rdf-sak 0.1.5
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +7 -0
- data/.gitignore +16 -0
- data/.rspec +2 -0
- data/.travis.yml +4 -0
- data/Gemfile +4 -0
- data/LICENSE +202 -0
- data/README.md +268 -0
- data/Rakefile +13 -0
- data/bin/console +14 -0
- data/bin/setup +8 -0
- data/example/cleanup.xsl +14 -0
- data/example/matches.xhtml +11 -0
- data/example/transforms.ttl +58 -0
- data/lib/rdf-sak.rb +1 -0
- data/lib/rdf/sak.rb +2506 -0
- data/lib/rdf/sak/ci.rb +827 -0
- data/lib/rdf/sak/cli.rb +35 -0
- data/lib/rdf/sak/docstats.rb +188 -0
- data/lib/rdf/sak/document.rb +772 -0
- data/lib/rdf/sak/ibis.rb +248 -0
- data/lib/rdf/sak/mimemagic.rb +73 -0
- data/lib/rdf/sak/pav.rb +479 -0
- data/lib/rdf/sak/qb.rb +280 -0
- data/lib/rdf/sak/scovo.rb +51 -0
- data/lib/rdf/sak/tfo.rb +301 -0
- data/lib/rdf/sak/transform.rb +1172 -0
- data/lib/rdf/sak/urlrunner.rb +602 -0
- data/lib/rdf/sak/util.rb +2081 -0
- data/lib/rdf/sak/version.rb +5 -0
- data/rdf-sak.gemspec +60 -0
- metadata +366 -0
data/Rakefile
ADDED
@@ -0,0 +1,13 @@
|
|
1
|
+
require "bundler/gem_tasks"
|
2
|
+
require "rspec/core/rake_task"
|
3
|
+
|
4
|
+
RSpec::Core::RakeTask.new(:spec)
|
5
|
+
|
6
|
+
task :default => :spec
|
7
|
+
|
8
|
+
#desc 'Generate Vocabularies'
|
9
|
+
#task :gen_vocabs => %w(ci).map { |v| "lib/rdf/sak/#{v}.rb" }
|
10
|
+
|
11
|
+
# XXX turn this into a rake task at some point :P
|
12
|
+
|
13
|
+
# rdf serialize --uri 'https://privatealpha.com/ontology/content-inventory/1#' --output-format vocabulary --module-name RDF::SAK --class-name CI -o lib/rdf/sak/ci.rb --strict 'https://privatealpha.com/ontology/content-inventory/1#'
|
data/bin/console
ADDED
@@ -0,0 +1,14 @@
|
|
1
|
+
#!/usr/bin/env ruby
|
2
|
+
|
3
|
+
require "bundler/setup"
|
4
|
+
require "rdf/sak"
|
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
|
12
|
+
|
13
|
+
require "irb"
|
14
|
+
IRB.start
|
data/bin/setup
ADDED
data/example/cleanup.xsl
ADDED
@@ -0,0 +1,14 @@
|
|
1
|
+
<?xml version="1.0" encoding="utf-8"?>
|
2
|
+
<xsl:stylesheet version="1.0"
|
3
|
+
xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
|
4
|
+
xmlns:html="http://www.w3.org/1999/xhtml"
|
5
|
+
xmlns="http://www.w3.org/1999/xhtml"
|
6
|
+
exclude-result-prefixes="html">
|
7
|
+
|
8
|
+
<xsl:key name="main" match="html:main" use="''"/>
|
9
|
+
|
10
|
+
<xsl:template match="/html:*">
|
11
|
+
<xsl:copy-of select="key('main', '')[1]"/>
|
12
|
+
</xsl:template>
|
13
|
+
|
14
|
+
</xsl:stylesheet>
|
@@ -0,0 +1,58 @@
|
|
1
|
+
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
|
2
|
+
@prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#> .
|
3
|
+
@prefix owl: <http://www.w3.org/2002/07/owl#> .
|
4
|
+
@prefix xsd: <http://www.w3.org/2001/XMLSchema#> .
|
5
|
+
@prefix dct: <http://purl.org/dc/terms/> .
|
6
|
+
@prefix skos: <http://www.w3.org/2004/02/skos/core#> .
|
7
|
+
@prefix ci: <https://privatealpha.com/ontology/content-inventory/1#> .
|
8
|
+
@prefix tfo: <https://privatealpha.com/ontology/transformation/1#> .
|
9
|
+
@prefix xf: <tag:makethingsmakesense.com,2020:transform/> .
|
10
|
+
|
11
|
+
xf:prefix a tfo:Parameter ;
|
12
|
+
skos:prefLabel "Prefix"@en ;
|
13
|
+
rdfs:comment "A compact prefix declaration of the form prefix:url"@en ;
|
14
|
+
dct:identifier "prefix"^^xsd:token ;
|
15
|
+
rdfs:range xsd:token .
|
16
|
+
|
17
|
+
xf:xpath a tfo:Parameter ;
|
18
|
+
skos:prefLabel "XPath"@en ;
|
19
|
+
rdfs:comment "An XPath expression"@en ;
|
20
|
+
dct:identifier "xpath"^^xsd:token ;
|
21
|
+
owl:cardinality 1 ;
|
22
|
+
rdfs:range xsd:string .
|
23
|
+
|
24
|
+
xf:reindent a tfo:Parameter ;
|
25
|
+
skos:prefLabel "Reindent"@en ;
|
26
|
+
rdfs:comment "Reindent the XML tree"@en ;
|
27
|
+
dct:identifier "reindent"^^xsd:token ;
|
28
|
+
tfo:default true ;
|
29
|
+
owl:cardinality 1 ;
|
30
|
+
rdfs:range xsd:boolean .
|
31
|
+
|
32
|
+
xf:subtree a tfo:Transform ;
|
33
|
+
skos:prefLabel "Subtree"@en ;
|
34
|
+
rdfs:comment "Isolate an X(HT)ML node using XPath."@en ;
|
35
|
+
tfo:implementation <urn:x-ruby:RDF::SAK::Transform::XPath> ;
|
36
|
+
tfo:accepts "application/xml"^^tfo:content-type ;
|
37
|
+
tfo:returns "application/xml"^^tfo:content-type ;
|
38
|
+
tfo:parameter xf:xpath, xf:prefix, xf:reindent ;
|
39
|
+
tfo:parameter-list ( xf:xpath xf:prefix xf:reindent ) .
|
40
|
+
|
41
|
+
xf:cleanup a tfo:Transform ;
|
42
|
+
skos:prefLabel "Cleanup"@en ;
|
43
|
+
rdfs:comment "Apply cleanup.xsl to the input."@en ;
|
44
|
+
tfo:implementation <file:example/cleanup.xsl> ;
|
45
|
+
tfo:accepts "application/xml"^^tfo:content-type ;
|
46
|
+
tfo:returns "application/xml"^^tfo:content-type .
|
47
|
+
|
48
|
+
<urn:uuid:78e6d8ce-a88a-4be0-8bfa-079136945816> a tfo:Partial ;
|
49
|
+
tfo:transform xf:subtree ;
|
50
|
+
xf:xpath "//html:main[1]"^^xsd:string ;
|
51
|
+
xf:prefix "html:http://www.w3.org/1999/xhtml"^^xsd:token .
|
52
|
+
|
53
|
+
<urn:uuid:4498eef5-1ca6-4034-937a-d50033dd6693> a tfo:Application ;
|
54
|
+
tfo:input <ni:///sha-256;0GHHmDtxh9CRZttXdr-cX78u72auS2P-O6tDXxvz2kU> ;
|
55
|
+
tfo:output <ni:///sha-256;_BbLbNSZl0TcQcjz-v3qF5fa5VL11rdha7c24K44pTc> ;
|
56
|
+
tfo:transform xf:subtree ;
|
57
|
+
xf:xpath "//html:main[1]"^^xsd:string ;
|
58
|
+
xf:prefix "html:http://www.w3.org/1999/xhtml"^^xsd:token .
|
data/lib/rdf-sak.rb
ADDED
@@ -0,0 +1 @@
|
|
1
|
+
require 'rdf/sak'
|
data/lib/rdf/sak.rb
ADDED
@@ -0,0 +1,2506 @@
|
|
1
|
+
# -*- coding: utf-8 -*-
|
2
|
+
require 'rdf/sak/version'
|
3
|
+
|
4
|
+
# basic stuff
|
5
|
+
require 'stringio'
|
6
|
+
require 'pathname'
|
7
|
+
require 'tempfile'
|
8
|
+
|
9
|
+
# rdf stuff
|
10
|
+
require 'uri'
|
11
|
+
require 'uri/urn'
|
12
|
+
require 'rdf'
|
13
|
+
require 'rdf/reasoner'
|
14
|
+
require 'linkeddata'
|
15
|
+
|
16
|
+
# my stuff
|
17
|
+
require 'xml-mixup'
|
18
|
+
require 'md-noko'
|
19
|
+
require 'uuid-ncname'
|
20
|
+
require 'rdf/sak/mimemagic'
|
21
|
+
require 'rdf/sak/util'
|
22
|
+
|
23
|
+
# ontologies, mine in particular
|
24
|
+
require 'rdf/sak/ci'
|
25
|
+
require 'rdf/sak/ibis'
|
26
|
+
# others not included in rdf.rb
|
27
|
+
require 'rdf/sak/pav'
|
28
|
+
require 'rdf/sak/qb'
|
29
|
+
|
30
|
+
module RDF::SAK
|
31
|
+
|
32
|
+
class Context
|
33
|
+
include XML::Mixup
|
34
|
+
include Util
|
35
|
+
|
36
|
+
private
|
37
|
+
|
38
|
+
# RDF::Reasoner.apply(:rdfs, :owl)
|
39
|
+
|
40
|
+
G_OK = [RDF::Repository, RDF::Dataset, RDF::Graph].freeze
|
41
|
+
C_OK = [Pathname, IO, String].freeze
|
42
|
+
|
43
|
+
def coerce_to_path_or_io obj
|
44
|
+
return obj if obj.is_a? IO
|
45
|
+
return obj.expand_path if obj.is_a? Pathname
|
46
|
+
raise "#{obj.inspect} is not stringable" unless obj.respond_to? :to_s
|
47
|
+
Pathname(obj.to_s).expand_path
|
48
|
+
end
|
49
|
+
|
50
|
+
def coerce_graph graph = nil, type: nil
|
51
|
+
# begin with empty graph
|
52
|
+
out = RDF::Repository.new
|
53
|
+
|
54
|
+
return out unless graph
|
55
|
+
return graph if G_OK.any? { |c| graph.is_a? c }
|
56
|
+
|
57
|
+
# now turn into an array
|
58
|
+
graph = [graph] unless graph.is_a? Array
|
59
|
+
|
60
|
+
graph.each do |g|
|
61
|
+
raise 'Graph must be some kind of RDF::Graph or RDF data file' unless
|
62
|
+
C_OK.any? { |c| g.is_a? c } || g.respond_to?(:to_s)
|
63
|
+
|
64
|
+
opts = {}
|
65
|
+
opts[:content_type] = type if type
|
66
|
+
|
67
|
+
if g.is_a? Pathname
|
68
|
+
opts[:filename] = g.expand_path.to_s
|
69
|
+
g = g.open
|
70
|
+
elsif g.is_a? File
|
71
|
+
opts[:filename] = g.path
|
72
|
+
end
|
73
|
+
|
74
|
+
g = StringIO.new(g.to_s) unless g.is_a? IO
|
75
|
+
reader = RDF::Reader.for(opts) do
|
76
|
+
g.rewind
|
77
|
+
sample = g.read 1000
|
78
|
+
g.rewind
|
79
|
+
sample
|
80
|
+
end or raise "Could not find an RDF::Reader for #{opts[:content_type]}"
|
81
|
+
|
82
|
+
reader = reader.new g, **opts
|
83
|
+
reader.each_statement do |stmt|
|
84
|
+
out << stmt
|
85
|
+
end
|
86
|
+
end
|
87
|
+
|
88
|
+
out
|
89
|
+
end
|
90
|
+
|
91
|
+
def normalize_hash h
|
92
|
+
return h unless h.is_a? Hash
|
93
|
+
out = {}
|
94
|
+
h.each do |k, v|
|
95
|
+
out[k.to_s.to_sym] = v.is_a?(Hash) ? normalize_hash(v) :
|
96
|
+
v.respond_to?(:to_a) ? v.to_a.map { |x| normalize_hash x } : v
|
97
|
+
end
|
98
|
+
out
|
99
|
+
end
|
100
|
+
|
101
|
+
def coerce_config config
|
102
|
+
# config must either be a hash or a file name/pathname/io object
|
103
|
+
unless config.respond_to? :to_h
|
104
|
+
# when in rome
|
105
|
+
require 'yaml'
|
106
|
+
config = if config.is_a? IO
|
107
|
+
YAML.load config
|
108
|
+
else
|
109
|
+
YAML.load_file Pathname.new(config).expand_path
|
110
|
+
end
|
111
|
+
end
|
112
|
+
|
113
|
+
config = normalize_hash config
|
114
|
+
|
115
|
+
# config MUST have source and target dirs
|
116
|
+
raise 'Config must have :source, :target, and :private directories' unless
|
117
|
+
([:source, :target, :private] - config.keys).empty?
|
118
|
+
[:source, :target].each do |path|
|
119
|
+
dir = config[path] = Pathname.new(config[path]).expand_path
|
120
|
+
raise "#{dir} is not a readable directory" unless
|
121
|
+
dir.directory? && dir.readable?
|
122
|
+
end
|
123
|
+
raise "Target directory #{config[:target]} is not writable" unless
|
124
|
+
config[:target].writable?
|
125
|
+
raise "Source and target directories are the same: #{config[:source]}" if
|
126
|
+
config[:source] == config[:target]
|
127
|
+
|
128
|
+
# we try to create the private directory
|
129
|
+
config[:private] = config[:target] + config[:private]
|
130
|
+
if config[:private].exist?
|
131
|
+
raise "#{config[:private]} is not a readable/writable directory" unless
|
132
|
+
[:directory?, :readable?, :writable?].all? do |m|
|
133
|
+
config[:private].send m
|
134
|
+
end
|
135
|
+
else
|
136
|
+
config[:private].mkpath
|
137
|
+
end
|
138
|
+
|
139
|
+
# config MAY have graph location(s) but we can test this other
|
140
|
+
# ways, same goes for base URI
|
141
|
+
if config[:graph]
|
142
|
+
g = config[:graph]
|
143
|
+
g = [g] unless g.is_a? Array
|
144
|
+
config[:graph] = g.map { |x| Pathname.new(x).expand_path }
|
145
|
+
end
|
146
|
+
|
147
|
+
# deal with prefix map
|
148
|
+
if config[:prefixes]
|
149
|
+
config[:prefixes] = config[:prefixes].transform_values do |p|
|
150
|
+
# we have to wrap this in case it fails
|
151
|
+
begin
|
152
|
+
RDF::Vocabulary.find_term(p) || RDF::URI(p)
|
153
|
+
rescue
|
154
|
+
RDF::URI(p)
|
155
|
+
end
|
156
|
+
end
|
157
|
+
end
|
158
|
+
|
159
|
+
if dups = config[:duplicate]
|
160
|
+
pfx = config[:prefixes] || {}
|
161
|
+
base = URI(uri_pp config[:base])
|
162
|
+
if dups.is_a? Hash
|
163
|
+
config[:duplicate] = dups.map do |ruri, preds|
|
164
|
+
preds = [preds] unless preds.is_a? Array
|
165
|
+
preds.map! do |p|
|
166
|
+
resolve_curie p, prefixes: pfx, scalar: true, coerce: :rdf
|
167
|
+
end
|
168
|
+
[RDF::URI((base + ruri.to_s).to_s), Set.new(preds)]
|
169
|
+
end.to_h
|
170
|
+
end
|
171
|
+
end
|
172
|
+
|
173
|
+
# rewrite maps
|
174
|
+
config[:maps] = {} unless config[:maps].is_a? Hash
|
175
|
+
%w(rewrite redirect gone).each do |type|
|
176
|
+
config[:maps][type.to_sym] ||= ".#{type}.map"
|
177
|
+
end
|
178
|
+
|
179
|
+
config
|
180
|
+
end
|
181
|
+
|
182
|
+
def cmp_label a, b, labels: nil, supplant: true, reverse: false
|
183
|
+
labels ||= {}
|
184
|
+
|
185
|
+
# try supplied label or fall back
|
186
|
+
pair = [a, b].map do |x|
|
187
|
+
if labels[x]
|
188
|
+
labels[x][1]
|
189
|
+
elsif supplant and y = label_for(x)
|
190
|
+
labels[x] = y
|
191
|
+
y[1]
|
192
|
+
else
|
193
|
+
x
|
194
|
+
end
|
195
|
+
end
|
196
|
+
|
197
|
+
pair.reverse! if reverse
|
198
|
+
# warn "#{pair[0]} <=> #{pair[1]}"
|
199
|
+
pair[0].to_s <=> pair[1].to_s
|
200
|
+
end
|
201
|
+
|
202
|
+
def term_list terms
|
203
|
+
return [] if terms.nil?
