tao_rdfizer 0.11.2 → 0.11.4
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/lib/tao_rdfizer/tao_rdfizer.rb +44 -56
- metadata +4 -4
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA256:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 59d7999a921ef33104728a1f86edbeb7f20161641b37a53435daf316772ffed8
|
4
|
+
data.tar.gz: d6cd66e38eececa7b7cbb251686d148a4924da16ad751eea3af2cef21b00121a
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 4e98ac931a0f2913f44e8e234222f55c93ff9bbd76173787393fcf1e866a14b7707168c8446f33cfd00f208599439cacdd67b4b6296c9ab8a1288a3963fd4f0c
|
7
|
+
data.tar.gz: 788aaf5c427d44783edadfc30f84a8d2cc7aac99641575dd3999f780520c374cf5c443fc9a5dd656424a4291940861fa268a37b6ca4f3dab9a2ae196f6df2bc8
|
@@ -14,14 +14,14 @@ class TAO::RDFizer
|
|
14
14
|
else
|
15
15
|
ERB_ANNOTATIONS_TTL
|
16
16
|
end
|
17
|
-
@tao_ttl_erb = ERB.new(template,
|
18
|
-
@prefix_ttl_erb = ERB.new(ERB_PREFIXES_TTL,
|
17
|
+
@tao_ttl_erb = ERB.new(template, trim_mode: '-')
|
18
|
+
@prefix_ttl_erb = ERB.new(ERB_PREFIXES_TTL, trim_mode: '-')
|
19
19
|
end
|
20
20
|
|
21
21
|
def rdfize(annotations_col, options = nil)
|
22
22
|
options ||= {}
|
23
|
-
only_prefixes = options
|
24
|
-
with_prefixes = options
|
23
|
+
only_prefixes = options[:only_prefixes] == true
|
24
|
+
with_prefixes = options[:with_prefixes] != true
|
25
25
|
|
26
26
|
# check the format
|
27
27
|
annotations_col.each do |annotations|
|
@@ -29,35 +29,33 @@ class TAO::RDFizer
|
|
29
29
|
end
|
30
30
|
|
31
31
|
# namespaces
|
32
|
-
namespaces =
|
33
|
-
|
34
|
-
anns = annotations_col.first
|
35
|
-
anns[:namespaces].each {|n| namespaces[n[:prefix]] = n[:uri]} unless anns[:namespaces].nil?
|
32
|
+
namespaces = get_namespaces(annotations_col.first)
|
36
33
|
|
37
34
|
prefixes_ttl = @prefix_ttl_erb.result_with_hash(namespaces:namespaces) if only_prefixes || with_prefixes
|
38
35
|
|
39
|
-
if only_prefixes
|
40
|
-
|
41
|
-
|
42
|
-
|
43
|
-
if with_prefixes
|
44
|
-
prefixes_ttl + annotations_ttl
|
45
|
-
else
|
46
|
-
annotations_ttl
|
47
|
-
end
|
48
|
-
end
|
36
|
+
return prefixes_ttl if only_prefixes
|
37
|
+
|
38
|
+
annotations_ttl = get_annotations_ttl(annotations_col, namespaces)
|
39
|
+
with_prefixes ? prefixes_ttl + annotations_ttl : annotations_ttl
|
49
40
|
end
|
50
41
|
|
51
42
|
private
|
52
43
|
|
53
|
-
def
|
54
|
-
|
44
|
+
def get_namespaces(annotations)
|
45
|
+
return {} if annotations[:namespaces].nil?
|
46
|
+
|
47
|
+
annotations[:namespaces].each_with_object({}) do |n, namespaces|
|
48
|
+
namespaces[n[:prefix]] = n[:uri]
|
49
|
+
end
|
50
|
+
end
|
55
51
|
|
52
|
+
def get_annotations_ttl(annotations_col, namespaces)
|
56
53
|
unless @mode ==:spans
|
54
|
+
anns = annotations_col.first
|
57
55
|
raise ArgumentError, "A project name has to be specified." unless anns.has_key?(:project)
|
58
56
|
prefix_for_this = anns[:project].downcase.gsub(/ /, '_')
|
59
57
|
raise ArgumentError, "'#{prefix_for_this}' is a reserved prefix for this project." if namespaces.has_key?(prefix_for_this)
|
60
|
-
project_uri = '
|
58
|
+
project_uri = 'https://pubannotation.org/projects/' + anns[:project]
|
61
59
|
namespaces[prefix_for_this] = project_uri + '/'
|
62
60
|
end
|
63
61
|
|
@@ -70,8 +68,8 @@ class TAO::RDFizer
|
|
70
68
|
text = annotations[:text]
|
71
69
|
text_uri = annotations[:target]
|
72
70
|
text_id = begin
|
73
|
-
sourcedb, sourceid
|
74
|
-
|
71
|
+
sourcedb, sourceid = get_target_info(text_uri)
|
72
|
+
"#{sourcedb}-#{sourceid}"
|
75
73
|
end
|
76
74
|
|
77
75
|
# denotations and relations
|
@@ -153,37 +151,29 @@ class TAO::RDFizer
|
|
153
151
|
end
|
154
152
|
end
|
155
153
|
|
156
|
-
_spans.
