Module: RDF::SAK::Util

Extended by:
Util
Included in:
Context, Context::Document, Document, Util
Defined in:
lib/rdf/sak/util.rb

Constant Summary collapse

SCHEME_RANK = 
{ https: 0, http: 1 }
XHTMLNS = 
'http://www.w3.org/1999/xhtml'.freeze
XHV = 
'http://www.w3.org/1999/xhtml/vocab#'.freeze
XPATHNS = 
{
  html:  XHTMLNS,
  svg:   'http://www.w3.org/2000/svg',
  atom:  'http://www.w3.org/2005/Atom',
  xlink: 'http://www.w3.org/1999/xlink',
}.freeze

Class Method Summary collapse

Instance Method Summary collapse

Class Method Details

.asserted_types(repo, subject, type = nil) ⇒ Array

Obtain all and only the rdf:types directly asserted on the subject.

Parameters:

  • repo (RDF::Queryable)
  • subject (RDF::Resource)
  • type (RDF::Term, :to_a) (defaults to: nil)

Returns:

  • (Array) 


431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
 
# File 'lib/rdf/sak/util.rb', line 431

def self.asserted_types repo, subject, type = nil
  asserted = nil

  if type
    type = type.respond_to?(:to_a) ? type.to_a : [type]
    asserted = type.select { |t| t.is_a? RDF::Value }.map do |t|
      RDF::Vocabulary.find_term t
    end
  end

  asserted ||= repo.query([subject, RDF.type, nil]).objects.map do |o|
    RDF::Vocabulary.find_term o
  end.compact

  asserted.select { |t| t && t.uri? }.uniq
end

.base_for(xmlnode, base) ⇒ Object 



1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
 
# File 'lib/rdf/sak/util.rb', line 1269

def self.base_for xmlnode, base
  base = URI(base.to_s) unless base.is_a? URI
  out  = base

  if xmlnode.at_xpath('self::html:*|/html', XPATHNS)
    b = URI(xmlnode.at_xpath(XPATH[:htmlbase], XPATHNS).to_s.strip)
    
    out = b if b.absolute?
  elsif b = xmlnode.root.at_xpath(XPATH[:xmlbase])
    b = URI(b.to_s.strip)
    out = b if b.absolute?
  end

  out
end

.canonical_uri(repo, subject, base: nil, unique: true, rdf: true, slugs: false, fragment: false) ⇒ RDF::URI, ...

Obtain the “best” dereferenceable URI for the subject. Optionally returns all candidates.

Parameters:

  • repo (RDF::Queryable)
  • subject (RDF::Resource)
  • unique (true, false) (defaults to: true) —

    flag for unique return value

  • rdf (true, false) (defaults to: true) —

    flag to specify RDF::URI vs URI

  • slugs (true, false) (defaults to: false) —

    flag to include slugs

  • fragment (true, false) (defaults to: false) —

    flag to include fragment URIs

Returns:

  • (RDF::URI, URI, Array) 


951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
 
# File 'lib/rdf/sak/util.rb', line 951

def self.canonical_uri repo, subject, base: nil,
    unique: true, rdf: true, slugs: false, fragment: false
  subject = coerce_resource subject, base
  out = []

  # try to find it first
  out = objects_for(repo, subject, [RDF::SAK::CI.canonical, RDF::OWL.sameAs],
    entail: false, only: :resource).select do |o|
    # only consider the subjects
    repo.has_subject? o
  end.sort { |a, b| cmp_resource a, b }

  # try to generate in lieu
  if subject.uri? and (out.empty? or slugs)

    out += objects_for(repo, subject,
      [RDF::SAK::CI['canonical-slug'], RDF::SAK::CI.slug],
      only: :literal).map do |o|
      base + o.value
    end if slugs

    uri = URI(uri_pp(subject.to_s))
    if base and uri.respond_to? :uuid
      b = base.clone
      b.query = b.fragment = nil
      b.path = '/' + uri.uuid
      out << RDF::URI.new(b.to_s)
    else
      out << subject
    end
  end

  # remove all URIs with fragments unless specified
  unless fragment
    tmp = out.reject(&:fragment)
    out = tmp unless tmp.empty?
  end

  # coerce to URI objects if specified
  out.map! { |u| URI(uri_pp u.to_s) } unless rdf

  unique ? out.first : out.uniq
end

.canonical_uuid(repo, uri, unique: true, published: false, scache: {}, ucache: {}, base: nil) ⇒ RDF::URI, Array

Obtain the canonical UUID for the given URI

Parameters:

  • repo (RDF::Queryable)
  • uri (RDF::URI, URI, to_s) —

    the subject of the inquiry

  • unique (true, false) (defaults to: true) —

    return a single resource/nil or an array

  • published (true, false) (defaults to: false) —

    whether to restrict to published docs

Returns:

  • (RDF::URI, Array) 


704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
 
# File 'lib/rdf/sak/util.rb', line 704

def self.canonical_uuid repo, uri, unique: true, published: false,
    scache: {}, ucache: {}, base: nil
  # make sure this is actually a uri
  orig = uri = coerce_resource uri, base
  unless uri.is_a? RDF::Node
    tu = URI(uri_pp(uri).to_s).normalize

    if tu.path && !tu.fragment &&
        UUID_RE.match?(uu = tu.path.delete_prefix(?/))
      tu = URI('urn:uuid:' + uu.downcase)
    end

    # unconditionally overwrite uri
    uri = RDF::URI(tu.to_s)

    # now check if it's a uuid
    if tu.respond_to? :uuid
      # warn "lol uuid #{orig}"
      # if it's a uuid, check that we have it as a subject
      # if we have it as a subject, return it
      return uri if scache[uri] ||= repo.has_subject?(uri)
      # note i don't want to screw around right now dealing with the
      # case that a UUID might not itself be canonical
    end
  end

  # spit up the cache if present
  if out = ucache[orig]
    # warn "lol cached #{orig}"
    return unique ? out.first : out
  end

  # otherwise we proceed:

  # goal: return the most "appropriate" UUID for the given URI

  # it is so lame i have to do this
  bits = { nil => 0, false => 0, true => 1 }

  # rank (0 is higher):
  # * (00) exact & canonical == 0,
  # * (01) exact == 1,
  # * (10) inexact & canonical == 2,
  # * (11) inexact == 3. 

  # warn "WTF URI #{uri}"

  # handle path parameters by generating a bunch of candidates
  uris = if uri.respond_to? :path and uri.path.start_with? ?/
           # split any path parameters off
           uu, *pp = split_pp uri
           if pp.empty?
             [uri] # no path parameters
           else
             uu = RDF::URI(uu.to_s)
             bp = uu.path # base path
             (0..pp.length).to_a.reverse.map do |i|
               u = uu.dup
               u.path = ([bp] + pp.take(i)).join(';')
               u
             end
           end
         else
           [uri] # not a pathful URI
         end

  # collect the candidates by URI
  sa = predicate_set [RDF::SAK::CI.canonical,
    RDF::SAK::CI.alias, RDF::OWL.sameAs]
  candidates = nil
  uris.each do |u|
    candidates = subjects_for(repo, sa, u, entail: false) do |s, f|
      # there is no #to_i for booleans and also we xor this number
      [s, { rank: bits[f.include?(RDF::SAK::CI.canonical)] ^ 1,
        published: published?(repo, s),
        mtime: dates_for(repo, s).last || DateTime.new }]
    end.compact.to_h
    break unless candidates.empty?
  end