|
204
|
+
terms = terms.respond_to?(:to_a) ? terms.to_a : [terms]
|
205
|
+
terms.uniq.map { |t| RDF::Vocabulary.find_term t }.compact
|
206
|
+
end
|
207
|
+
|
208
|
+
def coerce_resource arg
|
209
|
+
super arg, @base
|
210
|
+
end
|
211
|
+
|
212
|
+
def coerce_uuid_urn arg
|
213
|
+
super arg, @base
|
214
|
+
end
|
215
|
+
|
216
|
+
public
|
217
|
+
|
218
|
+
attr_reader :config, :graph, :base
|
219
|
+
|
220
|
+
# Initialize a context.
|
221
|
+
#
|
222
|
+
# @param graph
|
223
|
+
# @param base
|
224
|
+
# @param config
|
225
|
+
# @param type
|
226
|
+
#
|
227
|
+
# @return [RDF::SAK::Context] the new context object.
|
228
|
+
|
229
|
+
def initialize graph: nil, base: nil, config: nil, type: nil
|
230
|
+
# RDF::Reasoner.apply(:rdfs, :owl)
|
231
|
+
|
232
|
+
@config = coerce_config config
|
233
|
+
|
234
|
+
graph ||= @config[:graph] if @config[:graph]
|
235
|
+
base ||= @config[:base] if @config[:base]
|
236
|
+
|
237
|
+
@graph = coerce_graph graph, type: type
|
238
|
+
@base = RDF::URI.new base.to_s if base
|
239
|
+
@ucache = RDF::Util::Cache.new(-1)
|
240
|
+
@scache = {} # wtf rdf util cache doesn't like booleans
|
241
|
+
end
|
242
|
+
|
243
|
+
# Get the prefix mappings from the configuration.
|
244
|
+
#
|
245
|
+
# @return [Hash]
|
246
|
+
|
247
|
+
def prefixes
|
248
|
+
@config[:prefixes] || {}
|
249
|
+
end
|
250
|
+
|
251
|
+
# Abbreviate a set of terms against the registered namespace
|
252
|
+
# prefixes and optional default vocabulary, or otherwise return a
|
253
|
+
# string representation of the original URI.
|
254
|
+
|
255
|
+
# @param term [RDF::Term]
|
256
|
+
# @param prefixes [Hash]
|
257
|
+
#
|
258
|
+
# @return [String]
|
259
|
+
#
|
260
|
+
def abbreviate term, prefixes: @config[:prefixes],
|
261
|
+
vocab: nil, noop: true, sort: true
|
262
|
+
super term, prefixes: prefixes || {}, vocab: vocab, noop: noop, sort: sort
|
263
|
+
end
|
264
|
+
|
265
|
+
# Obtain a key-value structure for the given subject, optionally
|
266
|
+
# constraining the result by node type (:resource, :uri/:iri,
|
267
|
+
# :blank/:bnode, :literal)
|
268
|
+
#
|
269
|
+
# @param subject of the inquiry
|
270
|
+
# @param rev map in reverse
|
271
|
+
# @param only one or more node types
|
272
|
+
# @param uuids coerce resources to if possible
|
273
|
+
#
|
274
|
+
# @return [Hash]
|
275
|
+
#
|
276
|
+
def struct_for subject, rev: false, only: [], uuids: false, canon: false
|
277
|
+
Util.struct_for @graph, subject,
|
278
|
+
rev: rev, only: only, uuids: uuids, canon: canon
|
279
|
+
end
|
280
|
+
|
281
|
+
# Obtain everything in the graph that is an `rdf:type` of something.
|
282
|
+
#
|
283
|
+
# @return [Array]
|
284
|
+
#
|
285
|
+
def all_types
|
286
|
+
@graph.query([nil, RDF.type, nil]).objects.uniq
|
287
|
+
end
|
288
|
+
|
289
|
+
# Obtain every subject that is rdf:type the given type or its subtypes.
|
290
|
+
#
|
291
|
+
# @param rdftype [RDF::Term]
|
292
|
+
#
|
293
|
+
# @return [Array]
|
294
|
+
#
|
295
|
+
def all_of_type rdftype, exclude: []
|
296
|
+
exclude = term_list exclude
|
297
|
+
t = RDF::Vocabulary.find_term(rdftype) or raise "No type #{rdftype.to_s}"
|
298
|
+
out = []
|
299
|
+
(all_types & all_related(t) - exclude).each do |type|
|
300
|
+
out += @graph.query([nil, RDF.type, type]).subjects
|
301
|
+
end
|
302
|
+
|
303
|
+
out.uniq
|
304
|
+
end
|
305
|
+
|
306
|
+
# Obtain all and only the rdf:types directly asserted on the subject.
|
307
|
+
#
|
308
|
+
# @param subject [RDF::Resource]
|
309
|
+
# @param type [RDF::Term, :to_a]
|
310
|
+
#
|
311
|
+
# @return [Array]
|
312
|
+
#
|
313
|
+
def asserted_types subject, type = nil
|
314
|
+
Util.asserted_types @graph, subject, type
|
315
|
+
end
|
316
|
+
|
317
|
+
# Obtain the canonical UUID for the given URI
|
318
|
+
#
|
319
|
+
# @param uri [RDF::URI, URI, to_s] the subject of the inquiry
|
320
|
+
# @param unique [true, false] return a single resource/nil or an array
|
321
|
+
# @param published [true, false] whether to restrict to published docs
|
322
|
+
#
|
323
|
+
# @return [RDF::URI, Array]
|
324
|
+
#
|
325
|
+
def canonical_uuid uri, unique: true, published: false
|
326
|
+
Util.canonical_uuid @graph, uri, unique: unique,
|
327
|
+
published: published, scache: @scache, ucache: @ucache, base: @base
|
328
|
+
end
|
329
|
+
|
330
|
+
# Obtain the "best" dereferenceable URI for the subject.
|
331
|
+
# Optionally returns all candidates.
|
332
|
+
#
|
333
|
+
# @param subject [RDF::Resource]
|
334
|
+
# @param unique [true, false] flag for unique return value
|
335
|
+
# @param rdf [true, false] flag to specify RDF::URI vs URI
|
336
|
+
# @param slugs [true, false] flag to include slugs
|
337
|
+
# @param fragment [true, false] flag to include fragment URIs
|
338
|
+
#
|
339
|
+
# @return [RDF::URI, URI, Array]
|
340
|
+
#
|
341
|
+
def canonical_uri subject,
|
342
|
+
unique: true, rdf: true, slugs: false, fragment: false
|
343
|
+
Util.canonical_uri @graph, subject, base: @base,
|
344
|
+
unique: unique, rdf: rdf, slugs: slugs, fragment: fragment
|
345
|
+
end
|
346
|
+
|
347
|
+
# Returns subjects from the graph with entailment.
|
348
|
+
#
|
349
|
+
# @param predicate
|
350
|
+
# @param object
|
351
|
+
# @param entail
|
352
|
+
# @param only
|
353
|
+
#
|
354
|
+
# @return [RDF::Resource]
|
355
|
+
#
|
356
|
+
def subjects_for predicate, object, entail: true, only: []
|
357
|
+
Util.subjects_for @graph, predicate, object, entail: entail, only: only
|
358
|
+
end
|
359
|
+
|
360
|
+
# Returns objects from the graph with entailment.
|
361
|
+
#
|
362
|
+
# @param subject
|
363
|
+
# @param predicate
|
364
|
+
# @param entail
|
365
|
+
# @param only
|
366
|
+
# @param datatype
|
367
|
+
#
|
368
|
+
# @return [RDF::Term]
|
369
|
+
#
|
370
|
+
def objects_for subject, predicate, entail: true, only: [], datatype: nil
|
371
|
+
Util.objects_for @graph, subject, predicate,
|
372
|
+
entail: entail, only: only, datatype: datatype
|
373
|
+
end
|
374
|
+
|
375
|
+
# Find the terminal replacements for the given subject, if any exist.
|
376
|
+
#
|
377
|
+
# @param subject
|
378
|
+
# @param published indicate the context is published
|
379
|
+
#
|
380
|
+
# @return [Set]
|
381
|
+
#
|
382
|
+
def replacements_for subject, published: true
|
383
|
+
Util.replacements_for @graph, subject, published: published
|
384
|
+
end
|
385
|
+
|
386
|
+
# Obtain dates for the subject as instances of Date(Time). This is
|
387
|
+
# just shorthand for a common application of `objects_for`.
|
388
|
+
#
|
389
|
+
# @param subject
|
390
|
+
# @param predicate
|
391
|
+
# @param datatype
|
392
|
+
#
|
393
|
+
# @return [Array] of dates
|
394
|
+
def dates_for subject, predicate: RDF::Vocab::DC.date,
|
395
|
+
datatype: [RDF::XSD.date, RDF::XSD.dateTime]
|
396
|
+
Util.dates_for @graph, subject, predicate: predicate, datatype: datatype
|
397
|
+
end
|
398
|
+
|
399
|
+
# Obtain any specified MIME types for the subject. Just shorthand
|
400
|
+
# for a common application of `objects_for`.
|
401
|
+
#
|
402
|
+
# @param subject
|
403
|
+
# @param predicate
|
404
|
+
# @param datatype
|
405
|
+
#
|
406
|
+
# @return [Array] of internet media types
|
407
|
+
#
|
408
|
+
def formats_for subject, predicate: RDF::Vocab::DC.format,
|
409
|
+
datatype: [RDF::XSD.token]
|
410
|
+
Util.objects_for @graph, subject, predicate: predicate, datatype: datatype
|
411
|
+
end
|
412
|
+
|
413
|
+
# Assuming the subject is a thing that has authors, return the
|
414
|
+
# list of authors. Try bibo:authorList first for an explicit
|
415
|
+
# ordering, then continue to the various other predicates.
|
416
|
+
#
|
417
|
+
# @param subject [RDF::Resource]
|
418
|
+
# @param unique [false, true] only return the first author
|
419
|
+
# @param contrib [false, true] return contributors instead of authors
|
420
|
+
#
|
421
|
+
# @return [RDF::Value, Array]
|
422
|
+
#
|
423
|
+
def authors_for subject, unique: false, contrib: false
|
424
|
+
Util.authors_for @graph, subject, unique: unique, contrib: contrib
|
425
|
+
end
|
426
|
+
|
427
|
+
# Obtain the most appropriate label(s) for the subject's type(s).
|
428
|
+
# Returns one or more (depending on the `unique` flag)
|
429
|
+
# predicate-object pairs in order of preference.
|
430
|
+
#
|
431
|
+
# @param subject [RDF::Resource]
|
432
|
+
# @param unique [true, false] only return the first pair
|
433
|
+
# @param type [RDF::Term, Array] supply asserted types if already retrieved
|
434
|
+
# @param lang [nil] not currently implemented (will be conneg)
|
435
|
+
# @param desc [false, true] retrieve description instead of label
|
436
|
+
# @param alt [false, true] retrieve alternate instead of main
|
437
|
+
#
|
438
|
+
# @return [Array] either a predicate-object pair or an array of pairs.
|
439
|
+
#
|
440
|
+
def label_for subject, candidates: nil, unique: true, type: nil,
|
441
|
+
lang: nil, desc: false, alt: false
|
442
|
+
Util.label_for @graph, subject, candidates: candidates,
|
443
|
+
unique: unique, type: type, lang: lang, desc: desc, alt: alt
|
444
|
+
end
|
445
|
+
|
446
|
+
SKOS_HIER = [
|
447
|
+
{
|
448
|
+
element: :subject,
|
449
|
+
pattern: -> c, p { [nil, p, c] },
|
450
|
+
preds: [RDF::Vocab::SKOS.broader, RDF::Vocab::SKOS.broaderTransitive],
|
451
|
+
},
|
452
|
+
{
|
453
|
+
element: :object,
|
454
|
+
pattern: -> c, p { [c, p, nil] },
|
455
|
+
preds: [RDF::Vocab::SKOS.narrower, RDF::Vocab::SKOS.narrowerTransitive],
|
456
|
+
}
|
457
|
+
]
|
458
|
+
SKOS_HIER.each do |struct|
|
459
|
+
# lol how many times are we gonna cart this thing around
|
460
|
+
preds = struct[:preds]
|
461
|
+
i = 0
|
462
|
+
loop do
|
463
|
+
equiv = preds[i].entail(:equivalentProperty) - preds
|
464
|
+
preds.insert(i + 1, *equiv) unless equiv.empty?
|
465
|
+
i += equiv.length + 1;
|
466
|
+
break if i >= preds.length
|
467
|
+
end
|
468
|
+
end
|
469
|
+
|
470
|
+
def sub_concepts concept, extra: []
|
471
|
+
raise 'Concept must be exactly one concept' unless
|
472
|
+
concept.is_a? RDF::Resource
|
473
|
+
extra = term_list extra
|
474
|
+
|
475
|
+
# we need an array for a queue, and a set to accumulate the
|
476
|
+
# output as well as a separate 'seen' set
|
477
|
+
queue = [concept]
|
478
|
+
seen = Set.new queue.dup
|
479
|
+
out = seen.dup
|
480
|
+
|
481
|
+
# it turns out that the main SKOS hierarchy terms, while not
|
482
|
+
# being transitive themselves, are subproperties of transitive
|
483
|
+
# relations which means they are as good as being transitive.
|
484
|
+
|
485
|
+
while c = queue.shift
|
486
|
+
SKOS_HIER.each do |struct|
|
487
|
+
elem, pat, preds = struct.values_at(:element, :pattern, :preds)
|
488
|
+
preds.each do |p|
|
489
|
+
@graph.query(pat.call c, p).each do |stmt|
|
490
|
+
# obtain hierarchical element
|
491
|
+
hierc = stmt.send elem
|
492
|
+
|
493
|
+
# skip any further processing if we have seen this concept
|
494
|
+
next if seen.include? hierc
|
495
|
+
seen << hierc
|
496
|
+
|
497
|
+
next if !extra.empty? and !extra.any? do |t|
|
498
|
+
@graph.has_statement? RDF::Statement.new(hierc, RDF.type, t)
|
499
|
+
end
|
500
|
+
|
501
|
+
queue << hierc
|
502
|
+
out << hierc
|
503
|
+
end
|
504
|
+
end
|
505
|
+
end
|
506
|
+
end
|
507
|
+
|
508
|
+
out.to_a.sort
|
509
|
+
end
|
510
|
+
|
511
|
+
def audiences_for uuid, proximate: false, invert: false
|
512
|
+
p = invert ? CI['non-audience'] : RDF::Vocab::DC.audience
|
513
|
+
return @graph.query([uuid, p, nil]).objects if proximate
|
514
|
+
|
515
|
+
out = []
|
516
|
+
@graph.query([uuid, p, nil]).objects.each do |o|
|
517
|
+
out += sub_concepts o
|
518
|
+
end
|
519
|
+
|
520
|
+
out
|
521
|
+
end
|
522
|
+
|
523
|
+
# Get all "reachable" UUID-identified entities (subjects which are
|
524
|
+
# also objects)
|
525
|
+
def reachable published: false
|
526
|
+
p = published ? -> x { published?(x) } : -> x { true }
|
527
|
+
# now get the subjects which are also objects
|
528
|
+
@graph.subjects.select do |s|
|
529
|
+
s.uri? && s =~ /^urn:uuid:/ && @graph.has_object?(s) && p.call(s)
