|
73
42
|
# rubocop:disable Metrics/AbcSize, Metrics/MethodLength
|
@@ -107,50 +76,78 @@ module RelatonNist
|
|
107
76
|
)
|
108
77
|
end
|
109
78
|
end
|
79
|
+
# rubocop:enable Metrics/AbcSize, Metrics/MethodLength
|
110
80
|
|
111
81
|
# Fetches data form json
|
112
82
|
# @param docid [String]
|
83
|
+
# @param stage [String]
|
84
|
+
# @return [Array<RelatonNist::Hit>]
|
113
85
|
def from_json(docid, **opts)
|
86
|
+
select_data(docid, **opts).map do |h|
|
87
|
+
/(?<serie>(?<=-)\w+$)/ =~ h["series"]
|
88
|
+
title = [h["title-main"], h["title-sub"]].compact.join " - "
|
89
|
+
release_date = RelatonBib.parse_date h["published-date"]
|
90
|
+
Hit.new({ code: h["docidentifier"], serie: serie.upcase, title: title,
|
91
|
+
url: h["uri"], status: h["status"],
|
92
|
+
release_date: release_date, json: h }, self)
|
93
|
+
end
|
94
|
+
end
|
95
|
+
|
96
|
+
# @param docid [String]
|
97
|
+
# @param stage [String]
|
98
|
+
# @return [Array<Hach>]
|
99
|
+
def select_data(docid, **opts)
|
100
|
+
d = Date.strptime year, "%Y" if year
|
114
101
|
data.select do |doc|
|
115
|
-
|
116
|
-
|
117
|
-
idate = RelatonNist.parse_date doc["issued-date"]
|
118
|
-
next unless idate.between? d, d.next_year.prev_day
|
119
|
-
end
|
102
|
+
next unless match_year?(doc, d)
|
103
|
+
|
120
104
|
if /PD/ =~ opts[:stage]
|
121
105
|
next unless %w[draft-public draft-prelim].include? doc["status"]
|
122
106
|
else
|
123
107
|
next unless doc["status"] == "final"
|
124
108
|
end
|
125
109
|
doc["docidentifier"] =~ Regexp.new(docid)
|
126
|
-
end.map do |h|
|
127
|
-
/(?<serie>(?<=-)\w+$)/ =~ h["series"]
|
128
|
-
title = [h["title-main"], h["title-sub"]].compact.join " - "
|
129
|
-
release_date = RelatonNist.parse_date h["published-date"]
|
130
|
-
Hit.new(
|
131
|
-
{
|
132
|
-
code: h["docidentifier"], serie: serie.upcase, title: title,
|
133
|
-
url: h["uri"], status: h["status"], release_date: release_date,
|
134
|
-
json: h
|
135
|
-
}, self
|
136
|
-
)
|
137
110
|
end
|
138
111
|
end
|
139
112
|
|
140
|
-
#
|
113
|
+
# @param doc [Hash]
|
114
|
+
# @param date [Date] first day of year
|
115
|
+
# @return [TrueClass, FalseClass]
|
116
|
+
def match_year?(doc, date)
|
117
|
+
return true unless year
|
118
|
+
|
119
|
+
idate = RelatonBib.parse_date doc["issued-date"]
|
120
|
+
idate.between? date, date.next_year.prev_day
|
121
|
+
end
|
122
|
+
|
123
|
+
# Fetches json data form server
|
141
124
|
# @return [Hash]
|
142
125
|
def data
|
143
126
|
ctime = File.ctime DATAFILE if File.exist? DATAFILE
|
144
127
|
if !ctime || ctime.to_date < Date.today
|
145
|
-
|
146
|
-
|
147
|
-
|
148
|
-
|
149
|
-
|
150
|
-
|
151
|
-
|
128
|
+
fetch_data(ctime)
|
129
|
+
end
|
130
|
+
unzip
|
131
|
+
end
|