  # now collect by slug
  slug = terminal_slug uri, base: base
  if slug and !slug.empty?
    exact = uri == coerce_resource(slug, base) # slug represents exact match
    sl = [RDF::SAK::CI['canonical-slug'], RDF::SAK::CI.slug]
    [RDF::XSD.string, RDF::XSD.token].each do |t|
      subjects_for(repo, sl, RDF::Literal(slug, datatype: t)) do |s, f|
        # default to lowest rank if this candidate is new
        entry = candidates[s] ||= {
          published: published?(repo, s, base: base),
          rank: 0b11, mtime: dates_for(repo, s).last || DateTime.new }
        # true is 1 and false is zero so we xor this too
        rank  = (BITS[exact] << 1 | BITS[f.include?(sl[0])]) ^ 0b11
        # now amend the rank if we have found a better one
        entry[:rank] = rank if rank < entry[:rank]
      end
    end
  end

  candidates.delete_if { |s, _| !/^urn:uuid:/.match?(s.to_s)  }

  # scan all the candidates for replacements and remove any
  # candidates that have been replaced
  candidates.to_a.each do |k, v|
    # note that 
    reps = replacements_for(repo, k, published: published) - [k]
    unless reps.empty?
      v[:replaced] = true
      reps.each do |r|
        c = candidates[r] ||= { rank: v[:rank],
          published: published?(repo, r),
          mtime: dates_for(repo, r).last || v[:mtime] || DateTime.new }
        # we give the replacement the rank and mtime of the
        # resource being replaced if it scores better
        c[:rank]  = v[:rank]  if v[:rank]  < c[:rank]
        c[:mtime] = v[:mtime] if v[:mtime] > c[:mtime]
      end
    end
  end

  # now we can remove all unpublished candidates if the context is
  # published
  candidates.select! do |_, v|
    !v[:replaced] && (published ? v[:published] : true)
  end

  # now we sort by rank and date; the highest-ranking newest
  # candidate is the one

  out = candidates.sort do |a, b|
    _, va = a
    _, vb = b
    cb = published ? BITS[vb[:published]] <=> BITS[va[:published]] : 0
    cr = va[:rank] <=> vb[:rank]
    cb == 0 ? cr == 0 ? vb[:mtime] <=> va[:mtime] : cr : cb
  end.map { |x| x.first }.compact

  # set cache
  ucache[orig] = out

  #warn "lol not cached #{orig}"

  unique ? out.first : out

  # an exact match is better than an inexact one

  # a canonical match is better than non-canonical

  # note this is four bits: exact, canon(exact), inexact, canon(inexact)
  # !canon(exact) should rank higher than canon(inexact)

  # unreplaced is better than replaced

  # newer is better than older (though no reason an older item
  # can't replace a newer one)

  # published is better than not, unless the context is
  # unpublished and an unpublished document replaces a published one
end

.cmp_label(repo, a, b, labels: nil, supplant: true, reverse: false) ⇒ Object 



919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
 
# File 'lib/rdf/sak/util.rb', line 919

def self.cmp_label repo, a, b, labels: nil, supplant: true, reverse: false
  labels ||= {}

  # try supplied label or fall back
  pair = [a, b].map do |x|
    if labels[x]
      labels[x][1]
    elsif supplant and y = label_for(repo, x)
      labels[x] = y
      y[1]
    else
      x
    end
  end

  pair.reverse! if reverse
  # warn "#{pair[0]} <=> #{pair[1]}"
  pair[0].to_s <=> pair[1].to_s
end

.dates_for(repo, subject, predicate: RDF::Vocab::DC.date, datatype: [RDF::XSD.date, RDF::XSD.dateTime]) ⇒ Array

Obtain dates for the subject as instances of Date(Time). This is just shorthand for a common application of ‘objects_for`.

Parameters:

  • repo
  • subject
  • predicate (defaults to: RDF::Vocab::DC.date)
  • datatype (defaults to: [RDF::XSD.date, RDF::XSD.dateTime])

Returns:

  • (Array) —

    of dates



1242
1243
1244
1245
1246
1247
1248
 
# File 'lib/rdf/sak/util.rb', line 1242

def self.dates_for repo, subject, predicate: RDF::Vocab::DC.date,
    datatype: [RDF::XSD.date, RDF::XSD.dateTime]
  objects_for(
    repo, subject, predicate, only: :literal, datatype: datatype) do |o|
    o.object
  end.sort.uniq
end

.label_for(repo, subject, candidates: nil, unique: true, type: nil, lang: nil, desc: false, alt: false, base: nil) ⇒ Array

Obtain the most appropriate label(s) for the subject’s type(s). Returns one or more (depending on the ‘unique` flag) predicate-object pairs in order of preference.

Parameters:

  • repo (RDF::Queryable)
  • subject (RDF::Resource)
  • unique (true, false) (defaults to: true) —

    only return the first pair

  • type (RDF::Term, Array) (defaults to: nil) —

    supply asserted types if already retrieved

  • lang (nil) (defaults to: nil) —

    not currently implemented (will be conneg)

  • desc (false, true) (defaults to: false) —

    retrieve description instead of label

  • alt (false, true) (defaults to: false) —

    retrieve alternate instead of main

Returns:

  • (Array) —

    either a predicate-object pair or an array of pairs.

Raises:

  • (ArgumentError) 


1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
 
# File 'lib/rdf/sak/util.rb', line 1077

def self.label_for repo, subject, candidates: nil, unique: true, type: nil,
    lang: nil, desc: false, alt: false, base: nil
  raise ArgumentError, 'no repo!' unless repo.is_a? RDF::Queryable
  return unless subject.is_a? RDF::Value and subject.resource?
  
  asserted = asserted_types repo, subject, type

  # get all the inferred types by layer; add default class if needed
  strata = type_strata asserted
  strata.push [RDF::RDFS.Resource] if
    strata.empty? or not strata[-1].include?(RDF::RDFS.Resource)

  # get the key-value pairs for the subject
  candidates ||= struct_for repo, subject, only: :literal

  seen  = {}
  accum = []
  strata.each do |lst|
    lst.each do |cls|
      next unless STRINGS[cls] and
        preds = STRINGS[cls][desc ? :desc : :label][alt ? 1 : 0]
      # warn cls
      preds.each do |p|
        # warn p.inspect
        next unless vals = candidates[p]
        vals.each do |v|
          pair = [p, v]
          accum.push(pair) unless seen[pair]
          seen[pair] = true
        end
      end
    end
  end

  # try that for now
  unique ? accum[0] : accum.uniq
  
  # what we want to do is match the predicates from the subject to
  # the predicates in the label designation
  
  # get label predicate stack(s) for RDF type(s)
  
  # get all predicates in order (use alt stack if doubly specified)
  
  # filter out desired language(s)
  
  # XXX note we will probably want to return the predicate as well
end

.objects_for(repo, subject, predicate, entail: true, only: [], datatype: nil) ⇒ RDF::Term

Returns objects from the graph with entailment.