|
530
|
+
end
|
531
|
+
end
|
532
|
+
|
533
|
+
# holy cow this is actually a lot of stuff:
|
534
|
+
|
535
|
+
# turn markdown into xhtml (via md-noko)
|
536
|
+
|
537
|
+
# turn html into xhtml (trivial)
|
538
|
+
|
539
|
+
# generate triples from ordinary (x)html structure
|
540
|
+
|
541
|
+
# map vanilla (x)html metadata to existing graph (ie to get resource URIs)
|
542
|
+
|
543
|
+
# pull triples from rdfa
|
544
|
+
|
545
|
+
# stuff rdfa into rdfa-less xhtml
|
546
|
+
|
547
|
+
# basic nlp detection of terms + text-level markup (dfn, abbr...)
|
548
|
+
|
549
|
+
# markdown round-tripping (may as well store source in md if possible)
|
550
|
+
|
551
|
+
# add title attribute to all links
|
552
|
+
|
553
|
+
# add alt attribute to all images
|
554
|
+
|
555
|
+
# segmentation of composite documents into multiple files
|
556
|
+
|
557
|
+
# aggregation of simple documents into composites
|
558
|
+
|
559
|
+
# generate backlinks
|
560
|
+
|
561
|
+
# - resource (ie file) generation -
|
562
|
+
|
563
|
+
# generate indexes of people, groups, and organizations
|
564
|
+
|
565
|
+
# generate indexes of books, not-books, and other external links
|
566
|
+
|
567
|
+
def head_links subject, struct: nil, nodes: nil, prefixes: {},
|
568
|
+
ignore: [], uris: {}, labels: {}, vocab: nil
|
569
|
+
|
570
|
+
raise 'ignore must be Array or Set' unless
|
571
|
+
[Array, Set].any? { |c| ignore.is_a? c }
|
572
|
+
|
573
|
+
struct ||= struct_for subject
|
574
|
+
nodes ||= invert_struct struct
|
575
|
+
|
576
|
+
# make sure these are actually URI objects not RDF::URI
|
577
|
+
uris = uris.transform_values { |v| URI(uri_pp v.to_s) }
|
578
|
+
uri = uris[subject] || canonical_uri(subject, rdf: false)
|
579
|
+
|
580
|
+
ignore = ignore.to_set
|
581
|
+
|
582
|
+
# output
|
583
|
+
links = []
|
584
|
+
|
585
|
+
nodes.reject { |n, _| ignore.include?(n) || !n.uri? }.each do |k, v|
|
586
|
+
# first nuke rdf:type, that's never in there
|
587
|
+
v = v.dup.delete RDF::RDFV.type
|
588
|
+
next if v.empty?
|
589
|
+
|
590
|
+
unless uris[k]
|
591
|
+
cu = canonical_uri k
|
592
|
+
uris[k] = cu || uri_pp(k.to_s)
|
593
|
+
end
|
594
|
+
|
595
|
+
# munge the url and make the tag
|
596
|
+
rel = abbreviate v.to_a, vocab: vocab
|
597
|
+
ru = uri.route_to(uris[k])
|
598
|
+
ln = { nil => :link, rel: rel, href: ru.to_s }
|
599
|
+
|
600
|
+
# add the title
|
601
|
+
if lab = labels[k]
|
602
|
+
ln[:title] = lab[1].to_s
|
603
|
+
end
|
604
|
+
|
605
|
+
# add type attribute
|
606
|
+
unless (mts = formats_for k).empty?
|
607
|
+
ln[:type] = mts.first.to_s
|
608
|
+
|
609
|
+
if ln[:type] =~ /(java|ecma)script/i ||
|
610
|
+
!(v.to_set & Set[RDF::Vocab::DC.requires]).empty?
|
611
|
+
ln[:src] = ln.delete :href
|
612
|
+
# make sure we pass in an empty string so there is a closing tag
|
613
|
+
ln.delete nil
|
614
|
+
ln[['']] = :script
|
615
|
+
end
|
616
|
+
end
|
617
|
+
|
618
|
+
# finally add the link
|
619
|
+
links.push ln
|
620
|
+
end
|
621
|
+
|
622
|
+
links.sort! do |a, b|
|
623
|
+
# sort by rel, then by href
|
624
|
+
# warn a.inspect, b.inspect
|
625
|
+
s = 0
|
626
|
+
[nil, :rel, :rev, :href, :title].each do |k|
|
627
|
+
s = a.fetch(k, '').to_s <=> b.fetch(k, '').to_s
|
628
|
+
break if s != 0
|
629
|
+
end
|
630
|
+
s
|
631
|
+
end
|
632
|
+
|
633
|
+
links
|
634
|
+
end
|
635
|
+
|
636
|
+
def head_meta subject, struct: nil, nodes: nil, prefixes: {},
|
637
|
+
ignore: [], meta_names: {}, vocab: nil, lang: nil, xhtml: true
|
638
|
+
|
639
|
+
raise 'ignore must be Array or Set' unless
|
640
|
+
[Array, Set].any? { |c| ignore.is_a? c }
|
641
|
+
|
642
|
+
struct ||= struct_for subject
|
643
|
+
nodes ||= invert_struct struct
|
644
|
+
|
645
|
+
ignore = ignore.to_set
|
646
|
+
|
647
|
+
meta = []
|
648
|
+
nodes.select { |n| n.literal? && !ignore.include?(n) }.each do |k, v|
|
649
|
+
rel = abbreviate v.to_a, vocab: vocab
|
650
|
+
tag = { nil => :meta, property: rel, content: k.to_s }
|
651
|
+
|
652
|
+
lang = (k.language? && k.language != lang ? k.language : nil) ||
|
653
|
+
(k.datatype == RDF::XSD.string && lang ? '' : nil)
|
654
|
+
if lang
|
655
|
+
tag['xml:lang'] = lang if xhtml
|
656
|
+
tag[:lang] = lang
|
657
|
+
end
|
658
|
+
|
659
|
+
tag[:datatype] = abbreviate k.datatype, vocab: XHV if k.datatype?
|
660
|
+
tag[:name] = meta_names[k] if meta_names[k]
|
661
|
+
|
662
|
+
meta << tag
|
663
|
+
end
|
664
|
+
|
665
|
+
meta.sort! do |a, b|
|
666
|
+
s = 0
|
667
|
+
[:about, :property, :datatype, :content, :name].each do |k|
|
668
|
+
# warn a.inspect, b.inspect
|
669
|
+
s = a.fetch(k, '').to_s <=> b.fetch(k, '').to_s
|
670
|
+
break if s != 0
|
671
|
+
end
|
672
|
+
s
|
673
|
+
end
|
674
|
+
|
675
|
+
meta
|
676
|
+
end
|
677
|
+
|
678
|
+
def generate_backlinks subject, published: true, ignore: nil
|
679
|
+
uri = canonical_uri(subject, rdf: false) || URI(uri_pp subject)
|
680
|
+
ignore ||= Set.new
|
681
|
+
raise 'ignore must be amenable to a set' unless ignore.respond_to? :to_set
|
682
|
+
ignore = ignore.to_set
|
683
|
+
nodes = {}
|
684
|
+
labels = {}
|
685
|
+
types = {}
|
686
|
+
@graph.query([nil, nil, subject]).each do |stmt|
|
687
|
+
next if ignore.include?(sj = stmt.subject)
|
688
|
+
preds = nodes[sj] ||= Set.new
|
689
|
+
preds << (pr = stmt.predicate)
|
690
|
+
types[sj] ||= asserted_types sj
|
691
|
+
labels[sj] ||= label_for sj
|
692
|
+
labels[pr] ||= label_for pr
|
693
|
+
end
|
694
|
+
|
695
|
+
# prune out
|
696
|
+
nodes.select! { |k, _| published? k } if published
|
697
|
+
|
698
|
+
return if nodes.empty?
|
699
|
+
|
700
|
+
li = nodes.sort do |a, b|
|
701
|
+
cmp_label a[0], b[0], labels: labels
|
702
|
+
end.map do |rsrc, preds|
|
703
|
+
cu = canonical_uri(rsrc, rdf: false) or next
|
704
|
+
lab = labels[rsrc] || [nil, rsrc]
|
705
|
+
lp = abbreviate(lab[0]) if lab[0]
|
706
|
+
ty = abbreviate(types[rsrc]) if types[rsrc]
|
707
|
+
|
708
|
+
{ [{ [{ [lab[1].to_s] => :span, property: lp }] => :a,
|
709
|
+
href: uri.route_to(cu), typeof: ty, rev: abbreviate(preds) }] => :li }
|
710
|
+
end.compact
|
711
|
+
|
712
|
+
{ [{ li => :ul }] => :nav }
|
713
|
+
end
|
714
|
+
|
715
|
+
def generate_twitter_meta subject
|
716
|
+
# get author
|
717
|
+
author = authors_for(subject, unique: true) or return
|
718
|
+
|
719
|
+
# get author's twitter account
|
720
|
+
twitter = objects_for(author, RDF::Vocab::FOAF.account,
|
721
|
+
only: :resource).select { |t| t.to_s =~ /twitter\.com/
|
722
|
+
}.sort.first or return
|
723
|
+
twitter = URI(twitter.to_s).path.split(/\/+/)[1]
|
724
|
+
twitter = ?@ + twitter unless twitter.start_with? ?@
|
725
|
+
|
726
|
+
# get title
|
727
|
+
title = label_for(subject) or return
|
728
|
+
|
729
|
+
out = [
|
730
|
+
{ nil => :meta, name: 'twitter:card', content: :summary },
|
731
|
+
{ nil => :meta, name: 'twitter:site', content: twitter },
|
732
|
+
{ nil => :meta, name: 'twitter:title', content: title[1].to_s }
|
733
|
+
]
|
734
|
+
|
735
|
+
# get abstract
|
736
|
+
if desc = label_for(subject, desc: true)
|
737
|
+
out.push({ nil => :meta, name: 'twitter:description',
|
738
|
+
content: desc[1].to_s })
|
739
|
+
end
|
740
|
+
|
741
|
+
# get image (foaf:depiction)
|
742
|
+
img = objects_for(subject, RDF::Vocab::FOAF.depiction, only: :resource)
|
743
|
+
unless img.empty?
|
744
|
+
img = img[0].to_s
|
745
|
+
out.push({ nil => :meta, name: 'twitter:image', content: img })
|
746
|
+
out[0][:content] = :summary_large_image
|
747
|
+
end
|
748
|
+
|
749
|
+
# return the appropriate xml-mixup structure
|
750
|
+
out
|
751
|
+
end
|
752
|
+
|
753
|
+
AUTHOR_SPEC = [
|
754
|
+
['By:', [RDF::Vocab::BIBO.authorList, RDF::Vocab::DC.creator]],
|
755
|
+
['With:', [RDF::Vocab::BIBO.contributorList, RDF::Vocab::DC.contributor]],
|
756
|
+
['Edited by:', [RDF::Vocab::BIBO.editorList, RDF::Vocab::BIBO.editor]],
|
757
|
+
['Translated by:', [RDF::Vocab::BIBO.translator]],
|
758
|
+
].freeze
|
759
|
+
|
760
|
+
def generate_bibliography id, published: true
|
761
|
+
id = canonical_uuid id
|
762
|
+
uri = canonical_uri id
|
763
|
+
struct = struct_for id
|
764
|
+
nodes = Set[id] + smush_struct(struct)
|
765
|
+
bodynodes = Set.new
|
766
|
+
parts = {}
|
767
|
+
referents = {}
|
768
|
+
labels = { id => label_for(id, candidates: struct) }
|
769
|
+
canon = {}
|
770
|
+
|
771
|
+
# uggh put these somewhere
|
772
|
+
preds = {
|
773
|
+
hp: predicate_set(RDF::Vocab::DC.hasPart),
|
774
|
+
sa: predicate_set(RDF::RDFS.seeAlso),
|
775
|
+
canon: predicate_set([RDF::OWL.sameAs, CI.canonical]),
|
776
|
+
ref: predicate_set(RDF::Vocab::DC.references),
|
777
|
+
al: predicate_set(RDF::Vocab::BIBO.contributorList),
|
778
|
+
cont: predicate_set(RDF::Vocab::DC.contributor),
|
779
|
+
}
|
780
|
+
|
781
|
+
# collect up all the parts (as in dct:hasPart)
|
782
|
+
objects_for(id, preds[:hp], entail: false, only: :resource).each do |part|
|
783
|
+
bodynodes << part
|
784
|
+
|
785
|
+
# gather up all the possible alias urls this thing can have
|
786
|
+
sa = ([part] + objects_for(part,
|
787
|
+
preds[:sa], only: :uri, entail: false)).map do |x|
|
788
|
+
[x] + subjects_for(preds[:canon], x, only: :uri, entail: false)
|
789
|
+
end.flatten.uniq
|
790
|
+
|
791
|
+
# collect all the referents
|
792
|
+
reftmp = {}
|
793
|
+
sa.each do |u|
|
794
|
+
subjects_for preds[:ref], u, only: :uri, entail: false do |s, *p|
|
795
|
+
reftmp[s] ||= Set.new
|
796
|
+
reftmp[s] += p[0].to_set
|
797
|
+
end
|
798
|
+
end
|
799
|
+
|
800
|
+
# if we are producing a list of references identified by only
|
801
|
+
# published resources, prune out all the unpublished referents
|
802
|
+
reftmp.select! { |x, _| published? x } if published
|
803
|
+
|
804
|
+
# unconditionally skip this item if nothing references it
|
805
|
+
next if reftmp.empty?
|
806
|
+
|
807
|
+
referents[part] = reftmp
|
808
|
+
|
809
|
+
reftmp.each do |r, _|
|
810
|
+
labels[r] ||= label_for r
|
811
|
+
canon[r] ||= canonical_uri r
|
812
|
+
end
|
813
|
+
|
814
|
+
# collect all the authors and author lists
|
815
|
+
|
816
|
+
objects_for(part, preds[:al], only: :resource, entail: false) do |o|
|
817
|
+
RDF::List.new(subject: o, graph: @graph).each do |a|
|
818
|
+
labels[a] ||= label_for a
|
819
|
+
end
|
820
|
+
end
|
821
|
+
|
822
|
+
objects_for(part, preds[:cont], only: :uri, entail: false) do |a|
|
823
|
+
labels[a] ||= label_for a
|
824
|
+
end
|
825
|
+
|
826
|
+
ps = struct_for part
|
827
|
+
labels[part] = label_for part, candidates: ps
|
828
|
+
nodes |= smush_struct ps
|
829
|
+
|
830
|
+
parts[part] = ps
|
831
|
+
end
|
832
|
+
|
833
|
+
bmap = prepare_collation struct
|
834
|
+
pf = -> x { abbreviate bmap[x.literal? ? :literals : :resources][x] }
|
835
|
+
|
836
|
+
body = []
|
837
|
+
parts.sort { |a, b| cmp_label a[0], b[0], labels: labels }.each do |k, v|
|
838
|
+
mapping = prepare_collation v
|
839
|
+
p = -> x {
|
840
|
+
abbreviate mapping[x.literal? ? :literals : :resources][x] }
|
841
|
+
t = abbreviate mapping[:types]
|
842
|
+
|
843
|
+
lp = label_for k, candidates: v
|
844
|
+
h2c = [lp[1].to_s]
|
845
|
+
h2 = { h2c => :h2 }
|
846
|
+
cu = canonical_uri k
|
847
|
+
rel = nil
|
848
|
+
unless cu.scheme.downcase.start_with? 'http'
|
849
|
+
if sa = v[RDF::RDFS.seeAlso]
|
850
|
+
rel = p.call sa[0]
|
851
|
+
cu = canonical_uri sa[0]
|
852
|
+
else
|
853
|
+
cu = nil
|
854
|
+
end
|
855
|
+
end
|
856
|
+
|
857
|
+
if cu
|
858
|
+
h2c[0] = { [lp[1].to_s] => :a, rel: rel,
|
859
|
+
property: p.call(lp[1]), href: cu.to_s }
|
860
|
+
else
|
861
|
+
h2[:property] = p.call(lp[1])
|
862
|
+
end
|
863
|
+
|
864
|
+
# authors &c
|
865
|
+
# authors contributors editors translators
|
866
|
+
al = []
|
867
|
+
AUTHOR_SPEC.each do |label, pl|
|
868
|
+
dd = []
|
869
|
+
seen = Set.new
|
870
|
+
pl.each do |pred|
|
871
|
+
# first check if the struct has the predicate
|
872
|
+
next unless v[pred]
|
873
|
+
li = []
|
874
|
+
ul = { li => :ul, rel: abbreviate(pred) }
|
875
|
+
v[pred].sort { |a, b| cmp_label a, b, labels: labels }.each do |o|
|
876
|
+
# check if this is a list
|
877
|
+
tl = RDF::List.new subject: o, graph: @graph
|
878
|
+
if tl.empty? and !seen.include? o
|
879
|
+
seen << o
|
880
|
+
lab = labels[o] ? { [labels[o][1]] => :span,
|
881
|
+
property: abbreviate(labels[o][0]) } : o
|
882
|
+
li << { [lab] => :li, resource: o }
|
883
|
+
else
|
884
|
+
# XXX this will actually not be right if there are
|
885
|
+
# multiple lists but FINE FOR NOW
|
886
|
+
ul[:inlist] ||= ''
|
887
|
+
tl.each do |a|
|
888
|
+
seen << a
|
889
|
+
lab = labels[a] ? { [labels[a][1]] => :span,
|
890
|
+
property: abbreviate(labels[a][0]) } : a
|
891
|
+
li << { [lab] => :li, resource: a }
|
892
|
+
end
|
893
|
+
end
|
894
|
+
end
|
895
|
+
dd << ul unless li.empty?