132
|
+
|
133
|
+
# Fetch data form server and save it to file
|
134
|
+
#
|
135
|
+
# @prarm ctime [Time, NilClass]
|
136
|
+
def fetch_data(ctime)
|
137
|
+
resp = OpenURI.open_uri("https://csrc.nist.gov/CSRC/media/feeds/metanorma/pubs-export.meta")
|
138
|
+
if !ctime || ctime < resp.last_modified
|
139
|
+
@data = nil
|
140
|
+
zip = OpenURI.open_uri "https://csrc.nist.gov/CSRC/media/feeds/metanorma/pubs-export.zip"
|
141
|
+
zip.close
|
142
|
+
FileUtils.mv zip.path, DATAFILE
|
152
143
|
end
|
153
|
-
|
144
|
+
end
|
145
|
+
|
146
|
+
# upack zip file
|
147
|
+
#
|
148
|
+
# @return [Hash]
|
149
|
+
def unzip
|
150
|
+
return @data if @data
|
154
151
|
|
155
152
|
Zip::File.open(DATAFILE) do |zf|
|
156
153
|
zf.each do |f|
|
@@ -160,6 +157,5 @@ module RelatonNist
|
|
160
157
|
end
|
161
158
|
@data
|
162
159
|
end
|
163
|
-
# rubocop:enable Metrics/AbcSize, Metrics/MethodLength
|
164
160
|
end
|
165
161
|
end
|
@@ -51,11 +51,7 @@ module RelatonNist
|
|
51
51
|
end
|
52
52
|
|
53
53
|
code += "-1" if opts[:all_parts]
|
54
|
-
|
55
|
-
# return nil if ret.nil?
|
56
|
-
# ret.to_most_recent_reference unless year || opts[:keep_year]
|
57
|
-
# ret.to_all_parts if opts[:all_parts]
|
58
|
-
ret
|
54
|
+
nistbib_get1(code, year, opts)
|
59
55
|
end
|
60
56
|
|
61
57
|
private
|
@@ -0,0 +1,33 @@
|
|
1
|
+
require "relaton/processor"
|
2
|
+
|
3
|
+
module RelatonNist
|
4
|
+
class Processor < Relaton::Processor
|
5
|
+
def initialize
|
6
|
+
@short = :relaton_nist
|
7
|
+
@prefix = "NIST"
|
8
|
+
@defaultprefix = %r{^(NIST|NISTGCR|ITL Bulletin|JPCRD|NISTIR|CSRC)[ /]}
|
9
|
+
@idtype = "NIST"
|
10
|
+
end
|
11
|
+
|
12
|
+
# @param code [String]
|
13
|
+
# @param date [String, NilClass] year
|
14
|
+
# @param opts [Hash]
|
15
|
+
# @return [RelatonNist::GbBibliographicItem]
|
16
|
+
def get(code, date = nil, opts = {})
|
17
|
+
::RelatonNist::NistBibliography.get(code, date, opts)
|
18
|
+
end
|
19
|
+
|
20
|
+
# @param xml [String]
|
21
|
+
# @return [RelatonNist::GbBibliographicItem]
|
22
|
+
def from_xml(xml)
|
23
|
+
::RelatonNist::XMLParser.from_xml xml
|
24
|
+
end
|
25
|
+
|
26
|
+
# @param hash [Hash]
|
27
|
+
# @return [RelatonNist::GbBibliographicItem]
|
28
|
+
def hash_to_bib(hash)
|
29
|
+
item_hash = ::RelatonNist::HashConverter.hash_to_bib(hash)
|
30
|
+
::RelatonNist::NistBibliographicItem.new item_hash
|
31
|
+
end
|
32
|
+
end
|
33
|
+
end
|
@@ -44,11 +44,9 @@ module RelatonNist
|
|
44
44
|
edition: fetch_edition(json),
|
45
45
|
language: [json["language"]],
|
46
46
|
script: [json["script"]],
|
47
|
-
# abstract: fetch_abstract(doc),
|
48
47
|
docstatus: fetch_status(json, hit_data[:status]),
|
49
48
|
copyright: fetch_copyright(json["published-date"]),
|
50
49