Parameters:

  • repo
  • subject
  • predicate
  • entail (defaults to: true)
  • only (defaults to: [])
  • datatype (defaults to: nil)

Returns:

  • (RDF::Term) 


631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
 
# File 'lib/rdf/sak/util.rb', line 631

def self.objects_for repo, subject, predicate,
    entail: true, only: [], datatype: nil
  raise "Subject must be a resource, not #{subject.inspect}" unless
    subject.is_a? RDF::Resource
  predicate = predicate.respond_to?(:to_a) ? predicate.to_a : [predicate]
  raise "Predicate must be a term, not #{predicate.first.class}" unless
    predicate.all? { |p| p.is_a? RDF::URI }

  predicate = predicate.map { |x| RDF::Vocabulary.find_term x }.compact

  only = coerce_node_spec only

  datatype = (
    datatype.respond_to?(:to_a) ? datatype.to_a : [datatype]).compact
  raise 'Datatype must be some kind of term' unless
    datatype.all? { |p| p.is_a? RDF::URI }

  # fluff this out 
  predicate = predicate_set predicate if entail

  out = {}
  predicate.each do |p|
    repo.query([subject, p, nil]).objects.each do |o|

      # make sure it's in the spec
      next unless node_matches? o, only

      # constrain output
      next if o.literal? and
        !(datatype.empty? or datatype.include?(o.datatype))

      entry = out[o] ||= [Set.new, Set.new]
      entry.first << p
    end
  end

  # now we do the reverse
  unless only == [:literal]
    # generate reverse predicates
    revp = Set.new
    predicate.each do |p|
      revp += p.inverseOf.to_set
      revp << p if p.type.include? RDF::OWL.SymmetricProperty
    end
    revp = predicate_set revp if entail

    # now scan 'em
    revp.each do |p|
      repo.query([nil, p, subject]).subjects.each do |s|
        next unless node_matches? s, only
        # no need to check datatype; subject is never a literal

        entry = out[s] ||= [Set.new, Set.new]
        entry.last << p
      end
    end
  end

  # run this through a block to get access to the predicates
  return out.map { |p, v| yield p, *v } if block_given?

  out.keys
end

.published?(repo, uri, circulated: false, base: nil) ⇒ true, false

Determine whether the URI represents a published document.

Parameters:

  • repo
  • uri

Returns:

  • (true, false) 


1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
 
# File 'lib/rdf/sak/util.rb', line 1001

def self.published? repo, uri, circulated: false, base: nil
  uri = coerce_resource uri, base
  candidates = objects_for(
    repo, uri, RDF::Vocab::BIBO.status, only: :resource).to_set

  test = Set[RDF::Vocab::BIBO['status/published']]
  test << RDF::SAK::CI.circulated if circulated

  # warn candidates, test, candidates & test

  !(candidates & test).empty?
end

.replacements_for(repo, subject, published: true, base: nil) ⇒ Set

Find the terminal replacements for the given subject, if any exist.

Parameters:

  • repo
  • subject
  • published (defaults to: true) —

    indicate the context is published

Returns:

  • (Set) 


1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
 
# File 'lib/rdf/sak/util.rb', line 1173

def self.replacements_for repo, subject, published: true, base: nil
  subject = coerce_resource subject, base

  # `seen` is a hash mapping resources to publication status and
  # subsequent replacements. it collects all the resources in the
  # replacement chain in :fwd (replaces) and :rev (replaced-by)
  # members, along with a boolean :pub. `seen` also performs a
  # duty as cycle-breaking sentinel.

  seen  = {}
  queue = [subject]
  while (test = queue.shift)
    # fwd is "replaces", rev is "replaced by"
    entry = seen[test] ||= {
      pub: published?(repo, test), fwd: Set.new, rev: Set.new }
    queue += (
      subjects_for(repo, RDF::Vocab::DC.replaces, subject) +
        objects_for(repo, subject, RDF::Vocab::DC.isReplacedBy,
        only: :resource)
    ).uniq.map do |r| # r = replacement
      next if seen.include? r
      seen[r] ||= { pub: published?(repo, r), fwd: Set.new, rev: Set.new }
      seen[r][:fwd] << test
      entry[:rev] << r
      r
    end.compact.uniq
  end

  # if we're calling from a published context, we return the
  # (topologically) last published resource(s), even if they are
  # replaced ultimately by unpublished resources.
    
  out = seen.map { |k, v| v[:rev].empty? ? k : nil }.compact - [subject]

  # now we modify `out` based on the publication status of the context
  if published
    pubout = out.select { |o| seen[o][:pub] }
    # if there is anything left after this, return it
    return pubout unless pubout.empty?
    # now we want to find the penultimate elements of `seen` that
    # are farthest along the replacement chain but whose status is
    # published

    # start with `out`, take the union of their :fwd members, then
    # take the subset of those which are published. if the result
    # is empty, repeat. (this is walking backwards through the
    # graph we just walked forwards through to construct `seen`)
    loop do
      # XXX THIS NEEDS A TEST CASE
      out = seen.values_at(*out).map { |v| v[:fwd] }.reduce(:+).to_a
      break if out.empty?
      pubout = out.select { |o| seen[o][:pub] }
      return pubout unless pubout.empty?
    end
  end

  out
end

.struct_for(repo, subject, base: nil, rev: false, only: [], uuids: false, canon: false, ucache: {}, scache: {}) ⇒ Hash

Obtain a key-value structure for the given subject, optionally constraining the result by node type (:resource, :uri/:iri, :blank/:bnode, :literal)

Parameters:

  • repo
  • subject —

    of the inquiry

  • rev (defaults to: false) —

    map in reverse

  • only (defaults to: []) —

    one or more node types

  • uuids (defaults to: false) —

    coerce resources to if possible

Returns:

  • (Hash) 


1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
 
# File 'lib/rdf/sak/util.rb', line 1026

def self.struct_for repo, subject, base: nil,
    rev: false, only: [], uuids: false, canon: false, ucache: {}, scache: {}
  only = coerce_node_spec only

  # coerce the subject
  subject = canonical_uuid(repo, subject,
    base: base, scache: scache, ucache: ucache) || subject if uuids 

  rsrc = {}
  pattern = rev ? [nil, nil, subject] : [subject, nil, nil]
  repo.query(pattern) do |stmt|
    # this will skip over any term not matching the type
    node = rev ? stmt.subject : stmt.object
    next unless node_matches? node, only

    # coerce the node to uuid if told to
    if node.resource?
      if uuids
        uu = canonical_uuid(repo, node, scache: scache, ucache: ucache) unless
          ucache.key? node
        node = uu || (canon ? canonical_uri(repo, node) : node)
      elsif canon
        node = canonical_uri(repo, node)
      end
    end

    p = RDF::Vocabulary.find_term(stmt.predicate) || stmt.predicate
    o = rsrc[p] ||= []
    o.push node if node # may be nil
  end

  # XXX in here we can do fun stuff like filter/sort by language/datatype
  rsrc.values.each { |v| v.sort!.uniq! }

  rsrc
end

.subjects_for(repo, predicate, object, entail: true, only: []) ⇒ RDF::Resource

Returns subjects from the graph with entailment.

Parameters:

  • repo
  • predicate
  • object
  • entail (defaults to: true)
  • only (defaults to: [])

Returns:

  • (RDF::Resource) 


573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
 
# File 'lib/rdf/sak/util.rb', line 573

def self.subjects_for repo, predicate, object, entail: true, only: []
  raise 'Object must be a Term' unless object.is_a? RDF::Term
  predicate = predicate.respond_to?(:to_a) ? predicate.to_a : [predicate]
  raise 'Predicate must be some kind of term' unless
    predicate.all? { |p| p.is_a? RDF::URI }

  only = coerce_node_spec only, rev: true

  predicate = predicate.map { |x| RDF::Vocabulary.find_term x }.compact
  predicate = predicate_set predicate if entail

  out  = {}
  revp = Set.new
  predicate.each do |p|
    repo.query([nil, p, object]).subjects.each do |s|
      next unless node_matches? s, only

      entry = out[s] ||= [Set.new, Set.new]
      entry[0] << p
    end

    # do this here while we're at it
    unless object.literal?
      revp += p.inverseOf.to_set
      revp << p if p.type.include? RDF::OWL.SymmetricProperty
    end
  end

  unless object.literal?
    revp = predicate_set revp if entail

    revp.each do |p|
      repo.query([object, p, nil]).objects.each do |o|
        next unless node_matches? o, only

        entry = out[o] ||= [Set.new, Set.new]
        entry[1] << p
      end
    end
  end

  # run this through a block to get access to the predicates
  return out.map { |p, v| yield p, *v } if block_given?

  out.keys
end

.traverse_links(node, type: 'application/xhtml+xml', &block) ⇒ Object

Traverse links based on content type.