|
896
|
+
end
|
897
|
+
al += [{ [label] => :dt }, { dd => :dd }] unless dd.empty?
|
898
|
+
end
|
899
|
+
|
900
|
+
# ref list
|
901
|
+
rl = referents[k].sort do |a, b|
|
902
|
+
cmp_label a[0], b[0], labels: labels
|
903
|
+
end.map do |ref, pset|
|
904
|
+
lab = labels[ref] ? { [labels[ref][1]] => :span,
|
905
|
+
property: abbreviate(labels[ref][0]) } : ref
|
906
|
+
|
907
|
+
{ [{ [lab] => :a, rev: abbreviate(pset), href: canon[ref] }] => :li }
|
908
|
+
end
|
909
|
+
|
910
|
+
contents = [h2, {
|
911
|
+
al + [{ ['Referenced in:'] => :dt },
|
912
|
+
{ [{ rl => :ul }] => :dd }] => :dl }]
|
913
|
+
|
914
|
+
body << { contents => :section,
|
915
|
+
rel: pf.call(k), resource: k.to_s, typeof: t }
|
916
|
+
end
|
917
|
+
|
918
|
+
# prepend abstract to body if it exists
|
919
|
+
abs = label_for id, candidates: struct, desc: true
|
920
|
+
if abs
|
921
|
+
tag = { '#p' => abs[1], property: abbreviate(abs[0]) }
|
922
|
+
body.unshift tag
|
923
|
+
end
|
924
|
+
|
925
|
+
# add labels to nodes
|
926
|
+
nodes += smush_struct labels
|
927
|
+
|
928
|
+
# get prefixes
|
929
|
+
pfx = prefix_subset prefixes, nodes
|
930
|
+
|
931
|
+
# get title tag
|
932
|
+
title = title_tag labels[id][0], labels[id][1],
|
933
|
+
prefixes: prefixes, lang: 'en'
|
934
|
+
|
935
|
+
# get links
|
936
|
+
link = head_links id,
|
937
|
+
struct: struct, ignore: bodynodes, labels: labels, vocab: XHV
|
938
|
+
|
939
|
+
# get metas
|
940
|
+
mn = {}
|
941
|
+
mn[abs[1]] = :description if abs
|
942
|
+
mi = Set.new
|
943
|
+
mi << labels[id][1] if labels[id]
|
944
|
+
meta = head_meta id,
|
945
|
+
struct: struct, lang: 'en', ignore: mi, meta_names: mn, vocab: XHV
|
946
|
+
|
947
|
+
meta += generate_twitter_meta(id) || []
|
948
|
+
|
949
|
+
xhtml_stub(base: uri, prefix: pfx, lang: 'en', title: title, vocab: XHV,
|
950
|
+
link: link, meta: meta, transform: @config[:transform],
|
951
|
+
body: { body => :body, about: '',
|
952
|
+
typeof: abbreviate(struct[RDF::RDFV.type] || []) }).document
|
953
|
+
end
|
954
|
+
|
955
|
+
# generate skos concept schemes
|
956
|
+
|
957
|
+
CONCEPTS = Util.all_related(RDF::Vocab::SKOS.Concept).to_set
|
958
|
+
|
959
|
+
def generate_audience_csv file = nil, published: true
|
960
|
+
require 'csv'
|
961
|
+
file = coerce_to_path_or_io file if file
|
962
|
+
lab = {}
|
963
|
+
|
964
|
+
out = all_internal_docs(published: published,
|
965
|
+
exclude: RDF::Vocab::FOAF.Image).map do |s|
|
966
|
+
u = canonical_uri s
|
967
|
+
x = struct_for s
|
968
|
+
c = x[RDF::Vocab::DC.created] ? x[RDF::Vocab::DC.created][0] : nil
|
969
|
+
_, t = label_for s, candidates: x
|
970
|
+
_, d = label_for s, candidates: x, desc: true
|
971
|
+
|
972
|
+
# # audience(s)
|
973
|
+
# a = objects_for(s, RDF::Vocab::DC.audience).map do |au|
|
974
|
+
# next lab[au] if lab[au]
|
975
|
+
# _, al = label_for au
|
976
|
+
# lab[au] = al
|
977
|
+
# end.map(&:to_s).sort.join '; '
|
978
|
+
|
979
|
+
# # explicit non-audience(s)
|
980
|
+
# n = objects_for(s, RDF::SAK::CI['non-audience']).map do |au|
|
981
|
+
# next lab[au] if lab[au]
|
982
|
+
# _, al = label_for au
|
983
|
+
# lab[au] = al
|
984
|
+
# end.map(&:to_s).sort.join '; '
|
985
|
+
|
986
|
+
# audience and non-audience
|
987
|
+
a, n = [RDF::Vocab::DC.audience, CI['non-audience']].map do |ap|
|
988
|
+
objects_for(s, ap).map do |au|
|
989
|
+
next lab[au] if lab[au]
|
990
|
+
_, al = label_for au
|
991
|
+
lab[au] = al
|
992
|
+
end.map(&:to_s).sort.join '; '
|
993
|
+
end
|
994
|
+
|
995
|
+
# concepts???
|
996
|
+
concepts = [RDF::Vocab::DC.subject, CI.introduces,
|
997
|
+
CI.assumes, CI.mentions].map do |pred|
|
998
|
+
objects_for(s, pred, only: :resource).map do |o|
|
999
|
+
con = self.objects_for(o, RDF.type).to_set & CONCEPTS
|
1000
|
+
next if con.empty?
|
1001
|
+
next lab[o] if lab[o]
|
1002
|
+
_, ol = label_for o
|
1003
|
+
lab[o] = ol
|
1004
|
+
end.compact.map(&:to_s).sort.join '; '
|
1005
|
+
end
|
1006
|
+
|
1007
|
+
[s, u, c, t, d, a, n].map(&:to_s) + concepts
|
1008
|
+
end.sort { |a, b| a[2] <=> b[2] }
|
1009
|
+
|
1010
|
+
out.unshift ['ID', 'URL', 'Created', 'Title', 'Description', 'Audience',
|
1011
|
+
'Non-Audience', 'Subject', 'Introduces', 'Assumes', 'Mentions']
|
1012
|
+
|
1013
|
+
if file
|
1014
|
+
# don't open until now
|
1015
|
+
file = file.expand_path.open('wb') unless file.is_a? IO
|
1016
|
+
|
1017
|
+
csv = CSV.new file
|
1018
|
+
out.each { |x| csv << x }
|
1019
|
+
file.flush
|
1020
|
+
end
|
1021
|
+
|
1022
|
+
out
|
1023
|
+
end
|
1024
|
+
|
1025
|
+
CSV_PRED = {
|
1026
|
+
audience: RDF::Vocab::DC.audience,
|
1027
|
+
nonaudience: CI['non-audience'],
|
1028
|
+
subject: RDF::Vocab::DC.subject,
|
1029
|
+
assumes: CI.assumes,
|
1030
|
+
introduces: CI.introduces,
|
1031
|
+
mentions: CI.mentions,
|
1032
|
+
}
|
1033
|
+
|
1034
|
+
def ingest_csv file
|
1035
|
+
file = coerce_to_path_or_io file
|
1036
|
+
|
1037
|
+
require 'csv'
|
1038
|
+
|
1039
|
+
# key mapper
|
1040
|
+
km = { uuid: :id, url: :uri }
|
1041
|
+
kt = -> (k) { km[k] || k }
|
1042
|
+
|
1043
|
+
# grab all the concepts and audiences
|
1044
|
+
|
1045
|
+
audiences = {}
|
1046
|
+
all_of_type(CI.Audience).map do |c|
|
1047
|
+
s = struct_for c
|
1048
|
+
|
1049
|
+
# homogenize the labels
|
1050
|
+
lab = [false, true].map do |b|
|
1051
|
+
label_for(c, candidates: s, unique: false, alt: b).map { |x| x[1] }
|
1052
|
+
end.flatten.map { |x| x.to_s.strip.downcase }
|
1053
|
+
|
1054
|
+
# we want all the keys to share the same set
|
1055
|
+
set = nil
|
1056
|
+
lab.each { |t| set = audiences[t] ||= set || Set.new }
|
1057
|
+
set << c
|
1058
|
+
end
|
1059
|
+
|
1060
|
+
concepts = {}
|
1061
|
+
all_of_type(RDF::Vocab::SKOS.Concept).map do |c|
|
1062
|
+
s = struct_for c
|
1063
|
+
|
1064
|
+
# homogenize the labels
|
1065
|
+
lab = [false, true].map do |b|
|
1066
|
+
label_for(c, candidates: s, unique: false, alt: b).map { |x| x[1] }
|
1067
|
+
end.flatten.map { |x| x.to_s.strip.downcase }
|
1068
|
+
|
1069
|
+
# we want all the keys to share the same set
|
1070
|
+
set = nil
|
1071
|
+
lab.each { |t| set = concepts[t] ||= set || Set.new }
|
1072
|
+
set << c
|
1073
|
+
end
|
1074
|
+
|
1075
|
+
data = CSV.read(file, headers: true,
|
1076
|
+
header_converters: :symbol).map do |o|
|
1077
|
+
o = o.to_h.transform_keys(&kt)
|
1078
|
+
s = canonical_uuid(o.delete :id) or next
|
1079
|
+
|
1080
|
+
# LOLOL wtf
|
1081
|
+
|
1082
|
+
# handle audience
|
1083
|
+
[:audience, :nonaudience].each do |a|
|
1084
|
+
if o[a]
|
1085
|
+
o[a] = o[a].strip.split(/\s*[;,]+\s*/, -1).map do |t|
|
1086
|
+
if t =~ /^[a-z+-]+:[^[:space:]]+$/
|
1087
|
+
u = RDF::URI(t)
|
1088
|
+
canonical_uuid(u) || u
|
1089
|
+
elsif audiences[t.downcase]
|
1090
|
+
audiences[t.downcase].to_a
|
1091
|
+
end
|
1092
|
+
end.flatten.compact.uniq
|
1093
|
+
else
|
1094
|
+
o[a] = []
|
1095
|
+
end
|
1096
|
+
end
|
1097
|
+
|
1098
|
+
# handle concepts
|
1099
|
+
[:subject, :introduces, :assumes, :mentions].each do |a|
|
1100
|
+
if o[a]
|
1101
|
+
o[a] = o[a].strip.split(/\s*[;,]+\s*/, -1).map do |t|
|
1102
|
+
if t =~ /^[a-z+-]+:[^[:space:]]+$/
|
1103
|
+
u = RDF::URI(t)
|
1104
|
+
canonical_uuid(u) || u
|
1105
|
+
elsif concepts[t.downcase]
|
1106
|
+
concepts[t.downcase].to_a
|
1107
|
+
end
|
1108
|
+
end.flatten.compact.uniq
|
1109
|
+
else
|
1110
|
+
o[a] = []
|
1111
|
+
end
|
1112
|
+
|
1113
|
+
end
|
1114
|
+
|
1115
|
+
CSV_PRED.each do |sym, pred|
|
1116
|
+
o[sym].each do |obj|
|
1117
|
+
@graph << [s, pred, obj]
|
1118
|
+
end
|
1119
|
+
end
|
1120
|
+
|
1121
|
+
[s, o]
|
1122
|
+
end.compact.to_h
|
1123
|
+
data
|
1124
|
+
end
|
1125
|
+
|
1126
|
+
def generate_sitemap published: true
|
1127
|
+
urls = {}
|
1128
|
+
|
1129
|
+
# do feeds separately
|
1130
|
+
feeds = all_of_type RDF::Vocab::DCAT.Distribution
|
1131
|
+
#feeds.select! { |f| published? f } if published
|
1132
|
+
feeds.each do |f|
|
1133
|
+
uri = canonical_uri(f)
|
1134
|
+
f = generate_atom_feed f, published: published, related: feeds
|
1135
|
+
mt = f.at_xpath('/atom:feed/atom:updated[1]/text()',
|
1136
|
+
{ atom: 'http://www.w3.org/2005/Atom' })
|
1137
|
+
urls[uri] = { [{ [uri.to_s] => :loc }, { [mt] => :lastmod }] => :url }
|
1138
|
+
end
|
1139
|
+
|
1140
|
+
# build up hash of urls
|
1141
|
+
all_internal_docs(published: published).each do |doc|
|
1142
|
+
next if asserted_types(doc).include? RDF::Vocab::FOAF.Image
|
1143
|
+
uri = canonical_uri(doc)
|
1144
|
+
next unless uri.authority && @base && uri.authority == base.authority
|
1145
|
+
mods = objects_for(doc, [RDF::Vocab::DC.created,
|
1146
|
+
RDF::Vocab::DC.modified, RDF::Vocab::DC.issued],
|
1147
|
+
datatype: RDF::XSD.dateTime).sort
|
1148
|
+
nodes = [{ [uri.to_s] => :loc }]
|
1149
|
+
nodes << { [mods[-1].to_s] => :lastmod } unless mods.empty?
|
1150
|
+
urls[uri] = { nodes => :url }
|
1151
|
+
end
|
1152
|
+
|
1153
|
+
urls = urls.sort.map { |_, v| v }
|
1154
|
+
|
1155
|
+
markup(spec: { urls => :urlset,
|
1156
|
+
xmlns: 'http://www.sitemaps.org/schemas/sitemap/0.9' }).document
|
1157
|
+
end
|
1158
|
+
|
1159
|
+
def write_sitemap published: true
|
1160
|
+
sitemap = generate_sitemap published: published
|
1161
|
+
file = @config[:sitemap] || '.well-known/sitemap.xml'
|
1162
|
+
target = @config[published ? :target : :private]
|
1163
|
+
target.mkpath unless target.directory?