|
relation: fetch_relations_json(json),
|
51
|
-
# series: fetch_series(json),
|
52
50
|
keyword: fetch_keywords(json),
|
53
51
|
commentperiod: fetch_commentperiod_json(json),
|
54
52
|
}
|
@@ -179,14 +177,14 @@ module RelatonNist
|
|
179
177
|
dates = [{ type: "published", on: release_date.to_s }]
|
180
178
|
|
181
179
|
if doc.is_a? Hash
|
182
|
-
issued =
|
183
|
-
updated =
|
180
|
+
issued = RelatonBib.parse_date doc["issued-date"]
|
181
|
+
updated = RelatonBib.parse_date doc["updated-date"]
|
184
182
|
dates << { type: "updated", on: updated.to_s } if updated
|
185
|
-
obsoleted =
|
183
|
+
obsoleted = RelatonBib.parse_date doc["obsoleted-date"]
|
186
184
|
dates << { type: "obsoleted", on: obsoleted.to_s } if obsoleted
|
187
185
|
else
|
188
186
|
d = doc.at("//span[@id='pub-release-date']").text.strip
|
189
|
-
issued =
|
187
|
+
issued = RelatonBib.parse_date d
|
190
188
|
end
|
191
189
|
dates << { type: "issued", on: issued.to_s }
|
192
190
|
dates
|
data/lib/relaton_nist/version.rb
CHANGED
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: relaton-nist
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.3.
|
4
|
+
version: 0.3.4
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Ribose Inc.
|
8
8
|
autorequire:
|
9
9
|
bindir: exe
|
10
10
|
cert_chain: []
|
11
|
-
date: 2019-
|
11
|
+
date: 2019-09-12 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
14
|
name: bundler
|
@@ -211,7 +211,6 @@ files:
|
|
211
211
|
- appveyor.yml
|
212
212
|
- bin/console
|
213
213
|
- bin/setup
|
214
|
-
- lib/relaton/processor.rb
|
215
214
|
- lib/relaton_nist.rb
|
216
215
|
- lib/relaton_nist/comment_period.rb
|
217
216
|
- lib/relaton_nist/data/pubs-export.zip
|
@@ -222,6 +221,7 @@ files:
|
|
222
221
|
- lib/relaton_nist/keyword.rb
|
223
222
|
- lib/relaton_nist/nist_bibliographic_item.rb
|
224
223
|
- lib/relaton_nist/nist_bibliography.rb
|
224
|
+
- lib/relaton_nist/processor.rb
|
225
225
|
- lib/relaton_nist/scrapper.rb
|
226
226
|
- lib/relaton_nist/version.rb
|
227
227
|
- lib/relaton_nist/xml_parser.rb
|
data/lib/relaton/processor.rb
DELETED
@@ -1,23 +0,0 @@
|
|
1
|
-
require "relaton/processor"
|
2
|
-
|
3
|
-
module Relaton
|
4
|
-
module RelatonNist
|
5
|
-
class Processor < Relaton::Processor
|
6
|
-
|
7
|
-
def initialize
|
8
|
-
@short = :relaton_nist
|
9
|
-
@prefix = "NIST"
|
10
|
-
@defaultprefix = %r{^(NIST|NISTGCR|ITL Bulletin|JPCRD|NISTIR|CSRC)[ /]}
|
11
|
-
@idtype = "NIST"
|
12
|
-
end
|
13
|
-
|
14
|
-
def get(code, date = nil, opts = {})
|
15
|
-
::RelatonNist::NistBibliography.get(code, date, opts)
|
16
|
-
end
|
17
|
-
|
18
|
-
def from_xml(xml)
|
19
|
-
::RelatonNist::XMLParser.from_xml xml
|
20
|
-
end
|
21
|
-
end
|
22
|
-
end
|
23
|
-
end
|