1286
1287
1288
1289
1290
1291
 
# File 'lib/rdf/sak/util.rb', line 1286

def self.traverse_links node, type: 'application/xhtml+xml', &block
  enum_for :traverse_links, node, type: type unless block
  type  = type.strip.downcase.gsub(/\s*;.*/, '')
  xpath = LINK_MAP.fetch type, XPATH[:xlinks]
  node.xpath(xpath, XPATHNS).each { |node| block.call node }
end

Instance Method Details

#abbreviate(term, prefixes: {}, vocab: nil, noop: true, sort: true) ⇒ String, ...

Note:

Only noop: true can be guaranteed to return a value.

Abbreviate one or more URIs into one or more CURIEs if we can. Will through if noop: is true, or if false, return nil for any URI that can’t be abbreviated this way. Takes a hash of prefix-URI mappings where the keys are assumed to be symbols or nil to express the current vocabulary, which can be overridden via vocab:.

Parameters:

  • term (Array<#to_s>, #to_s) —

    the term(s)

  • prefixes (Hash<Symbol,nil>, #to_h) (defaults to: {}) —

    the prefix mappings

  • vocab (#to_s) (defaults to: nil) —

    current vocabulary, overrides prefixes[nil]

  • noop (true, false) (defaults to: true) —

    whether or not to pass terms through

  • sort (true, false) (defaults to: true) —

    whether or not to sort (only if noop:)

Returns:

  • (String, nil, Array<String,nil>) —

    the (maybe) abbreviated term(s)

Raises:

  • (ArgumentError) 


1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
 
# File 'lib/rdf/sak/util.rb', line 1643

def abbreviate term, prefixes: {}, vocab: nil, noop: true, sort: true
  # this returns a duplicate that we can mess with
  prefixes = sanitize_prefixes prefixes

  # sanitize vocab
  raise ArgumentError, 'vocab must be nil or stringable' unless
    vocab.nil? or vocab.respond_to? :to_s
  prefixes[nil] = vocab.to_s if vocab
  scalar = true

  term = if term.respond_to? :to_a
           scalar = false
           term.to_a
         else [term]; end

  rev = prefixes.invert

  term.map! do |t|
    t = t.to_s
    slug = nil # we want this value to be nil if no match and !noop

    # try matching each prefix URI from longest to shortest
    rev.sort { |a, b| b.first.length <=> a.first.length }.each do |uri, pfx|
      slug = t.delete_prefix uri
      # this is saying the URI either doesn't match or abbreviates to ""
      if slug == t or pfx.nil? && slug.empty?
        slug = nil
      else
        # it's already a slug so we add a prefix if there is one
        slug = '%s:%s' % [pfx, slug] unless pfx.nil?
        break # we have our match
      end
    end

    # at this point slug is either an abbreviated term or nil, so:
    slug ||= t if noop
    slug
  end

  # only sort if noop is set
  term.sort! if noop && sort

  scalar ? term.first : term
end

#all_related(rdftype) ⇒ Array

Obtain everything that is an owl:equivalentClass or rdfs:subClassOf the given type.

Parameters:

  • rdftype (RDF::Term)

Returns:

  • (Array) 


2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
 
# File 'lib/rdf/sak/util.rb', line 2046

def all_related rdftype
  t = RDF::Vocabulary.find_term(rdftype) or raise "No type #{rdftype.to_s}"
  q = [t] # queue
  c = {}  # cache

  while term = q.shift
    # add term to cache
    c[term] = term

    # keep this from tripping up
    next unless term.uri? and term.respond_to? :class?

    # entail equivalent classes
    term.entail(:equivalentClass).each do |ec|
      # add equivalent classes to queue (if not already cached)
      q.push ec unless c[ec]
      c[ec] = ec unless ec == term
    end

    # entail subclasses
    term.subClass.each do |sc|
      # add subclasses to queue (if not already cached)
      q.push sc unless c[sc]
      c[sc] = sc unless sc == term
    end
  end

  # smush the result 
  c.keys
end

#authors_for(repo, subject, unique: false, contrib: false, base: nil) ⇒ RDF::Value, Array

Assuming the subject is a thing that has authors, return the list of authors. Try bibo:authorList first for an explicit ordering, then continue to the various other predicates.

Parameters:

  • repo (RDF::Queryable)
  • subject (RDF::Resource)
  • unique (false, true) (defaults to: false) —

    only return the first author

  • contrib (false, true) (defaults to: false) —

    return contributors instead of authors

Returns:

  • (RDF::Value, Array) 


1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
 
# File 'lib/rdf/sak/util.rb', line 1137

def authors_for repo, subject, unique: false, contrib: false, base: nil
  authors = []

  # try the author list
  lp = [RDF::Vocab::BIBO[contrib ? :contributorList : :authorList]]
  lp += lp.first.entail(:equivalentProperty) # XXX cache this
  lp.each do |pred|
    o = repo.first_object([subject, pred, nil])
    next unless o
    # note this use of RDF::List is not particularly well-documented
    authors += RDF::List.from(repo, o).to_a
  end

  # now try various permutations of the author/contributor predicate
  unsorted = []
  preds = contrib ? CONTRIB : AUTHOR
  preds.each do |pred|
    unsorted += repo.query([subject, pred, nil]).objects
  end

  # prefetch the author names
  labels = authors.map { |a| [a, label_for(repo, a)] }.to_h

  authors += unsorted.uniq.sort { |a, b| labels[a] <=> labels[b] }

  unique ? authors.first : authors.uniq
end

#cmp_resource(a, b, www: nil) ⇒ Object 



866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
 
# File 'lib/rdf/sak/util.rb', line 866

def cmp_resource a, b, www: nil
  raise 'Comparands must be instances of RDF::Value' unless
    [a, b].all? { |x| x.is_a? RDF::Value }

  # URI beats non-URI
  if a.uri?
    if b.uri?
      # https beats http beats other
      as = a.scheme.downcase.to_sym
      bs = b.scheme.downcase.to_sym
      cmp = SCHEME_RANK.fetch(as, 2) <=> SCHEME_RANK.fetch(bs, 2)

      # bail out early
      return cmp unless cmp == 0

      # this would have returned if the schemes were different, as
      # such we only need to test one of them
      if [:http, :https].any?(as) and not www.nil?
        # if www is non-nil, prefer www or no-www depending on
        # truthiness of `www` parameter
        pref = [false, true].zip(www ? [1, 0] : [0, 1]).to_h
        re = /^(?:(www)\.)?(.*?)$/

        ah = re.match(a.host.to_s.downcase)[1,2]
        bh = re.match(b.host.to_s.downcase)[1,2]

        # compare hosts sans www
        cmp = ah[1] <=> bh[1]
        return cmp unless cmp == 0

        # now compare presence of www
        cmp = pref[ah[0] == 'www'] <=> pref[bh[0] == 'www']
        return cmp unless cmp == 0

        # if we're still here, compare the path/query/fragment
        re = /^.*?\/\/.*?(\/.*)$/
        al = re.match(a.to_s)[1].to_s
        bl = re.match(b.to_s)[1].to_s

        return al <=> bl
      end

      return a <=> b
    else
      return -1
    end
  elsif b.uri?
    return 1
  else
    return a <=> b
  end
end

#coerce_node_spec(spec, rev: false) ⇒ Object 



409
410
411
412
413
414
415
416
417
418
 
# File 'lib/rdf/sak/util.rb', line 409

def coerce_node_spec spec, rev: false
  spec = [spec] unless spec.respond_to? :to_a
  spec = spec - [:resource] + [:uri, :blank] if spec.include? :resource
  raise 'Subjects are never literals' if rev and spec.include? :literal

  spec = NMAP.values_at(*spec).reject(&:nil?).uniq
  spec = NTESTS.keys if spec.empty?
  spec.delete :literal if rev
  spec.uniq
end

#coerce_resource(arg, base = nil, as: :rdf) ⇒ URI, ...