|
1164
|
+
|
1165
|
+
fh = (target + file).open(?w)
|
1166
|
+
sitemap.write_to fh
|
1167
|
+
fh.close
|
1168
|
+
end
|
1169
|
+
|
1170
|
+
# generate atom feed
|
1171
|
+
|
1172
|
+
#
|
1173
|
+
def all_internal_docs published: true, exclude: []
|
1174
|
+
# find all UUIDs that are documents
|
1175
|
+
docs = all_of_type(RDF::Vocab::FOAF.Document,
|
1176
|
+
exclude: exclude).select { |x| x =~ /^urn:uuid:/ }
|
1177
|
+
|
1178
|
+
# prune out all but the published documents if specified
|
1179
|
+
if published
|
1180
|
+
p = RDF::Vocab::BIBO.status
|
1181
|
+
o = RDF::Vocabulary.find_term(
|
1182
|
+
'http://purl.org/ontology/bibo/status/published')
|
1183
|
+
docs = docs.select do |s|
|
1184
|
+
@graph.has_statement? RDF::Statement(s, p, o)
|
1185
|
+
end
|
1186
|
+
end
|
1187
|
+
|
1188
|
+
docs
|
1189
|
+
end
|
1190
|
+
|
1191
|
+
def generate_atom_feed id, published: true, related: []
|
1192
|
+
raise 'ID must be a resource' unless id.is_a? RDF::Resource
|
1193
|
+
|
1194
|
+
# prepare relateds
|
1195
|
+
raise 'related must be an array' unless related.is_a? Array
|
1196
|
+
related -= [id]
|
1197
|
+
|
1198
|
+
# feed = struct_for id
|
1199
|
+
|
1200
|
+
faudy = audiences_for id
|
1201
|
+
faudn = audiences_for id, invert: true
|
1202
|
+
faudy -= faudn
|
1203
|
+
|
1204
|
+
docs = all_internal_docs published: published
|
1205
|
+
|
1206
|
+
# now we create a hash keyed by uuid containing the metadata
|
1207
|
+
authors = {}
|
1208
|
+
titles = {}
|
1209
|
+
dates = {}
|
1210
|
+
entries = {}
|
1211
|
+
latest = nil
|
1212
|
+
docs.each do |uu|
|
1213
|
+
# basically make a jsonld-like structure
|
1214
|
+
#rsrc = struct_for uu
|
1215
|
+
|
1216
|
+
indexed = objects_for uu, RDF::SAK::CI.indexed, only: :literal
|
1217
|
+
next if !indexed.empty? and indexed.any? { |f| f == false }
|
1218
|
+
|
1219
|
+
# get id (got it already duh)
|
1220
|
+
|
1221
|
+
# get audiences
|
1222
|
+
audy = audiences_for uu, proximate: true
|
1223
|
+
audn = audiences_for uu, proximate: true, invert: true
|
1224
|
+
|
1225
|
+
#warn "#{faudy.to_s} & #{faud"
|
1226
|
+
|
1227
|
+
skip = false
|
1228
|
+
if audy.empty?
|
1229
|
+
# an unspecified audience implies "everybody", but if the
|
1230
|
+
# feed's audience *is* specified, then it's not for everybody
|
1231
|
+
skip = true unless faudy.empty?
|
1232
|
+
else
|
1233
|
+
# if document audience matches feed non-audience, disqualify
|
1234
|
+
skip = true unless (faudn & audy).empty?
|
1235
|
+
|
1236
|
+
# absence of an explicit feed audience implies "everybody"
|
1237
|
+
if faudy.empty?
|
1238
|
+
# if document audience minus feed non-audience has
|
1239
|
+
# members, re-qualify
|
1240
|
+
skip = false unless (audy - faudn).empty?
|
1241
|
+
else
|
1242
|
+
# if document audience matches feed audience, re-qualify
|
1243
|
+
skip = false unless (faudy & audy).empty?
|
1244
|
+
end
|
1245
|
+
end
|
1246
|
+
|
1247
|
+
# if document non-audience matches feed audience, re-disqualify
|
1248
|
+
skip = true if !(audn.empty? || faudy.empty?) && !(faudy & audn).empty?
|
1249
|
+
|
1250
|
+
next if skip
|
1251
|
+
|
1252
|
+
canon = URI.parse(canonical_uri(uu).to_s)
|
1253
|
+
|
1254
|
+
xml = { '#entry' => [
|
1255
|
+
{ '#link' => nil, rel: :alternate, href: canon, type: 'text/html' },
|
1256
|
+
{ '#id' => uu.to_s }
|
1257
|
+
] }
|
1258
|
+
|
1259
|
+
# get published date first
|
1260
|
+
published = (objects_for uu,
|
1261
|
+
[RDF::Vocab::DC.issued, RDF::Vocab::DC.created],
|
1262
|
+
datatype: RDF::XSD.dateTime)[0]
|
1263
|
+
|
1264
|
+
# get latest updated date
|
1265
|
+
updated = (objects_for uu, RDF::Vocab::DC.modified,
|
1266
|
+
datatype: RDF::XSD.dateTime).sort[-1]
|
1267
|
+
updated ||= published || RDF::Literal::DateTime.new(DateTime.now)
|
1268
|
+
updated = Time.parse(updated.to_s).utc
|
1269
|
+
latest = updated if !latest or latest < updated
|
1270
|
+
|
1271
|
+
xml['#entry'].push({ '#updated' => updated.iso8601 })
|
1272
|
+
|
1273
|
+
if published
|
1274
|
+
published = Time.parse(published.to_s).utc
|
1275
|
+
xml['#entry'].push({ '#published' => published.iso8601 })
|
1276
|
+
dates[uu] = [published, updated]
|
1277
|
+
else
|
1278
|
+
dates[uu] = [updated, updated]
|
1279
|
+
end
|
1280
|
+
|
1281
|
+
# get author(s)
|
1282
|
+
al = []
|
1283
|
+
authors_for(uu).each do |a|
|
1284
|
+
unless authors[a]
|
1285
|
+
n = label_for a
|
1286
|
+
x = authors[a] = { '#author' => [{ '#name' => n[1].to_s }] }
|
1287
|
+
|
1288
|
+
hp = @graph.first_object [a, RDF::Vocab::FOAF.homepage, nil]
|
1289
|
+
hp ||= canonical_uri a
|
1290
|
+
|
1291
|
+
x['#author'].push({ '#uri' => hp.to_s }) if hp
|
1292
|
+
end
|
1293
|
+
|
1294
|
+
al.push authors[a]
|
1295
|
+
end
|
1296
|
+
|
1297
|
+
xml['#entry'] += al unless al.empty?
|
1298
|
+
|
1299
|
+
# get title (note unshift)
|
1300
|
+
if (t = label_for uu)
|
1301
|
+
titles[uu] = t[1].to_s
|
1302
|
+
xml['#entry'].unshift({ '#title' => t[1].to_s })
|
1303
|
+
else
|
1304
|
+
titles[uu] = uu.to_s
|
1305
|
+
end
|
1306
|
+
|
1307
|
+
# get abstract
|
1308
|
+
if (d = label_for uu, desc: true)
|
1309
|
+
xml['#entry'].push({ '#summary' => d[1].to_s })
|
1310
|
+
end
|
1311
|
+
|
1312
|
+
entries[uu] = xml
|
1313
|
+
end
|
1314
|
+
|
1315
|
+
# note we overwrite the entries hash here with a sorted array
|
1316
|
+
entrycmp = -> a, b {
|
1317
|
+
# first we sort by published date
|
1318
|
+
p = dates[a][0] <=> dates[b][0]
|
1319
|
+
# if the published dates are the same, sort by updated date
|
1320
|
+
u = dates[a][1] <=> dates[b][1]
|
1321
|
+
# to break any ties, finally sort by title
|
1322
|
+
p == 0 ? u == 0 ? titles[a] <=> titles[b] : u : p }
|
1323
|
+
entries = entries.values_at(
|
1324
|
+
*entries.keys.sort { |a, b| entrycmp.call(a, b) })
|
1325
|
+
# ugggh god forgot the asterisk and lost an hour
|
1326
|
+
|
1327
|
+
# now we punt out the doc
|
1328
|
+
|
1329
|
+
preamble = [
|
1330
|
+
{ '#id' => id.to_s },
|
1331
|
+
{ '#updated' => latest.iso8601 },
|
1332
|
+
{ '#generator' => 'RDF::SAK', version: RDF::SAK::VERSION,
|
1333
|
+
uri: "https://github.com/doriantaylor/rb-rdf-sak" },
|
1334
|
+
{ nil => :link, rel: :self, type: 'application/atom+xml',
|
1335
|
+
href: canonical_uri(id) },
|
1336
|
+
{ nil => :link, rel: :alternate, type: 'text/html',
|
1337
|
+
href: @base },
|
1338
|
+
] + related.map do |r|
|
1339
|
+
{ nil => :link, rel: :related, type: 'application/atom+xml',
|
1340
|
+
href: canonical_uri(r) }
|
1341
|
+
end
|
1342
|
+
|
1343
|
+
if (t = label_for id)
|
1344
|
+
preamble.unshift({ '#title' => t[1].to_s })
|
1345
|
+
end
|
1346
|
+
|
1347
|
+
if (r = @graph.first_literal [id, RDF::Vocab::DC.rights, nil])
|
1348
|
+
rh = { '#rights' => r.to_s, type: :text }
|
1349
|
+
rh['xml:lang'] = r.language if r.has_language?
|
1350
|
+
preamble.push rh
|
1351
|
+
end
|
1352
|
+
|
1353
|
+
markup(spec: { '#feed' => preamble + entries,
|
1354
|
+
xmlns: 'http://www.w3.org/2005/Atom' }).document
|
1355
|
+
end
|
1356
|
+
|
1357
|
+
def write_feeds type: RDF::Vocab::DCAT.Distribution, published: true
|
1358
|
+
feeds = all_of_type type
|
1359
|
+
target = @config[published ? :target : :private]
|
1360
|
+
feeds.each do |feed|
|
1361
|
+
tu = URI(feed.to_s)
|
1362
|
+
doc = generate_atom_feed feed, published: published, related: feeds
|
1363
|
+
fh = (target + "#{tu.uuid}.xml").open('w')
|
1364
|
+
doc.write_to fh
|
1365
|
+
fh.close
|
1366
|
+
end
|
1367
|
+
end
|
1368
|
+
|
1369
|
+
# generate sass palettes
|
1370
|
+
|
1371
|
+
# generate rewrite map(s)
|
1372
|
+
def generate_rewrite_map published: false, docs: nil
|
1373
|
+
docs ||= reachable published: published
|
1374
|
+
base = URI(@base.to_s)
|
1375
|
+
rwm = {}
|
1376
|
+
docs.each do |doc|
|
1377
|
+
tu = URI(doc.to_s)
|
1378
|
+
cu = canonical_uri doc, rdf: false
|
1379
|
+
next unless tu.respond_to?(:uuid) and cu.respond_to?(:request_uri)
|
1380
|
+
|
1381
|
+
# skip external links obvs
|
1382
|
+
next unless base.route_to(cu).relative?
|
1383
|
+
|
1384
|
+
# skip /uuid form
|
1385
|
+
cp = cu.request_uri.delete_prefix '/'
|
1386
|
+
next if cu.host == base.host and tu.uuid == cp
|
1387
|
+
|
1388
|
+
rwm[cp] = tu.uuid
|
1389
|
+
end
|
1390
|
+
|
1391
|
+
rwm
|
1392
|
+
end
|
1393
|
+
|
1394
|
+
# give me all UUIDs of all documents, filter for published if
|
1395
|
+
# applicable
|
1396
|
+
#
|
1397
|
+
# find the "best" (relative) URL for the UUID and map the pair
|
1398
|
+
# together
|
1399
|
+
def generate_uuid_redirect_map published: false, docs: nil
|
1400
|
+
docs ||= reachable published: published
|
1401
|
+
|
1402
|
+
base = URI(@base.to_s)
|
1403
|
+
|
1404
|
+
# keys are /uuid, values are
|
1405
|
+
out = {}
|
1406
|
+
docs.each do |doc|
|
1407
|
+
tu = URI(doc.to_s)
|
1408
|
+
cu = canonical_uri doc, rdf: false
|
1409
|
+
next unless tu.respond_to?(:uuid) and cu.respond_to?(:request_uri)
|
1410
|
+
|
1411
|
+
# skip /uuid form
|
1412
|
+
cp = cu.request_uri.delete_prefix '/'
|
1413
|
+
next if cu.host == base.host && tu.uuid == cp
|
1414
|
+
|
1415
|
+
# all redirect links are absolute
|
1416
|
+
out[tu.uuid] = cu.to_s
|
1417
|
+
end
|
1418
|
+
out
|
1419
|
+
end
|
1420
|
+
|
1421
|
+
# find all URIs/slugs that are *not* canonical, map them to slugs
|
1422
|
+
# that *are* canonical
|
1423
|
+
def generate_slug_redirect_map published: false, docs: nil
|
1424
|
+
docs ||= reachable published: published
|
1425
|
+
base = URI(@base.to_s)
|
1426
|
+
|
1427
|
+
# for redirects we collect all the docs, plus all their URIs,
|
1428
|
+
# separate canonical from the rest
|
1429
|
+
|
1430
|
+
# actually an easy way to do this is just harvest all the
|
1431
|
+
# multi-addressed docs, remove the first one, then ask for the
|
1432
|
+
# canonical uuid back,
|
1433
|
+
|
1434
|
+
fwd = {}
|
1435
|
+
rev = {}
|
1436
|
+
out = {}
|
1437
|
+
|
1438
|
+
docs.each do |doc|
|
1439
|
+
uris = canonical_uri doc, unique: false, rdf: false
|
1440
|
+
canon = uris.shift
|
1441
|
+
next unless canon.respond_to? :request_uri
|
1442
|
+
|
1443
|
+
# cache the forward direction
|
1444
|
+
fwd[doc] = canon
|
1445
|
+
|
1446
|
+
unless uris.empty?
|
1447
|
+
uris.each do |uri|
|
1448
|
+
next unless uri.respond_to? :request_uri
|
1449
|
+
next if canon == uri
|
1450
|
+
next unless base.route_to(uri).relative?
|
1451
|
+
|
1452
|
+
# warn "#{canon} <=> #{uri}"
|
1453
|
+
|
1454
|
+
requri = uri.request_uri.delete_prefix '/'
|
1455
|
+
next if requri == '' ||
|
1456
|
+
requri =~ /^[0-9a-f]{8}(?:-[0-9a-f]{4}){4}[0-9a-f]{8}$/
|
1457
|
+
|
1458
|
+
# cache the reverse direction
|
1459
|
+
rev[uri] = requri
|
1460
|
+
end
|
1461
|
+
end
|
1462
|
+
end
|
1463
|
+
|
1464
|
+
rev.each do |uri, requri|
|
1465
|
+
if (doc = canonical_uuid(uri, published: published)) and
|
1466
|
+
fwd[doc] and fwd[doc] != uri
|
1467
|
+
out[requri] = fwd[doc].to_s
|
1468
|
+
end
|
1469
|
+
end
|
1470
|
+
|
1471
|
+
out
|
1472
|
+
end
|
1473
|
+
|
1474
|
+
# you know what, it's entirely possible that these ought never be
|
1475
|
+
# called individually and the work to get one would duplicate the
|
1476
|
+
# work of getting the other, so maybe just do 'em both at once
|
1477
|
+
|
1478
|
+
def generate_redirect_map published: false, docs: nil
|
1479
|
+
generate_uuid_redirect_map(published: published, docs: docs).merge(
|
1480
|
+
generate_slug_redirect_map(published: published, docs: docs))
|
1481
|
+
end
|
1482
|
+
|
1483
|
+
def generate_gone_map published: false, docs: nil
|
1484
|
+
# published is a no-op for this one because these docs are by
|
1485
|
+
# definition not published
|
1486
|
+
docs ||= reachable published: false
|
1487
|
+
p = RDF::Vocab::BIBO.status
|
1488
|
+
base = URI(@base.to_s)
|
1489
|
+
out = {}
|
1490
|
+
docs.select { |s|
|
1491
|
+
@graph.has_statement? RDF::Statement(s, p, CI.retired) }.each do |doc|
|
1492
|
+
canon = canonical_uri doc, rdf: false
|
1493
|
+
next unless base.route_to(canon).relative?