Coerce a stringlike argument into a URI. Raises an exception if the string can’t be turned into a valid URI. Optionally resolves against a base, and the coercion can be tuned to either URI or RDF::URI via :as.

Parameters:

  • arg (URI, RDF::URI, #to_s) —

    The input string

  • base (URI, RDF::URI, #to_s) (defaults to: nil) —

    The optional base URI

  • as (:rdf, :uri, nil) (defaults to: :rdf) —

    The optional coercion type

Returns:

  • (URI, RDF::URI, String)

Raises:

  • (ArgumentError) 


1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
 
# File 'lib/rdf/sak/util.rb', line 1487

def coerce_resource arg, base = nil, as: :rdf
  as = assert_uri_coercion as
  return arg if as and arg.is_a?({ uri: URI, rdf: RDF::URI }[as])
  raise ArgumentError, 'arg must be stringable' unless arg.respond_to? :to_s

  arg = arg.to_s.strip

  if arg.start_with? '_:' and as
    # override the coercion if this is a blank node
    as = :rdf
  elsif base
    begin
      arg = (base.is_a?(URI) ? base : URI(uri_pp base.to_s.strip)).merge arg
    rescue URI::InvalidURIError => e
      warn "attempted to coerce #{arg} which turned out to be invalid: #{e}"
      return
    end
  end

  URI_COERCIONS[as].call arg
end

#coerce_uuid_urn(arg, base = nil) ⇒ Object

Coerce a stringlike argument into a UUID URN. Will

Raises:

  • (ArgumentError) 


1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
 
# File 'lib/rdf/sak/util.rb', line 1510

def coerce_uuid_urn arg, base = nil
  # if this is an ncname then change it
  if ([URI, RDF::URI] & arg.class.ancestors).empty? &&
      arg.respond_to?(:to_s)
    arg = arg.to_s

    # coerce ncname to uuid
    arg = UUID::NCName::from_ncname(arg, version: 1) if arg =~
      /^[A-P](?:[0-9A-Z_-]{20}|[2-7A-Z]{24})[A-P]$/i

    # now the string is either a UUID or it isn't
    arg = "urn:uuid:#{arg}" unless arg.start_with? 'urn:uuid:'
  else
    arg = arg.class.new arg.to_s.downcase unless arg == arg.to_s.downcase
  end

  raise ArgumentError, 'not a UUID' unless
    arg.to_s =~ /^urn:uuid:[0-9a-f]{8}(?:-[0-9a-f]{4}){4}[0-9a-f]{8}$/

  arg = coerce_resource arg, base
end

#dehydrate(doc) ⇒ Object

Strip all the links surrounding and RDFa attributes off dfn/abbr/span tags. Assuming a construct like <a rel=“some:relation” href=“#…” typeof=“skos:Concept”><dfn property=“some:property”>Term</dfn></a> is a link to a glossary entry, this method returns the term back to an undecorated state (+<dfn>Term</dfn>+).



1823
1824
1825
1826
1827
1828
1829
1830
 
# File 'lib/rdf/sak/util.rb', line 1823

def dehydrate doc
  doc.xpath(XPATH[:dehydrate], XPATHNS).each do |e|
    e = e.replace e.elements.first.dup
    %w[about resource typeof rel rev property datatype].each do |a|
      e.delete a if e.key? a
    end
  end
end

#formats_for(repo, subject, predicate: RDF::Vocab::DC.format, datatype: [RDF::XSD.token]) ⇒ Array

Obtain any specified MIME types for the subject. Just shorthand for a common application of ‘objects_for`.

Parameters:

  • repo
  • subject
  • predicate (defaults to: RDF::Vocab::DC.format)
  • datatype (defaults to: [RDF::XSD.token])

Returns:

  • (Array) —

    of internet media types



1260
1261
1262
1263
1264
1265
1266
1267
 
# File 'lib/rdf/sak/util.rb', line 1260

def formats_for repo, subject, predicate: RDF::Vocab::DC.format,
    datatype: [RDF::XSD.token]
  objects_for(
    repo, subject, predicate, only: :literal, datatype: datatype) do |o|
    t = o.object
    t =~ /\// ? RDF::SAK::MimeMagic.new(t.to_s.downcase) : nil
  end.compact.sort.uniq
end

#get_base(elem, default: nil, coerce: nil) ⇒ nil, ...

Returns the base URI from the perspective of the given element. Can optionally be coerced into either a URI or RDF::URI. Also takes a default value.

Parameters:

  • elem (Nokogiri::XML::Node) —

    the context element

  • default (nil, #to_s) (defaults to: nil) —

    the default URI

  • coerce (nil, :uri, :rdf) (defaults to: nil) —

    the coercion scheme, if any

Returns:

  • (nil, String, URI, RDF::URI) —

    the context’s base URI



1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
 
# File 'lib/rdf/sak/util.rb', line 1698

def get_base elem, default: nil, coerce: nil
  assert_uri_coercion coerce

  if elem.document?
    elem = elem.root
    return unless elem
  end

  # get the xpath
  xpath = (elem.namespace && elem.namespace.href == XHTMLNS or
    elem.at_xpath('/html')) ? :htmlbase : :xmlbase

  # now we go looking for the attribute
  if base = elem.at_xpath(XPATH[xpath], XPATHNS)
    base = base.value.strip
  else
    base = default.to_s.strip if default
  end

  # clear it out if it's the empty string
  base = nil if base and base.empty?

  # eh that's about all the input sanitation we're gonna get
  base && coerce ? URI_COERCIONS[coerce].call(base) : base
end

#get_prefixes(elem, traverse: true, coerce: nil, descend: false) ⇒ Hash

Note:

The descend: true parameter assumes we are trying to collect all the namespaces in use in the entire subtree, rather than resolve any particular CURIE. As such, the first prefix mapping in document order is preserved over subsequent/descendant ones.

Given an X(HT)ML element, returns a hash of prefixes of the form { prefix: “vocab” }, where the current @vocab is represented by the nil key. An optional :traverse parameter can be set to false to prevent ascending the node tree. Any XML namespace declarations are superseded by the @prefix attribute. Returns any @vocab declaration found as the nil key.

Parameters:

  • elem (Nokogiri::XML::Node) —

    The context element

  • traverse (true, false) (defaults to: true) —

    whether or not to traverse the tree

  • coerce (nil, :rdf, :uri) (defaults to: nil) —

    a type coercion for the URIs, if any

  • descend (false, true) (defaults to: false) —

    go down the tree instead of up

Returns:

  • (Hash) —

    Depending on :traverse, either all prefixes merged, or just the ones asserted in the element.



1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
 
# File 'lib/rdf/sak/util.rb', line 1743

def get_prefixes elem, traverse: true, coerce: nil, descend: false
  coerce = assert_uri_coercion coerce

  # deal with a common phenomenon
  elem = elem.root if elem.is_a? Nokogiri::XML::Document

  # get namespace definitions first
  prefix = elem.namespaces.reject do |k, _| k == 'xmlns'
  end.transform_keys { |k| k.split(?:)[1].to_sym }

  # now do the prefix attribute
  if elem.key? 'prefix'
    # XXX note this assumes largely that the input is clean
    elem['prefix'].strip.split.each_slice(2) do |k, v|
      pfx = k.split(?:)[0] or next # otherwise error
      prefix[pfx.to_sym] = v
    end
  end

  # encode the vocab as the null prefix
  if vocab = elem['vocab']
    vocab.strip!
    # note that a specified but empty @vocab means kill any existing vocab
    prefix[nil] = vocab.empty? ? nil : vocab
  end

  # don't forget we can coerce
  prefix.transform_values! { |v| COERCIONS[coerce].call v } if coerce

  # don't proceed if `traverse` is false
  return prefix unless traverse

  # save us having to recurse in ruby by using xpath implemented in c
  xpath = '%s::*[namespace::*|@prefix|@vocab]' %
    (descend ? :descendant : :ancestor)
  elem.xpath(xpath).each do |e|
    # this will always merge our prefix on top irrespective of direction
    prefix = get_prefix(e, traverse: false, coerce: coerce).merge prefix
  end

  prefix
end

#invert_struct(struct) ⇒ Object 



2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
 
# File 'lib/rdf/sak/util.rb', line 2000

def invert_struct struct
  nodes = {}

  struct.each do |p, v|
    v.each do |o|
      nodes[o] ||= Set.new
      nodes[o] << p
    end
  end

  nodes
end

#modernize(doc) ⇒ Object 



1809
1810
1811
1812
1813
1814
 
# File 'lib/rdf/sak/util.rb', line 1809

def modernize doc
  doc.xpath(XPATH[:modernize], XPATHNS).each do |e|
    # gotta instance_exec because `markup` is otherwise unbound
    instance_exec e, &MODERNIZE[e.name.to_sym]
  end
end

#node_matches?(node, spec) ⇒ Boolean

Returns:

  • (Boolean) 


420
421
422
 
# File 'lib/rdf/sak/util.rb', line 420

def node_matches? node, spec
  spec.any? { |k| node.send NTESTS[k] }
end

#predicate_set(predicates, seen: Set.new) ⇒ Array

Obtain the objects for a given subject-predicate pair.

Parameters:

  • subject (RDF::Resource)
  • predicate (RDF::URI)
  • entail (false, true)
  • only (:uri, :iri, :resource, :blank, :bnode, :literal)
  • datatype (RDF::Term)

Returns:

  • (Array) 


532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
 
# File 'lib/rdf/sak/util.rb', line 532

def predicate_set predicates, seen: Set.new
  predicates = Set[predicates] if predicates.is_a? RDF::URI
  unless predicates.is_a? Set
    raise "predicates must be a set" unless predicates.respond_to? :to_set
    predicates = predicates.to_set
    end

  # shortcut
  return predicates if predicates.empty?

  raise 'predicates must all be RDF::URI' unless predicates.all? do |p|
    p.is_a? RDF::URI
  end

  # first we generate the set of equivalent properties for the given
  # properties
  predicates += predicates.map do |p|
    p.entail :equivalentProperty
  end.flatten.to_set

  # then we take the resulting set of properties and
  # compute their subproperties
  subp = Set.new
  (predicates - seen).each do |p|
    subp += p.subProperty.flatten.to_set
  end

  # uhh this whole "seen" business might not be necessary
  predicates + predicate_set(subp - predicates - seen, seen: predicates)
end

#prefix_subset(prefixes, nodes) ⇒ Hash

Given a hash of prefixes and an array of nodes, obtain the the subset of prefixes that abbreviate the nodes. Scans RDF URIs as well as RDF::Literal datatypes.

Parameters:

  • prefixes (#to_h) —

    The prefixes, of the form { k: “v” }

  • nodes (Array<RDF::Term>) —

    The nodes to supply

Returns:

  • (Hash) —

    The prefix subset



1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
 
# File 'lib/rdf/sak/util.rb', line 1961

def prefix_subset prefixes, nodes
  prefixes = sanitize_prefixes prefixes, true

  raise 'nodes must be arrayable' unless nodes.respond_to? :to_a

  # sniff out all the URIs and datatypes
  resources = Set.new
  nodes.each do |n|
    next unless n.is_a? RDF::Term
    if n.literal? && n.datatype?
      resources << n.datatype
    elsif n.uri?
      resources << n
    end
  end

  # now we abbreviate all the resources
  pfx = abbreviate(resources.to_a,
    prefixes: prefixes, noop: false, sort: false).uniq.compact.map do |p|
    p.split(?:).first.to_sym
  end.uniq.to_set

  # now we return the subset 
  prefixes.select { |k, _| pfx.include? k.to_sym }
end

#prepare_collation(struct) {|p, o| ... } ⇒ Hash

Given a structure of the form { predicate => [objects] }, rearrange the structure into one more amenable to rendering RDFa. Returns a hash of the form { resources: { r1 => Set[p1, pn] }, literals: { l1 => Set[p2, pm] }, types: Set[t1, tn], datatypes: Set[d1, dn] }. This inverted structure can then be conveniently traversed to generate the RDFa. An optional block lets us examine the predicate-object pairs as they go by.

Parameters:

  • struct (Hash) —

    The struct of the designated form

Yields:

  • (p, o) —

    An optional block is given the predicate-object pair

Returns:

  • (Hash) —

    The inverted structure, as described.



1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
 
# File 'lib/rdf/sak/util.rb', line 1922

def prepare_collation struct, &block
  resources = {}
  literals  = {}
  datatypes = Set.new
  types     = Set.new

  struct.each do |p, v|
    v.each do |o|
      block.call p, o if block

      if o.literal?
        literals[o] ||= Set.new
        literals[o].add p
        # collect the datatype
        datatypes.add o.datatype if o.has_datatype?
      else
        if  p == RDF::RDFV.type
          # separate the type
          types.add o
        else
          # collect the resource
          resources[o] ||= Set.new
          resources[o].add p
        end
      end
    end
  end

  { resources: resources, literals: literals,
    datatypes: datatypes, types: types }
end

#rehydrate(doc, graph, &block) ⇒ Object

(maybe add code/kbd/samp/var/time one day too)



1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
 
# File 'lib/rdf/sak/util.rb', line 1842

def rehydrate doc, graph, &block
  doc.xpath(XPATH[:rehydrate], XPATHNS).each do |e|
    lang = e.xpath(XPATH[:lang]).to_s.strip
    # dt   = e['datatype'] # XXX no datatype rn
    text = (e['content'] || e.xpath('.//text()').to_a.join).strip

    # now we have the literal
    lit = [RDF::Literal(text)]
    lit.unshift RDF::Literal(text, language: lang) unless lang.empty?

    # candidates
    cand = {}
    lit.map do |t|
      graph.query(object: t).to_a
    end.flatten.each do |x|
      y = cand[x.subject] ||= {}
      (y[:stmts] ||= []) << x
      y[:types]  ||= graph.query([x.subject, RDF.type, nil]).objects.sort
    end

    # if there's only one candidate, this is basically a noop
    chosen = cand.keys.first if cand.size == 1

    # call the block to reconcile any gaps or conflicts
    if block_given? and cand.size != 1
      # the block is expected to return one of the candidates or
      # nil. we call the block with the graph so that the block can
      # manipulate its contents.
      chosen = block.call cand, graph
      raise ArgumentError, 'block must return nil or a term' unless
        chosen.nil? or chosen.is_a? RDF::Term
    end

    if chosen
      # we assume this has been retrieved from the graph
      cc = cand[chosen]
      unless cc
        cc = cand[chosen] = {}
        cc[:stmts] = graph.query([chosen, nil, lit[0]]).to_a.sort
        cc[:types] = graph.query([chosen, RDF.type, nil]).objects.sort
        # if either of these are empty then the graph was not
        # appropriately populated
        raise 'Missing a statement relating #{chosen} to #{text}' if
          cc[:stmts].empty?
      end