|
1494
|
+
canon = canon.request_uri.delete_prefix '/'
|
1495
|
+
# value of the gone map doesn't matter
|
1496
|
+
out[canon] = canon
|
1497
|
+
end
|
1498
|
+
|
1499
|
+
out
|
1500
|
+
end
|
1501
|
+
|
1502
|
+
# private?
|
1503
|
+
|
1504
|
+
def map_location type
|
1505
|
+
# find file name in config
|
1506
|
+
fn = @config[:maps][type] or return
|
1507
|
+
|
1508
|
+
# concatenate to target directory
|
1509
|
+
@config[:target] + fn
|
1510
|
+
end
|
1511
|
+
|
1512
|
+
# private?
|
1513
|
+
|
1514
|
+
def write_map_file location, data
|
1515
|
+
# open file
|
1516
|
+
fh = File.new location, 'w'
|
1517
|
+
data.sort.each { |k, v| fh.write "#{k}\t#{v}\n" }
|
1518
|
+
fh.close # return value is return value from close
|
1519
|
+
end
|
1520
|
+
|
1521
|
+
# public again
|
1522
|
+
|
1523
|
+
def write_rewrite_map published: false, docs: nil
|
1524
|
+
data = generate_rewrite_map published: published, docs: docs
|
1525
|
+
loc = map_location :rewrite
|
1526
|
+
write_map_file loc, data
|
1527
|
+
end
|
1528
|
+
|
1529
|
+
def write_redirect_map published: false, docs: nil
|
1530
|
+
data = generate_redirect_map published: published, docs: docs
|
1531
|
+
loc = map_location :redirect
|
1532
|
+
write_map_file loc, data
|
1533
|
+
end
|
1534
|
+
|
1535
|
+
def write_gone_map published: false, docs: nil
|
1536
|
+
data = generate_gone_map published: published, docs: docs
|
1537
|
+
loc = map_location :gone
|
1538
|
+
write_map_file loc, data
|
1539
|
+
end
|
1540
|
+
|
1541
|
+
def write_maps published: true, docs: nil
|
1542
|
+
docs ||= reachable published: false
|
1543
|
+
# slug to uuid (internal)
|
1544
|
+
write_rewrite_map docs: docs
|
1545
|
+
# uuid/slug to canonical slug (308)
|
1546
|
+
write_redirect_map docs: docs
|
1547
|
+
# retired slugs/uuids (410)
|
1548
|
+
write_gone_map docs: docs
|
1549
|
+
true
|
1550
|
+
end
|
1551
|
+
|
1552
|
+
# whoops lol we forgot the book list
|
1553
|
+
|
1554
|
+
def reading_lists published: true
|
1555
|
+
out = all_of_type RDF::Vocab::SiocTypes.ReadingList
|
1556
|
+
return out unless published
|
1557
|
+
out.select { |r| published? r }
|
1558
|
+
end
|
1559
|
+
|
1560
|
+
def generate_reading_list subject, published: true
|
1561
|
+
# struct = struct_for subject
|
1562
|
+
|
1563
|
+
# find all the books, sort them by title
|
1564
|
+
|
1565
|
+
# for each book, give title, authors, inbound references
|
1566
|
+
|
1567
|
+
# punt out xhtml
|
1568
|
+
end
|
1569
|
+
|
1570
|
+
def write_reading_lists published: true
|
1571
|
+
reading_lists(published: published).each do |rl|
|
1572
|
+
tu = URI(rl.to_s)
|
1573
|
+
doc = generate_reading_list rl, published: published
|
1574
|
+
fh = (target + "#{tu.uuid}.xml").open('w')
|
1575
|
+
doc.write_to fh
|
1576
|
+
fh.close
|
1577
|
+
end
|
1578
|
+
end
|
1579
|
+
|
1580
|
+
DSD_SEQ = %i[characters words blocks sections
|
1581
|
+
min low-quartile median high-quartile max mean sd].freeze
|
1582
|
+
TH_SEQ = %w[Document Abstract Created Modified Characters Words Blocks
|
1583
|
+
Sections Min Q1 Median Q3 Max Mean SD].map { |t| { [t] => :th } }
|
1584
|
+
|
1585
|
+
def generate_stats published: true
|
1586
|
+
out = {}
|
1587
|
+
all_of_type(QB.DataSet).map do |s|
|
1588
|
+
base = canonical_uri s, rdf: false
|
1589
|
+
types = abbreviate asserted_types(s)
|
1590
|
+
title = if t = label_for(s)
|
1591
|
+
[t[1].to_s, abbreviate(t[0])]
|
1592
|
+
end
|
1593
|
+
cache = {}
|
1594
|
+
subjects_for(QB.dataSet, s, only: :resource).each do |o|
|
1595
|
+
if d = objects_for(o, CI.document, only: :resource).first
|
1596
|
+
if !published or published?(d)
|
1597
|
+
# include a "sort" time that defaults to epoch zero
|
1598
|
+
c = cache[o] ||= {
|
1599
|
+
doc: d, stime: Time.at(0).getgm, struct: struct_for(o) }
|
1600
|
+
|
1601
|
+
if t = label_for(d)
|
1602
|
+
c[:title] = t
|
1603
|
+
end
|
1604
|
+
if a = label_for(d, desc: true)
|
1605
|
+
c[:abstract] = a
|
1606
|
+
end
|
1607
|
+
if ct = objects_for(d,
|
1608
|
+
RDF::Vocab::DC.created, datatype: RDF::XSD.dateTime).first
|
1609
|
+
c[:stime] = c[:ctime] = ct.object.to_time.getgm
|
1610
|
+
end
|
1611
|
+
if mt = objects_for(d,
|
1612
|
+
RDF::Vocab::DC.modified, datatype:RDF::XSD.dateTime)
|
1613
|
+
c[:mtime] = mt.map { |m| m.object.to_time.getgm }.sort
|
1614
|
+
c[:stime] = c[:mtime].last unless mt.empty?
|
1615
|
+
end
|
1616
|
+
end
|
1617
|
+
end
|
1618
|
+
end
|
1619
|
+
|
1620
|
+
# sort lambda closure
|
1621
|
+
sl = -> a, b do
|
1622
|
+
x = cache[b][:stime] <=> cache[a][:stime]
|
1623
|
+
return x unless x == 0
|
1624
|
+
x = cache[b][:ctime] <=> cache[a][:ctime]
|
1625
|
+
return x unless x == 0
|
1626
|
+
ta = cache[a][:title] || Array.new(2, cache[a][:uri])
|
1627
|
+
tb = cache[b][:title] || Array.new(2, cache[b][:uri])
|
1628
|
+
ta[1].to_s <=> tb[1].to_s
|
1629
|
+
end
|
1630
|
+
|
1631
|
+
rows = []
|
1632
|
+
cache.keys.sort(&sl).each do |k|
|
1633
|
+
c = cache[k]
|
1634
|
+
href = base.route_to canonical_uri(c[:doc], rdf: false)
|
1635
|
+
dt = abbreviate asserted_types(c[:doc])
|
1636
|
+
uu = URI(k.to_s).uuid
|
1637
|
+
nc = UUID::NCName.to_ncname uu, version: 1
|
1638
|
+
tp, tt = c[:title] || []
|
1639
|
+
ab = if c[:abstract]
|
1640
|
+
{ [c[:abstract][1].to_s] => :th, about: href,
|
1641
|
+
property: abbreviate(c[:abstract].first) }
|
1642
|
+
else
|
1643
|
+
{ [] => :th }
|
1644
|
+
end
|
1645
|
+
|
1646
|
+
td = [{ { { [tt.to_s] => :span, property: abbreviate(tp) } => :a,
|
1647
|
+
rel: 'ci:document', href: href } => :th },
|
1648
|
+
ab,
|
1649
|
+
{ [c[:ctime].iso8601] => :th, property: 'dct:created',
|
1650
|
+
datatype: 'xsd:dateTime', about: href, typeof: dt },
|
1651
|
+
{ c[:mtime].reverse.map { |m| { [m.iso8601] => :span,
|
1652
|
+
property: 'dct:modified', datatype: 'xsd:dateTime' } } => :th,
|
1653
|
+
about: href
|
1654
|
+
},
|
1655
|
+
] + DSD_SEQ.map do |f|
|
1656
|
+
h = []
|
1657
|
+
x = { h => :td }
|
1658
|
+
p = CI[f]
|
1659
|
+
if y = c[:struct][p] and !y.empty?
|
1660
|
+
h << y = y.first
|
1661
|
+
x[:property] = abbreviate p
|
1662
|
+
x[:datatype] = abbreviate y.datatype if y.datatype?
|
1663
|
+
end
|
1664
|
+
x
|
1665
|
+
end
|
1666
|
+
rows << { td => :tr, id: nc, about: "##{nc}",
|
1667
|
+
typeof: 'qb:Observation' }
|
1668
|
+
end
|
1669
|
+
|
1670
|
+
out[s] = xhtml_stub(base: base, title: title,
|
1671
|
+
transform: config[:transform], attr: { about: '', typeof: types },
|
1672
|
+
prefix: prefixes, content: {
|
1673
|
+
[{ [{ [{ ['About'] => :th, colspan: 4 },
|
1674
|
+
{ ['Counts'] => :th, colspan: 4 },
|
1675
|
+
{ ['Words per Block'] => :th, colspan: 7 }] => :tr },
|
1676
|
+
{ TH_SEQ => :tr } ] => :thead },
|
1677
|
+
{ rows => :tbody, rev: 'qb:dataSet' }] => :table }).document
|
1678
|
+
end
|
1679
|
+
|
1680
|
+
out
|
1681
|
+
end
|
1682
|
+
|
1683
|
+
def write_stats published: true
|
1684
|
+
target = @config[published ? :target : :private]
|
1685
|
+
target.mkpath unless target.directory?
|
1686
|
+
generate_stats(published: published).each do |uu, doc|
|
1687
|
+
bn = URI(uu.to_s).uuid + '.xml'
|
1688
|
+
fh = (target + bn).open ?w
|
1689
|
+
doc.write_to fh
|
1690
|
+
fh.flush
|
1691
|
+
fh.close
|
1692
|
+
end
|
1693
|
+
end
|
1694
|
+
|
1695
|
+
# - io stuff -
|
1696
|
+
|
1697
|
+
# Locate the file in the source directory associated with the given URI.
|
1698
|
+
#
|
1699
|
+
# @param [RDF::URI, URI, :to_s] the URI requested
|
1700
|
+
#
|
1701
|
+
# @return [Pathname] of the corresponding file or nil if no file was found.
|
1702
|
+
|
1703
|
+
def locate uri
|
1704
|
+
uri = coerce_resource uri
|
1705
|
+
|
1706
|
+
base = URI(@base.to_s)
|
1707
|
+
|
1708
|
+
tu = URI(uri) # copy of uri for testing content
|
1709
|
+
unless tu.scheme == 'urn' and tu.nid == 'uuid'
|
1710
|
+
raise "could not find UUID for #{uri}" unless uuid = canonical_uuid(uri)
|
1711
|
+
tu = URI(uri = uuid)
|
1712
|
+
end
|
1713
|
+
|
1714
|
+
# xxx bail if the uri isn't a subject in the graph
|
1715
|
+
|
1716
|
+
candidates = [@config[:source] + tu.uuid]
|
1717
|
+
|
1718
|
+
# try all canonical URIs
|
1719
|
+
(canonical_uri uri, unique: false, slugs: true).each do |u|
|
1720
|
+
u = URI(u.to_s)
|
1721
|
+
next unless u.hostname == base.hostname
|
1722
|
+
p = URI.unescape u.path[/^\/*(.*?)$/, 1]
|
1723
|
+
candidates.push(@config[:source] + p)
|
1724
|
+
end
|
1725
|
+
|
1726
|
+
# warn candidates
|
1727
|
+
|
1728
|
+
files = candidates.uniq.map do |c|
|
1729
|
+
Pathname.glob(c.to_s + '{,.*,/index{,.*}}')
|
1730
|
+
end.reduce(:+).reject do |x|
|
1731
|
+
x.directory? or RDF::SAK::MimeMagic.by_path(x).to_s !~
|
1732
|
+
/.*(?:markdown|(?:x?ht|x)ml).*/i
|
1733
|
+
end.uniq
|
1734
|
+
|
1735
|
+
#warn files
|
1736
|
+
|
1737
|
+
# XXX implement negotiation algorithm
|
1738
|
+
return files[0]
|
1739
|
+
|
1740
|
+
# return the filename from the source
|
1741
|
+
# nil
|
1742
|
+
end
|
1743
|
+
|
1744
|
+
# Visit (open) the document at the given URI.
|
1745
|
+
#
|
1746
|
+
# @param uri [RDF::URI, URI, :to_s]
|
1747
|
+
#
|
1748
|
+
# @return [RDF::SAK::Context::Document] or nil
|
1749
|
+
|
1750
|
+
def visit uri
|
1751
|
+
uri = canonical_uuid uri
|
1752
|
+
path = locate uri
|
1753
|
+
return unless path
|
1754
|
+
Document.new self, uri, uri: canonical_uri(uri), doc: path
|
1755
|
+
end
|
1756
|
+
|
1757
|
+
# resolve documents from source
|
1758
|
+
def resolve_documents
|
1759
|
+
src = @config[:source]
|
1760
|
+
out = []
|
1761
|
+
src.find do |f|
|
1762
|
+
Find.prune if f.basename.to_s[0] == ?.
|
1763
|
+
next if f.directory?
|
1764
|
+
out << f
|
1765
|
+
end
|
1766
|
+
|
1767
|
+
out
|
1768
|
+
end
|
1769
|
+
|
1770
|
+
def resolve_file path
|
1771
|
+
return unless path.file?
|
1772
|
+
path = Pathname('/') + path.relative_path_from(@config[:source])
|
1773
|
+
base = URI(@base.to_s)
|
1774
|
+
uri = base + path.to_s
|
1775
|
+
|
1776
|
+
#warn "trying #{uri}"
|
1777
|
+
|
1778
|
+
until (out = canonical_uuid uri)
|
1779
|
+
# iteratively strip off
|
1780
|
+
break if uri.path.end_with? '/'
|
1781
|
+
|
1782
|
+
dn = path.dirname
|
1783
|
+
bn = path.basename '.*'
|
1784
|
+
|
1785
|
+
# try index first
|
1786
|
+
if bn.to_s == 'index'
|
1787
|
+
p = dn.to_s
|
1788
|
+
p << '/' unless p.end_with? '/'