      # we should actually probably move any prefix/vocab/xmlns
      # declarations from the inner node to the outer one (although
      # in practice this will be an unlikely configuration)
      pfx = get_prefixes e

      # here we have pretty much everything except for the prefixes
      # and wherever we want to actually link to.

      inner = e.dup
      spec  = { [inner] => :a, href: '' }
      # we should have types
      spec[:typeof] = abbreviate cc[:types], prefixes: pfx unless
        cc[:types].empty?

      markup replace: e, spec: spec
    end
  end
  # return maybe the elements that did/didn't get changed?
end

#reindent(node, depth = 0, indent = ' ') ⇒ Object

reindent text nodes



1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
 
# File 'lib/rdf/sak/util.rb', line 1318

def reindent node, depth = 0, indent = '  '
  kids = node.children
  if kids and child = kids.first
    loop do
      if child.element?
        # recurse into the element
        reindent child, depth + 1, indent
      elsif child.text?
        text = child.content || ''

        # optional horizontal whitespace followed by at least
        # one newline (we don't care what kind), followed by
        # optional horizontal or vertical whitespace
        preamble = !!text.gsub!(/\A[ \t]*[\r\n]+\s*/, '')

        # then we don't care what's in the middle, but hey let's get
        # rid of dos newlines because we can always put them back
        # later if we absolutely have to
        text.gsub!(/\r+/, '')

        # then optionally any whitespace followed by at least
        # another newline again, followed by optional horizontal
        # whitespace and then the end of the string
        epilogue = !!text.gsub!(/\s*[\r\n]+[ \t]*\z/, '')

        # if we prune these off we'll have a text node that is
        # either the empty string or it isn't (note we will only
        # register an epilogue if the text has some non-whitespace
        # in it, because otherwise the first regex would have
        # snagged everything, so it's probably redundant)

        # if it's *not* empty then we *prepend* indented whitespace
        if preamble and !text.empty?
          d = depth + (child.previous ? 1 : 0)
          text = "\n" + (indent * d) + text
        end
 
        # then we unconditionally *append*, (modulo there being a
        # newline in the original at all), but we have to check by
        # how much: if this is *not* the last node then depth + 1,
        # otherwise depth
        if preamble or epilogue
          d = depth + (child.next ? 1 : 0)
          text << "\n" + (indent * d)
        end

        child.content = text
      end

      break unless child = child.next
    end
  end

  node
end

#resolve_curie(curie, prefixes: {}, vocab: nil, base: nil, refnode: nil, term: false, noop: true, scalar: false, coerce: nil) ⇒ nil, ...

Note:

:vocab overrides, and is the same as supplying prefix[nil]. It is only meaningful when :term (i.e., when we expect the input to be an RDFa term) is true.

Resolve a string or array or attribute node containing one or more terms/CURIEs against a set of prefixes. The CURIE can be a string, Nokogiri::XML::Attr, or an array thereof. Strings are stripped and split on whitespace. :prefixes and :base can be supplied or gleaned from :refnode, which itself can be gleaned if curie is a Nokogiri::XML::Attr. Returns an array of (attempted) resolved terms unless :scalar is true, in which case only the first URI is returned. When :noop is true, this method will always return a value. Can coerce results to either RDF::URI or URI objects.

Parameters:

  • curie (#to_s, Nokogiri::XML::Attr, Array) —

    One or more CURIEs

  • prefixes (#to_h) (defaults to: {}) —

    The hash of prefixes (nil key is equivalent to vocab)

  • vocab (nil, #to_s) (defaults to: nil) —

    An optional base URI

  • refnode (nil, Nokogiri::XML::Element) (defaults to: nil) —

    A reference node for resolution

  • term (false, true) (defaults to: false) —

    Whether to treat the input as an RDFa term

  • noop (true, false) (defaults to: true) —

    Whether to skip if the CURIE can’t be resolved

  • scalar (false, true) (defaults to: false) —

    Whether to return a scalar value

  • coerce (nil, :rdf, :uri) (defaults to: nil) —

    Desired type coercion for the output

Returns:

  • (nil, URI, RDF::URI, Array<nil,URI,RDF::URI>) 


1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
 
# File 'lib/rdf/sak/util.rb', line 1577

def resolve_curie curie, prefixes: {}, vocab: nil, base: nil,
    refnode: nil, term: false, noop: true, scalar: false, coerce: nil
  prefixes = sanitize_prefixes prefixes

  raise 'coerce must be either :uri or :rdf' if coerce and
    not %i[uri rdf].include?(coerce)

  # coerce curie to its value and set refnode if not present
  if curie.is_a? Nokogiri::XML::Attr
    refnode ||= curie.parent
    curie = curie.value.strip.split
  elsif curie.respond_to? :to_a
    curie = curie.to_a
    raise ArgumentError,
      'if curie is an array, it has to be all strings' unless
      curie.all? { |x| x.respond_to? :to_s }
    curie = curie.map { |x| x.to_s.strip.split }.flatten
  else
    raise ArgumentError, 'curie must be stringable' unless
      curie.respond_to? :to_s
    curie = curie.to_s.strip.split
  end

  if refnode
    raise ArgumentError, 'refnode must be an element' unless
      refnode.is_a? Nokogiri::XML::Element
    prefixes = get_prefixes refnode if prefixes.empty?
  end

  # now we overwrite the vocab
  if vocab
    raise ArgumentError, 'vocab must be stringable' unless
      vocab.respond_to? :to_s
    prefixes[nil] = vocab.to_s.strip
  end

  out = curie.map do |c|
    prefix, slug = /^\[?(?:([^:]+):)?(.*?)\]?$/.match(c).captures
    prefix = prefix.to_sym if prefix
    tmp = if prefixes[prefix]
            prefixes[prefix] + slug
          else
            noop ? c : nil
          end
    tmp && coerce ? URI_COERCIONS[coerce].call(tmp) : tmp
  end

  scalar ? out.first : out
end

#smush_struct(struct) ⇒ Object

turns any data structure into a set of nodes



1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
 
# File 'lib/rdf/sak/util.rb', line 1988

def smush_struct struct
  out = Set.new

  if struct.is_a? RDF::Term
    out << struct
  elsif struct.respond_to? :to_a
    out |= struct.to_a.map { |s| smush_struct(s).to_a }.flatten.to_set
  end

  out
end

#split_pp(uri, only: false) ⇒ Array

Given a URI as input, split any path parameters out of the last path segment. Works the same way as #split_pp.

Parameters:

  • uri (URI, #to_s) —

    The URI to extract parameters from

  • only (false, true) (defaults to: false) —

    whether to only return the parameters

Returns:

  • (Array) —

    (See description)



1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
 
# File 'lib/rdf/sak/util.rb', line 1437

def split_pp uri, only: false
  begin
    u = (uri.is_a?(URI) ? uri : URI(uri_pp uri.to_s)).normalize

  rescue URI::InvalidURIError => e
    # these stock error messages don't even tell you what the uri is
    raise URI::InvalidURIError, "#{e.message} (#{uri.to_s})"
  end

  return only ? [] : [uri] unless u.path
  uri = u

  ps = uri.path.split '/', -1
  pp = ps.pop.split ';', -1
  bp = (ps + [pp.shift]).join '/'
  uri = uri.dup

  begin
    uri.path = bp
  rescue URI::InvalidURIError => e
    # these stock error messages don't even tell you what the uri is
    m = e.message
    raise URI::InvalidURIError, "#{m} (#{uri.to_s}, #{bp})"
  end

  return pp if only
  [uri] + pp
end

#split_pp2(path, only: false) ⇒ Object 



1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
 
# File 'lib/rdf/sak/util.rb', line 1466

def split_pp2 path, only: false
  # ugh apparently we need a special case for ''.split
  return only ? [] : [''] if !path or path.empty?

  ps = path.to_s.split ?/, -1    # path segments
  pp = ps.pop.to_s.split ?;, -1  # path parameters
  bp = (ps + [pp.shift]).join ?/ # base path

  only ? pp : [bp] + pp
end

#split_qp(uri, only: false) ⇒ Array

Given a URI as input, split any query parameters into an array of key-value pairs. If :only is true, this will just return the pairs. Otherwise it will prepend the query-less URI to the array, and can be captured with an idiom like uri, *qp = split_qp uri.