|
1789
|
+
uri = base + p
|
1790
|
+
elsif bn == path.basename
|
1791
|
+
break
|
1792
|
+
else
|
1793
|
+
path = dn + bn
|
1794
|
+
uri = base + path.to_s
|
1795
|
+
end
|
1796
|
+
|
1797
|
+
# warn "trying #{uri}"
|
1798
|
+
end
|
1799
|
+
|
1800
|
+
out
|
1801
|
+
end
|
1802
|
+
|
1803
|
+
# Determine whether the URI represents a published document.
|
1804
|
+
#
|
1805
|
+
# @param uri
|
1806
|
+
#
|
1807
|
+
# @return [true, false]
|
1808
|
+
def published? uri, circulated: false
|
1809
|
+
RDF::SAK::Util.published? @graph, uri,
|
1810
|
+
circulated: circulated, base: @base
|
1811
|
+
end
|
1812
|
+
|
1813
|
+
# Find a destination pathname for the document
|
1814
|
+
#
|
1815
|
+
# @param uri
|
1816
|
+
# @param published
|
1817
|
+
#
|
1818
|
+
# @return [Pathname]
|
1819
|
+
def target_for uri, published: false
|
1820
|
+
uri = coerce_resource uri
|
1821
|
+
uri = canonical_uuid uri
|
1822
|
+
target = @config[published?(uri) && published ? :target : :private]
|
1823
|
+
|
1824
|
+
# target is a pathname so this makes a pathname
|
1825
|
+
target + "#{URI(uri.to_s).uuid}.xml"
|
1826
|
+
end
|
1827
|
+
|
1828
|
+
# read from source
|
1829
|
+
|
1830
|
+
# write (manipulated (x|x?ht)ml) back to source
|
1831
|
+
|
1832
|
+
# write public and private variants to target
|
1833
|
+
|
1834
|
+
def write_xhtml published: true
|
1835
|
+
end
|
1836
|
+
|
1837
|
+
# write modified rdf
|
1838
|
+
|
1839
|
+
# - internet stuff -
|
1840
|
+
|
1841
|
+
# verify external links for upness
|
1842
|
+
|
1843
|
+
# collect triples for external links
|
1844
|
+
|
1845
|
+
# fetch references for people/companies/concepts/etc from dbpedia/wikidata
|
1846
|
+
|
1847
|
+
# - document context class -
|
1848
|
+
|
1849
|
+
class Document
|
1850
|
+
include XML::Mixup
|
1851
|
+
include Util
|
1852
|
+
|
1853
|
+
private
|
1854
|
+
|
1855
|
+
C_OK = [Nokogiri::XML::Node, IO, Pathname].freeze
|
1856
|
+
|
1857
|
+
public
|
1858
|
+
|
1859
|
+
attr_reader :doc, :uuid, :uri
|
1860
|
+
|
1861
|
+
def initialize context, uuid, doc: nil, uri: nil, mtime: nil
|
1862
|
+
raise 'context must be a RDF::SAK::Context' unless
|
1863
|
+
context.is_a? RDF::SAK::Context
|
1864
|
+
raise 'uuid must be an RDF::URI' unless
|
1865
|
+
uuid.is_a? RDF::URI and uuid.to_s.start_with? 'urn:uuid:'
|
1866
|
+
|
1867
|
+
doc ||= context.locate uuid
|
1868
|
+
raise 'doc must be Pathname, IO, or Nokogiri node' unless
|
1869
|
+
C_OK.any? { |c| doc.is_a? c } || doc.respond_to?(:to_s)
|
1870
|
+
|
1871
|
+
# set some instance variables
|
1872
|
+
@context = context
|
1873
|
+
@uuid = uuid
|
1874
|
+
@mtime = mtime || doc.respond_to?(:mtime) ? doc.mtime : Time.now
|
1875
|
+
@target = context.target_for uuid
|
1876
|
+
|
1877
|
+
# now process the document
|
1878
|
+
|
1879
|
+
# turn the document into an XML::Document
|
1880
|
+
if doc.is_a? Nokogiri::XML::Node
|
1881
|
+
# a node that is not a document should be wrapped with one
|
1882
|
+
unless doc.is_a? Nokogiri::XML::Document
|
1883
|
+
d = doc.dup 1
|
1884
|
+
doc = Nokogiri::XML::Document.new
|
1885
|
+
doc << d
|
1886
|
+
end
|
1887
|
+
else
|
1888
|
+
type = nil
|
1889
|
+
|
1890
|
+
# pathnames turned into IO objects
|
1891
|
+
if doc.is_a? Pathname
|
1892
|
+
type = RDF::SAK::MimeMagic.by_path doc
|
1893
|
+
doc = doc.open # this may raise if the file isn't there
|
1894
|
+
end
|
1895
|
+
|
1896
|
+
# squash everything else to a string
|
1897
|
+
doc = doc.to_s unless doc.is_a? IO
|
1898
|
+
|
1899
|
+
# check type by content
|
1900
|
+
type ||= RDF::SAK::MimeMagic.by_magic(doc)
|
1901
|
+
|
1902
|
+
# can you believe there is a special bookmarks mime type good grief
|
1903
|
+
type = 'text/html' if type == 'application/x-mozilla-bookmarks'
|
1904
|
+
|
1905
|
+
# now we try to parse the blob
|
1906
|
+
if type.to_s =~ /xml/i
|
1907
|
+
doc = Nokogiri.XML doc
|
1908
|
+
elsif type == 'text/html'
|
1909
|
+
# if the detected type is html, try it as strict xml first
|
1910
|
+
attempt = nil
|
1911
|
+
begin
|
1912
|
+
attempt = Nokogiri.XML doc, nil, nil, (1 << 11) # NONET
|
1913
|
+
rescue Nokogiri::XML::SyntaxError
|
1914
|
+
# do not wrap this a second time; let it fail if it's gonna
|
1915
|
+
tmp = Nokogiri.HTML doc
|
1916
|
+
attempt = Nokogiri::XML::Document.new
|
1917
|
+
attempt << tmp.root.dup(1)
|
1918
|
+
end
|
1919
|
+
doc = attempt
|
1920
|
+
elsif type.to_s =~ /^text\/(?:plain|(?:x-)?markdown)/i
|
1921
|
+
# just assume plain text is markdown
|
1922
|
+
doc = ::MD::Noko.new.ingest doc
|
1923
|
+
else
|
1924
|
+
raise "Don't know what to do with #{uuid} (#{type})"
|
1925
|
+
end
|
1926
|
+
end
|
1927
|
+
|
1928
|
+
# now fix the namespaces for mangled html documents
|
1929
|
+
root = doc.root
|
1930
|
+
if root.name == 'html'
|
1931
|
+
unless root.namespace
|
1932
|
+
# clear this off or it will be duplicated in the output
|
1933
|
+
root.remove_attribute('xmlns')
|
1934
|
+
# now generate a new ns object
|
1935
|
+
ns = root.add_namespace(nil, XHTMLNS)
|
1936
|
+
# *now* scan the document and add the namespace declaration
|
1937
|
+
root.traverse do |node|
|
1938
|
+
if node.element? && node.namespace.nil?
|
1939
|
+
# downcasing the name may be cargo culting; need to check
|
1940
|
+
# node.name = node.name.downcase # yup it is
|
1941
|
+
node.namespace = ns
|
1942
|
+
end
|
1943
|
+
end
|
1944
|
+
end
|
1945
|
+
|
1946
|
+
# also add the magic blank doctype declaration if it's missing
|
1947
|
+
unless doc.internal_subset
|
1948
|
+
doc.create_internal_subset('html', nil, nil)
|
1949
|
+
end
|
1950
|
+
end
|
1951
|
+
|
1952
|
+
# aaand set some more instance variables
|
1953
|
+
|
1954
|
+
@uri = URI(uri || @context.canonical_uri(uuid))
|
1955
|
+
|
1956
|
+
# voilà
|
1957
|
+
@doc = doc
|
1958
|
+
end
|
1959
|
+
|
1960
|
+
# proxy for context published
|
1961
|
+
def published?
|
1962
|
+
@context.published? @uuid
|
1963
|
+
end
|
1964
|
+
|
1965
|
+
def base_for node = nil
|
1966
|
+
node ||= @doc
|
1967
|
+
doc = node.document
|
1968
|
+
base = @uri.to_s
|
1969
|
+
if doc.root.name.to_sym == :html
|
1970
|
+
b = doc.at_xpath(
|
1971
|
+
'(/html:html/html:head/html:base[@href])[1]/@href',
|
1972
|
+
{ html: XHTMLNS }).to_s.strip
|
1973
|
+
base = b if URI(b).absolute?
|
1974
|
+
elsif b = doc.at_xpath('ancestor-or-self::*[@xml:base][1]/@xml:base')
|
1975
|
+
b = b.to_s.strip
|
1976
|
+
base = b if URI(b).absolute?
|
1977
|
+
end
|
1978
|
+
|
1979
|
+
URI(base)
|
1980
|
+
end
|
1981
|
+
|
1982
|
+
# notice these are only RDFa attributes that take URIs
|
1983
|
+
RDFA_ATTR = [:about, :resource, :typeof].freeze
|
1984
|
+
LINK_ATTR = [:href, :src, :data, :action, :longdesc].freeze
|
1985
|
+
LINK_XPATH = ('.//html:*[not(self::html:base)][%s]' %
|
1986
|
+
(LINK_ATTR + RDFA_ATTR).map { |a| "@#{a.to_s}" }.join('|')).freeze
|
1987
|
+
|
1988
|
+
def rewrite_links node = @doc, uuids: {}, uris: {}, &block
|
1989
|
+
base = base_for node
|
1990
|
+
count = 0
|
1991
|
+
cache = {}
|
1992
|
+
node.xpath(LINK_XPATH, { html: XHTMLNS }).each do |elem|
|
1993
|
+
LINK_ATTR.each do |attr|
|
1994
|
+
attr = attr.to_s
|
1995
|
+
next unless elem.has_attribute? attr
|
1996
|
+
|
1997
|
+
abs = base.merge uri_pp(elem[attr].strip)
|
1998
|
+
|
1999
|
+
# fix e.g. http->https
|
2000
|
+
if abs.host == @uri.host and abs.scheme != @uri.scheme
|
2001
|
+
tmp = @uri.dup
|
2002
|
+
tmp.path = abs.path
|
2003
|
+
tmp.query = abs.query
|
2004
|
+
tmp.fragment = abs.fragment
|
2005
|
+
abs = tmp
|
2006
|
+
end
|
2007
|
+
|
2008
|
+
# harvest query string
|
2009
|
+
pp = split_pp abs, only: true
|
2010
|
+
|
2011
|
+
abs = RDF::URI(abs.to_s)
|
2012
|
+
|
2013
|
+
# round-trip to uuid and back if we can
|
2014
|
+
if uuid = uuids[abs] ||= @context.canonical_uuid(abs)
|
2015
|
+
abs = cache[abs] ||= @context.canonical_uri(uuid)
|
2016
|
+
else
|
2017
|
+
abs = cache[abs] ||= @context.canonical_uri(abs)
|
2018
|
+
end
|
2019
|
+
|
2020
|
+
# reinstate the path parameters
|
2021
|
+
if !pp.empty? && split_pp(abs, only: true).empty?
|
2022
|
+
abs = abs.dup
|
2023
|
+
abs.path = ([abs.path] + pp).join(';')
|
2024
|
+
end
|
2025
|
+
|
2026
|
+
|
2027
|
+
elem[attr] = @uri.route_to(abs.to_s).to_s
|
2028
|
+
count += 1
|
2029
|
+
end
|
2030
|
+
|
2031
|
+
block.call elem if block
|
2032
|
+
end
|
2033
|
+
|
2034
|
+
count
|
2035
|
+
end
|
2036
|
+
|
2037
|
+
# sponge the document for rdfa
|
2038
|
+
def triples_for
|
2039
|
+
end
|
2040
|
+
|
2041
|
+
OBJS = [:href, :src].freeze
|
2042
|
+
|
2043
|
+
# ancestor node always with (@property and not @content) and
|
2044
|
+
# not @resource|@href|@src unless @rel|@rev
|
2045
|
+
LITXP = ['(ancestor::*[@property][not(@content)]',
|
2046
|
+
'[not(@resource|@href|@src) or @rel|@rev])[1]' ].join('').freeze
|
2047
|
+
# note parentheses cause the index to be counted from the root
|
2048
|
+
|
2049
|
+
def vocab_for node
|
2050
|
+
if node[:vocab]
|
2051
|
+
vocab = node[:vocab].strip
|
2052
|
+
return nil if vocab == ''
|
2053
|
+
return vocab
|
2054
|
+
end
|
2055
|
+
parent = node.parent
|
2056
|
+
vocab_for parent if parent and parent.element?
|
2057
|
+
end
|
2058
|
+
|
2059
|
+
def prefixes_for node, prefixes = {}
|
2060
|
+
# start with namespaces
|
2061
|
+
pfx = node.namespaces.select do |k, _|
|
2062
|
+
k.start_with? 'xmlns:'
|
2063
|
+
end.transform_keys do |k|
|
2064
|
+
k.delete_prefix 'xmlns:'
|
2065
|
+
end
|
2066
|
+
|
2067
|
+
# then add @prefix overtop of the namespaces
|
2068
|
+
if node[:prefix]
|
2069
|
+
x = node[:prefix].strip.split(/\s+/)
|
2070
|
+
a = []
|
2071
|
+
b = []
|
2072
|
+
x.each_index { |i| (i % 2 == 0 ? a : b).push x[i] }
|
2073
|
+
# if the size is uneven the values will be nil, so w drop em
|
2074
|
+
pfx.merge! a.zip(b).to_h.reject { |_, v| v.nil? }
|
2075
|
+
end
|
2076
|
+
|
2077
|
+
# since we're ascending the tree, input takes precedence
|
2078
|
+
prefixes = pfx.merge prefixes
|
2079
|
+
|
2080
|
+
if node.parent and node.parent.element?
|
2081
|
+
prefixes_for(node.parent, prefixes)
|
2082
|
+
else
|
2083
|
+
prefixes
|
2084
|
+
end
|
2085
|
+
end
|
2086
|
+
|
2087
|
+
# give us the rdf subject of the node itself
|
2088
|
+
def subject_for node = nil, rdf: false, is_ancestor: false
|
2089
|
+
node ||= @doc.root
|
2090
|
+
raise 'Node must be an element' unless
|
2091
|
+
node.is_a? Nokogiri::XML::Element
|
2092
|
+
|
2093
|
+
# first we check for an ancestor element with @property and no
|
2094
|
+
# @content; if we find one then we reevaluate with that
|
2095
|
+
# element as the starting point
|
2096
|
+
if n = node.at_xpath(LITXP)
|
2097
|
+
return subject_for n
|
2098
|
+
end
|
2099
|
+
|
2100
|
+
# answer a bunch of helpful questions about this element
|
2101
|
+
subject = nil
|
2102
|
+
base = base_for node
|
2103
|
+
parent = node.parent
|
2104
|
+
ns_href = node.namespace.href if node.namespace
|
2105
|
+
up_ok = %i{rel rev}.none? { |a| node[a] }
|
2106
|
+
is_root = !parent or parent.document?
|
2107
|
+
special = /^(?:[^:]+:)?(?:head|body)$/i === node.name and
|
2108
|
+
(ns_href == 'http://www.w3.org/1999/xhtml' or
|
2109
|
+
/^(?:[^:]+:)?html$/xi === parent.name)
|
2110
|
+
|
2111
|
+
# if the node is being inspected as an ancestor to the
|
2112
|
+
# original node, we have to check it backwards.
|
2113
|
+
if is_ancestor
|
2114
|
+
# ah right @resource gets special treatment
|
2115
|
+
if subject = node[:resource]
|
2116
|
+
subject.strip!
|
2117
|
+
if m = /^\[(.*?)\]$/.match(subject)
|
2118
|
+
end
|
2119
|
+
else
|
2120
|
+
OBJS.each do |attr|
|
2121
|
+
if node[attr]
|
2122
|
+
# merge with the root and return it
|
2123
|
+
subject = base + node[attr]
|
2124
|
+
break
|
2125
|
+
end
|
2126
|
+
end
|
2127
|
+
end
|
2128
|
+
|
2129
|
+
return rdf ? RDF::URI(subject.to_s) : subject
|
2130
|
+
|
2131
|
+
# note if we are being called with is_ancestor, that means
|
2132
|
+
# the original node (or indeed any of the nodes previously
|
2133
|
+
# tested) have anything resembling a resource in them. this
|
2134
|
+
# means @rel/@rev should be ignored, and we should keep
|
2135
|
+
# looking for a subject.
|
2136
|
+
end
|
2137
|
+
|
2138
|
+
if node[:about]
|
2139
|
+
|
2140
|
+
if m = /^_:(.*)$/.match(node[:about])
|
2141
|
+
return RDF::Node(m[1])
|
2142
|
+
end
|
2143
|
+
|
2144
|
+
# XXX resolve @about against potential curie
|
2145
|
+
subject = base + node[:about]
|
2146
|
+
|
2147
|
+
elsif is_root
|
2148
|
+
subject = base
|
2149
|
+
elsif special
|
2150
|
+
subject = subject_for parent
|
2151
|
+
elsif node[:resource]
|
2152
|
+
# XXX resolve @about against potential curie
|
2153
|
+
subject = base + node[:resource]
|
2154
|
+
elsif node[:href]
|
2155
|
+
subject = base + node[:href]
|
2156
|
+
elsif node[:src]
|
2157
|
+
subject = base + node[:src]
|
2158
|
+
elsif node[:typeof]
|
2159
|
+
# bnode the typeof attr
|
2160
|
+
|
2161
|
+
# note we return bnodes irrespective of the rdf flag
|
2162
|
+
return RDF::Node('id-%016x' % node.attributes['typeof'].pointer_id)
|
2163
|
+
elsif node[:inlist]
|
2164
|
+
# bnode the inlist attr
|
2165
|
+
return RDF::Node('id-%016x' % node.attributes['inlist'].pointer_id)
|
2166
|
+
elsif (parent[:inlist] && OBJS.none? { |a| parent[a] }) ||
|
2167
|
+
(is_ancestor && !up_ok)
|
2168
|
+
# bnode the element
|
2169
|
+
return RDF::Node('id-%016x' % node.pointer_id)
|
2170
|
+
# elsif node[:id]
|
2171
|
+
else
|
2172
|
+
subject = subject_for parent, is_ancestor: true
|
2173
|
+
end
|
2174
|
+
|
2175
|
+
rdf ? RDF::URI(subject.to_s) : URI(subject.to_s)
|
2176
|
+
|
2177
|
+
end
|
2178
|
+
|
2179
|
+
# backlink structure
|
2180
|
+
def generate_backlinks published: true, ignore: nil
|
2181
|
+
@context.generate_backlinks @uuid, published: published, ignore: ignore
|
2182
|
+
end
|
2183
|
+
|
2184
|
+
# goofy twitter-specific metadata
|
2185
|
+
def generate_twitter_meta
|
2186
|
+
@context.generate_twitter_meta @uuid
|
2187
|
+
end
|
2188
|
+
|
2189
|
+
def transform_xhtml published: true
|
2190
|
+
# before we do any more work make sure this is html
|
2191
|
+
doc = @doc.dup 1
|
2192
|
+
body = doc.at_xpath('//html:body[1]', { html: XHTMLNS }) or return
|
2193
|
+
|
2194
|
+
# eliminate comments
|
2195
|
+
doc.xpath('//comment()[not(ancestor::html:script)]',
|
2196
|
+
{ html: XHTMLNS }).each { |c| c.unlink }
|
2197
|
+
|
2198
|
+
# initial stuff
|
2199
|
+
struct = @context.struct_for @uuid, uuids: true, canon: true
|
2200
|
+
# rstruct = @context.struct_for @uuid, uuids: true, rev: true
|
2201
|
+
resources = {}
|
2202
|
+
literals = {}
|
2203
|
+
ufwd = {} # uuid -> uri
|
2204
|
+
urev = {} # uri -> uuid
|
2205
|
+
datatypes = Set.new
|
2206
|
+
types = Set.new
|
2207
|
+
authors = @context.authors_for(@uuid)
|
2208
|
+
title = @context.label_for @uuid, candidates: struct
|
2209
|
+
desc = @context.label_for @uuid, candidates: struct, desc: true
|
2210
|
+
|
2211
|
+
# rewrite content
|
2212
|
+
title = title[1] if title
|
2213
|
+
desc = desc[1] if desc
|
2214
|
+
|
2215
|
+
# `struct` and `rstruct` will contain all the links and
|
2216
|
+
# metadata for forward and backward neighbours, respectively,
|
2217
|
+
# which we need to mine (predicates, classes, datatypes) for
|
2218
|
+
# prefixes among other things.
|
2219
|
+
|
2220
|
+
struct.each do |p, v|
|
2221
|
+
v.each do |o|
|
2222
|
+
if o.literal?