Parameters:

  • uri (URI, #to_s) —

    The URI to extract parameters from

  • only (false, true) (defaults to: false) —

    whether to only return the parameters

Returns:

  • (Array) —

    (See description)



1422
1423
1424
1425
1426
1427
1428
 
# File 'lib/rdf/sak/util.rb', line 1422

def split_qp uri, only: false
  uri = URI(uri_pp uri.to_s) unless uri.is_a? URI
  qp  = URI::decode_www_form(uri.query)
  return qp if only
  uri.query = nil
  [uri] + qp
end

#subject_for(node, prefixes: nil, base: nil, coerce: :rdf) ⇒ URI, ...

Given an X(HT)ML element, return the nearest RDFa subject. Optionally takes :prefix and :base parameters which override anything found in the document tree.

Parameters:

  • node (Nokogiri::XML::Element) —

    the node

  • prefixes (Hash) (defaults to: nil) —

    Prefix mapping. Overrides derived values.

  • base (#to_s, URI, RDF::URI) (defaults to: nil) —

    Base URI, overrides as well.

  • coerce (nil, :rdf, :uri) (defaults to: :rdf) —

    the coercion regime

Returns:

  • (URI, RDF::URI, String) —

    the subject



1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
 
# File 'lib/rdf/sak/util.rb', line 1797

def subject_for node, prefixes: nil, base: nil, coerce: :rdf
  assert_xml_node node
  coerce = assert_uri_coercion coerce

  if n = node.at_xpath(XPATH[:literal])
    return internal_subject_for n,
      prefixes: prefixes, base: base, coerce: coerce
  end

  internal_subject_for node, prefixes: prefixes, base: base, coerce: coerce
end

#subtree(doc, xpath = '/*', reindent: true, prefixes: {}) ⇒ Object

isolate an element into a new document



1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
 
# File 'lib/rdf/sak/util.rb', line 1297

def subtree doc, xpath = '/*', reindent: true, prefixes: {}
  # at this time we shouldn't try to do anything cute with the xpath
  # even though it is attractive to want to prune out prefixes

  # how about we start with a noop
  return doc.root.dup if xpath == '/*'

  begin
    nodes = doc.xpath xpath, prefixes
    return unless
      nodes and nodes.is_a?(Nokogiri::XML::NodeSet) and !nodes.empty?
    out = Nokogiri::XML::Document.new
    out << nodes.first.dup
    reindent out.root if reindent
    out
  rescue Nokogiri::SyntaxError
    return
  end
end

#terminal_slug(uri, base: nil) ⇒ String

Get the last non-empty path segment of the URI

Parameters:

  • uri

Returns:

  • (String) 


1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
 
# File 'lib/rdf/sak/util.rb', line 1537

def terminal_slug uri, base: nil
  uri = coerce_resource uri, base
  return unless uri.respond_to? :path
  if p = uri.path
    if p = /^\/+(.*?)\/*$/.match(p)
      if p = p[1].split(/\/+/).last
        # we need to escape colons or it will think it's absolute
        return uri_pp(p.split(/;+/).first || '', ':')
      end
    end        
  end
  ''
end

#title_tag(predicates, content, prefixes: {}, vocab: nil, lang: nil, xhtml: true) ⇒ Object 



2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
 
# File 'lib/rdf/sak/util.rb', line 2013

def title_tag predicates, content,
    prefixes: {}, vocab: nil, lang: nil, xhtml: true

  # begin with the tag
  tag = { '#title' => content.to_s,
    property: abbreviate(predicates, prefixes: prefixes, vocab: vocab) }

  # we set the language if it exists and is different from the
  # body OR if it is xsd:string we set it to the empty string
  lang = (content.language? && content.language != lang ?
    content.language : nil) || (content.datatype == RDF::XSD.string &&
    lang ? '' : nil)
  if lang
    tag['xml:lang'] = lang if xhtml
    tag[:lang] = lang
  end
  if content.datatype? && content.datatype != RDF::XSD.string
    tag[:datatype] = abbreviate(content.datatype,
      prefixes: prefixes, vocab: vocab)
  end

  tag
end

#type_strata(rdftype) ⇒ Array

Obtain a stack of types for an asserted initial type or set thereof. Returns an array of arrays, where the first is the asserted types and their inferred equivalents, and subsequent elements are immediate superclasses and their equivalents. A given URI will only appear once in the entire structure.

Parameters:

  • rdftype (RDF::Term, :to_a)

Returns:

  • (Array) 


458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
 
# File 'lib/rdf/sak/util.rb', line 458

def type_strata rdftype
  # first we coerce this to an array
  if rdftype.respond_to? :to_a
    rdftype = rdftype.to_a
  else
    rdftype = [rdftype]
  end
    
  # now squash and coerce
  rdftype = rdftype.uniq.map { |t| RDF::Vocabulary.find_term t }.compact

  # bail out early
  return [] if rdftype.empty?

  # essentially what we want to do is construct a layer of
  # asserted classes and their inferred equivalents, then probe
  # the classes in the first layer for subClassOf assertions,
  # which will form the second layer, and so on.

  queue  = [rdftype]
  strata = []
  seen   = Set.new

  while qin = queue.shift
    qwork = []

    qin.each do |q|
      qwork << q # entail doesn't include q
      qwork += q.entail(:equivalentClass) if q.uri?
    end

    # grep and flatten
    qwork = qwork.map do |t|
      next t if t.is_a? RDF::Vocabulary::Term
      RDF::Vocabulary.find_term t
    end.compact.uniq - seen.to_a
    seen |= qwork

    # warn "qwork == #{qwork.inspect}"

    # push current layer out
    strata.push qwork.dup unless qwork.empty?
   
    # now deal with subClassOf
    qsuper = []
    qwork.each { |q| qsuper += q.subClassOf }

    # grep and flatten this too
    qsuper = qsuper.map do |t|
      next t if t.is_a? RDF::Vocabulary::Term
      RDF::Vocabulary.find_term t
    end.compact.uniq - seen.to_a
    # do not append qsuper to seen!

    # warn "qsuper == #{qsuper.inspect}"

    # same deal, conditionally push the input queue
    queue.push qsuper.dup unless qsuper.empty?
  end

  # voila
  strata
end

#uri_pp(uri, extra = '') ⇒ Object

really gotta stop carting this thing around



1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
 
# File 'lib/rdf/sak/util.rb', line 1393

def uri_pp uri, extra = ''
  # take care of malformed escapes
  uri = uri.to_s.b.gsub(/%(?![0-9A-Fa-f]{2})/n, '%25')
  uri.gsub!(/([#{Regexp.quote extra}])/) do |s|
    sprintf('%%%02X', s.ord)
  end unless extra.empty?
  # we want the minimal amount of escaping so we split out the separators
  out = ''
  parts = RFC3986.match(uri).captures
  parts.each_index do |i|
    next if parts[i].nil?
    out << SEPS[i].first
    out << parts[i].b.gsub(SF) { |s| sprintf('%%%02X', s.ord) }
    out << SEPS[i].last
  end

  # make sure escaped hex is upper case like the rfc says
  out.gsub(/(%[0-9A-Fa-f]{2})/) { |x| x.upcase }
end