|
2223
|
+
literals[o] ||= Set.new
|
2224
|
+
literals[o].add p
|
2225
|
+
|
2226
|
+
# collect the datatype
|
2227
|
+
datatypes.add o.datatype if o.has_datatype?
|
2228
|
+
else
|
2229
|
+
# normalize URIs
|
2230
|
+
if o.to_s.start_with? 'urn:uuid:'
|
2231
|
+
ufwd[o] ||= @context.canonical_uri o
|
2232
|
+
elsif cu = @context.canonical_uuid(o)
|
2233
|
+
o = urev[o] ||= cu
|
2234
|
+
end
|
2235
|
+
|
2236
|
+
|
2237
|
+
# collect the resource
|
2238
|
+
resources[o] ||= Set.new
|
2239
|
+
resources[o].add p
|
2240
|
+
|
2241
|
+
# add to type
|
2242
|
+
types.add o if p == RDF::RDFV.type
|
2243
|
+
end
|
2244
|
+
end
|
2245
|
+
end
|
2246
|
+
urev.merge! ufwd.invert
|
2247
|
+
|
2248
|
+
labels = resources.keys.map do |k|
|
2249
|
+
# turn this into a pair which subsequently gets turned into a hash
|
2250
|
+
[k, @context.label_for(k) ]
|
2251
|
+
end.to_h
|
2252
|
+
|
2253
|
+
#warn labels
|
2254
|
+
|
2255
|
+
# handle the title
|
2256
|
+
title ||= RDF::Literal('')
|
2257
|
+
tm = { '#title' => title,
|
2258
|
+
property: @context.abbreviate(literals[title].to_a, vocab: XHV) }
|
2259
|
+
if tl = title.language
|
2260
|
+
tm['xml:lang'] = tl # if xmlns
|
2261
|
+
tm['lang'] = tl
|
2262
|
+
elsif tdt = title.datatype and tdt != RDF::XSD.string
|
2263
|
+
tm[:datatype] = @context.abbreviate(tdt)
|
2264
|
+
end
|
2265
|
+
|
2266
|
+
# we accumulate a record of the links in the body so we know
|
2267
|
+
# which ones to skip in the head
|
2268
|
+
bodylinks = {}
|
2269
|
+
rewrite_links body, uuids: ufwd, uris: urev do |elem|
|
2270
|
+
vocab = elem.at_xpath('ancestor-or-self::*[@vocab][1]/@vocab')
|
2271
|
+
vocab = uri_pp(vocab.to_s) if vocab
|
2272
|
+
|
2273
|
+
if elem.key?('href') or elem.key?('src')
|
2274
|
+
vu = uri_pp(elem['href'] || elem['src'])
|
2275
|
+
ru = RDF::URI(@uri.merge(vu))
|
2276
|
+
bodylinks[urev[ru] || ru] = true
|
2277
|
+
|
2278
|
+
if rel = resources[urev[ru] || ru]
|
2279
|
+
elem['rel'] = (@context.abbreviate rel, vocab: vocab).join ' '
|
2280
|
+
end
|
2281
|
+
|
2282
|
+
label = labels[urev[ru] || ru]
|
2283
|
+
if label and (!elem.key?('title') or elem['title'].strip == '')
|
2284
|
+
elem['title'] = label[1].to_s
|
2285
|
+
end
|
2286
|
+
end
|
2287
|
+
end
|
2288
|
+
|
2289
|
+
# and now we do the head
|
2290
|
+
links = []
|
2291
|
+
resources.reject { |k, _| bodylinks[k] }.each do |k, v|
|
2292
|
+
v = v.dup.delete RDF::RDFV.type
|
2293
|
+
next if v.empty?
|
2294
|
+
mts = @context.formats_for k
|
2295
|
+
|
2296
|
+
# warn k, v.inspect
|
2297
|
+
|
2298
|
+
# warn k, mts.inspect
|
2299
|
+
|
2300
|
+
rel = @context.abbreviate v.to_a, vocab: XHV
|
2301
|
+
ru = @uri.route_to(uri_pp (ufwd[k] || k).to_s)
|
2302
|
+
ln = { nil => :link, rel: rel, href: ru.to_s }
|
2303
|
+
if (label = labels[urev[k] || k])
|
2304
|
+
ln[:title] = label[1].to_s
|
2305
|
+
end
|
2306
|
+
|
2307
|
+
# add type=lol/wut
|
2308
|
+
ln[:type] = mts.first.to_s unless mts.empty?
|
2309
|
+
|
2310
|
+
if !ln[:type] and v.include?(RDF::Vocab::XHV.stylesheet)
|
2311
|
+
ln[:type] = 'text/css'
|
2312
|
+
elsif ln[:type] =~ /(java|ecma)script/i or
|
2313
|
+
v.include?(RDF::Vocab::DC.requires)
|
2314
|
+
ln[nil] = :script
|
2315
|
+
ln[:src] = ln.delete :href
|
2316
|
+
ln[:type] ||= 'text/javascript'
|
2317
|
+
end
|
2318
|
+
links.push ln
|
2319
|
+
end
|
2320
|
+
|
2321
|
+
links.sort! do |a, b|
|
2322
|
+
# sort by rel, then by href
|
2323
|
+
# warn a.inspect, b.inspect
|
2324
|
+
s = 0
|
2325
|
+
[nil, :rel, :rev, :href, :title].each do |k|
|
2326
|
+
s = a.fetch(k, '').to_s <=> b.fetch(k, '').to_s
|
2327
|
+
break if s != 0
|
2328
|
+
end
|
2329
|
+
s
|
2330
|
+
end
|
2331
|
+
|
2332
|
+
# we want to duplicate links from particular subjects (eg the root)
|
2333
|
+
(@context.config[:duplicate] || {}).sort do |a, b|
|
2334
|
+
a.first <=> b.first
|
2335
|
+
end.each do |s, preds|
|
2336
|
+
|
2337
|
+
o = {}
|
2338
|
+
u = ufwd[s] ||= @context.canonical_uuid s
|
2339
|
+
s = urev[u] ||= @context.canonical_uri u if u
|
2340
|
+
f = {}
|
2341
|
+
|
2342
|
+
# do not include this subject as these links are already included!
|
2343
|
+
next if u == @uuid
|
2344
|
+
|
2345
|
+
# gather up the objects, then gather up the predicates
|
2346
|
+
|
2347
|
+
@context.objects_for u || s, preds, only: :resource do |obj, rel|
|
2348
|
+
# XXX do not know why += |= etc does not work
|
2349
|
+
x = @context.canonical_uuid(obj) || obj
|
2350
|
+
urev[x] ||= @context.canonical_uri x
|
2351
|
+
y = o[x] ||= Set.new
|
2352
|
+
o[x] = y | rel
|
2353
|
+
f[x] = @context.formats_for x
|
2354
|
+
end
|
2355
|
+
|
2356
|
+
srel = @uri.route_to((u ? urev[u] || s : s).to_s)
|
2357
|
+
|
2358
|
+
# now collect all the other predicates
|
2359
|
+
o.keys.each do |obj|
|
2360
|
+
hrel = @uri.route_to((urev[obj] || obj).to_s)
|
2361
|
+
o[obj] |= @context.graph.query([u || s, nil, obj]).predicates.to_set
|
2362
|
+
rels = @context.abbreviate o[obj].to_a, vocab: XHV
|
2363
|
+
ln = { nil => :link, about: srel, rel: rels, href: hrel }
|
2364
|
+
ln[:type] = f[obj].first if f[obj]
|
2365
|
+
|
2366
|
+
# add to links
|
2367
|
+
links << ln
|
2368
|
+
end
|
2369
|
+
end
|
2370
|
+
|
2371
|
+
meta = []
|
2372
|
+
|
2373
|
+
# include author names as old school meta tags
|
2374
|
+
authors.each do |a|
|
2375
|
+
name = labels[urev[a] || a] or next
|
2376
|
+
datatypes.add name[0] # a convenient place to chuck this
|
2377
|
+
prop = @context.abbreviate(name[0])
|
2378
|
+
name = name[1]
|
2379
|
+
about = @uri.route_to((ufwd[a] || a).to_s)
|
2380
|
+
tag = { nil => :meta, about: about.to_s, name: :author,
|
2381
|
+
property: prop, content: name.to_s }
|
2382
|
+
|
2383
|
+
if name.has_datatype? and name.datatype != RDF::XSD.string
|
2384
|
+
tag[:datatype] = @context.abbreviate(name.datatype)
|
2385
|
+
elsif name.has_language?
|
2386
|
+
tag['xml:lang'] = tag[:lang] = name.language
|
2387
|
+
end
|
2388
|
+
meta.push tag
|
2389
|
+
end
|
2390
|
+
|
2391
|
+
literals.each do |k, v|
|
2392
|
+
next if k == title
|
2393
|
+
rel = @context.abbreviate v.to_a, vocab: XHV
|
2394
|
+
elem = { nil => :meta, property: rel, content: k.to_s }
|
2395
|
+
elem[:name] = :description if k == desc
|
2396
|
+
|
2397
|
+
if k.has_datatype?
|
2398
|
+
datatypes.add k.datatype # so we get the prefix
|
2399
|
+
elem[:datatype] = @context.abbreviate k.datatype, vocab: XHV
|
2400
|
+
end
|
2401
|
+
|
2402
|
+
meta.push(elem)
|
2403
|
+
end
|
2404
|
+
|
2405
|
+
meta.sort! do |a, b|
|
2406
|
+
s = 0
|
2407
|
+
[:about, :property, :datatype, :content, :name].each do |k|
|
2408
|
+
# warn a.inspect, b.inspect
|
2409
|
+
s = a.fetch(k, '').to_s <=> b.fetch(k, '').to_s
|
2410
|
+
break if s != 0
|
2411
|
+
end
|
2412
|
+
s
|
2413
|
+
end
|
2414
|
+
|
2415
|
+
# don't forget style tag
|
2416
|
+
style = doc.xpath('/html:html/html:head/html:style', { html: XHTMLNS })
|
2417
|
+
|
2418
|
+
body = body.dup 1
|
2419
|
+
body = { '#body' => body.children.to_a, about: '' }
|
2420
|
+
body[:typeof] = @context.abbreviate(types.to_a, vocab: XHV) unless
|
2421
|
+
types.empty?
|
2422
|
+
|
2423
|
+
# prepare only the prefixes we need to resolve the data we need
|
2424
|
+
rsc = @context.abbreviate(
|
2425
|
+
(struct.keys + resources.keys + datatypes.to_a + types.to_a).uniq,
|
2426
|
+
noop: false).map do |x|
|
2427
|
+
next if x.nil?
|
2428
|
+
x.split(?:)[0].to_sym
|
2429
|
+
end.select { |x| not x.nil? }.to_set
|
2430
|
+
|
2431
|
+
pfx = @context.prefixes.select do |k, _|
|
2432
|
+
rsc.include? k
|
2433
|
+
end.transform_values { |v| v.to_s }
|
2434
|
+
|
2435
|
+
# XXX deal with the qb:Observation separately (just nuke it for now)
|
2436
|
+
extra = generate_twitter_meta || []
|
2437
|
+
if bl = generate_backlinks(published: published,
|
2438
|
+
ignore: @context.graph.query(
|
2439
|
+
[nil, CI.document, @uuid]).subjects.to_set)
|
2440
|
+
extra << { [bl] => :object }
|
2441
|
+
end
|
2442
|
+
|
2443
|
+
# and now for the document
|
2444
|
+
xf = @context.config[:transform]
|
2445
|
+
doc = xhtml_stub(
|
2446
|
+
base: @uri, prefix: pfx, vocab: XHV, lang: 'en', title: tm,
|
2447
|
+
link: links, meta: meta, style: style, transform: xf,
|
2448
|
+
extra: extra, body: body).document
|
2449
|
+
|
2450
|
+
# goddamn script tags and text/html
|
2451
|
+
doc.xpath('//html:script[@src][not(node())]',
|
2452
|
+
{ html: XHTMLNS }).each do |script|
|
2453
|
+
script << doc.create_text_node('')
|
2454
|
+
end
|
2455
|
+
|
2456
|
+
doc
|
2457
|
+
end
|
2458
|
+
|
2459
|
+
# Actually write the transformed document to the target
|
2460
|
+
#
|
2461
|
+
# @param published [true, false]
|
2462
|
+
#
|
2463
|
+
# @return [Array] pathname(s) written
|
2464
|
+
def write_to_target published: true
|
2465
|
+
|
2466
|
+
# in all cases we write to private target
|
2467
|
+
states = [false]
|
2468
|
+
# document has to be publishable
|
2469
|
+
states.push true if published && @context.published?(@uuid)
|
2470
|
+
|
2471
|
+
ok = []
|
2472
|
+
states.each do |state|
|
2473
|
+
target = @context.config[state ? :target : :private]
|
2474
|
+
|
2475
|
+
# XXX this is dumb; it should do something more robust if it
|
2476
|
+
# fails
|
2477
|
+
doc = transform_xhtml(published: state) or next
|
2478
|
+
|
2479
|
+
begin
|
2480
|
+
fh = Tempfile.create('xml-', target)
|
2481
|
+
path = Pathname(fh.path)
|
2482
|
+
|
2483
|
+
# write the doc to the target
|
2484
|
+
doc.write_to fh
|
2485
|
+
fh.close
|
2486
|
+
|
2487
|
+
uuid = URI(@uuid.to_s)
|
2488
|
+
newpath = path.dirname + "#{uuid.uuid}.xml"
|
2489
|
+
ok.push newpath
|
2490
|
+
|
2491
|
+
File.chmod(0644, path)
|
2492
|
+
File.rename(path, newpath)
|
2493
|
+
File.utime(@mtime, @mtime, newpath)
|
2494
|
+
rescue Exception => e
|
2495
|
+
# XXX this should only rescue a specific class of errors
|
2496
|
+
warn e.class, e
|
2497
|
+
File.unlink path if path.exist?
|
2498
|
+
end
|
2499
|
+
end
|
2500
|
+
|
2501
|
+
ok
|
2502
|
+
end
|
2503
|
+
|
2504
|
+
end
|
2505
|
+
end
|
2506
|
+
end
|