|
154
|
+
_spans.sort_by! { |s| [s[:begin], -s[:end]] }
|
157
155
|
|
158
156
|
## begin indexing
|
159
157
|
len = text.length
|
160
158
|
num = _spans.length
|
161
159
|
|
162
|
-
#
|
163
|
-
|
164
|
-
|
165
|
-
|
166
|
-
end
|
160
|
+
# create the indexes for followings and children
|
161
|
+
_spans.each_with_index do |span, i|
|
162
|
+
span[:followings] = []
|
163
|
+
span[:children] = []
|
167
164
|
|
168
|
-
(0 ... num).each do |i|
|
169
|
-
# find the first following span
|
170
165
|
j = i + 1
|
171
|
-
|
172
|
-
|
173
|
-
unless include_parent?(_spans[i][:children], _spans[j])
|
174
|
-
_spans[i][:children] << _spans[j]
|
175
|
-
end
|
166
|
+
while j < num && _spans[j][:begin] < span[:end]
|
167
|
+
span[:children] << _spans[j] unless include_parent?(span[:children], _spans[j])
|
176
168
|
j += 1
|
177
169
|
end
|
178
170
|
|
179
|
-
|
180
|
-
fp = _spans[i][:end]
|
171
|
+
fp = span[:end]
|
181
172
|
fp += 1 while fp < len && text[fp].match(/\s/)
|
182
173
|
next if fp == len
|
183
174
|
|
184
|
-
# index adjacent spans
|
185
175
|
while j < num && _spans[j][:begin] == fp
|
186
|
-
|
176
|
+
span[:followings] << _spans[j]
|
187
177
|
j += 1
|
188
178
|
end
|
189
179
|
end
|
@@ -206,33 +196,31 @@ class TAO::RDFizer
|
|
206
196
|
def get_target_info (text_uri)
|
207
197
|
sourcedb = (text_uri =~ %r|/sourcedb/([^/]+)|)? $1 : nil
|
208
198
|
sourceid = (text_uri =~ %r|/sourceid/([^/]+)|)? $1 : nil
|
209
|
-
divid = (text_uri =~ %r|/divs/([^/]+)|)? $1 : nil
|
210
199
|
|
211
|
-
return sourcedb, sourceid
|
200
|
+
return sourcedb, sourceid
|
212
201
|
end
|
213
202
|
|
214
|
-
def find_uri
|
203
|
+
def find_uri(label, namespaces, prefix_for_this)
|
204
|
+
@uri_cache ||= {}
|
205
|
+
|
206
|
+
return @uri_cache[label] if @uri_cache.key?(label)
|
207
|
+
|
215
208
|
if label.match(/\s/)
|
216
|
-
|
217
|
-
label.gsub(/\s/, '_')
|
209
|
+
label.gsub!(/\s/, '_')
|
218
210
|
end
|
219
211
|
delimiter_position = label.index(':')
|
220
|
-
if
|
221
|
-
label.gsub('(', '\(').gsub(')', '\)')
|
212
|
+
if delimiter_position && namespaces.key?(label[0...delimiter_position])
|
213
|
+
@uri_cache[label] = label.gsub('(', '\(').gsub(')', '\)')
|
222
214
|
elsif label =~ %r[^https?://]
|
223
|
-
"<#{label}>"
|
215
|
+
@uri_cache[label] = "<#{label}>"
|
224
216
|
else
|
225
|
-
clabel =
|
226
|
-
|
227
|
-
else
|
228
|
-
label.sub(/^\W+/, '').sub(/[^a-zA-Z0-9_)]+$/, '').gsub(/ +/, '_')
|
229
|
-
end
|
230
|
-
namespaces.has_key?('_base') ? "<#{clabel}>" : "#{prefix_for_this}:#{clabel.gsub('(', '\(').gsub(')', '\)')}"
|
217
|
+
clabel = label.match(/^\W+$/) ? 'SYM' : label.sub(/^\W+/, '').sub(/[^a-zA-Z0-9_)]+$/, '').gsub(/ +/, '_')
|
218
|
+
@uri_cache[label] = namespaces.key?('_base') ? "<#{clabel}>" : "#{prefix_for_this}:#{clabel.gsub('(', '\(').gsub(')', '\)')}"
|
231
219
|
end
|
232
220
|
end
|
233
221
|
|
234
222
|
def rdf_literal_escape(string)
|
235
|
-
string.gsub('\\', '
|
223
|
+
string.gsub('\\', '\\\\\\').
|
236
224
|
gsub("\t", '\\t').
|
237
225
|
gsub("\b", '\\b').
|
238
226
|
gsub("\n", '\\n').
|
@@ -274,7 +262,7 @@ class TAO::RDFizer
|
|
274
262
|
# variable: namespaces
|
275
263
|
ERB_PREFIXES_TTL = <<~HEREDOC
|
276
264
|
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .
|
277
|
-
@prefix tao: <
|
265
|
+
@prefix tao: <https://pubannotation.org/ontology/tao.owl#> .
|
278
266
|
<%# namespaces -%>
|
279
267
|
<% namespaces.each_key do |p| -%>
|
280
268
|
<% if p == '_base' -%>
|
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: tao_rdfizer
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.11.
|
4
|
+
version: 0.11.4
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Jin-Dong Kim
|
8
8
|
autorequire:
|
9
9
|
bindir: bin
|
10
10
|
cert_chain: []
|
11
|
-
date:
|
11
|
+
date: 2024-09-18 00:00:00.000000000 Z
|
12
12
|
dependencies: []
|
13
13
|
description: It uses TAO (text annotation ontology) for representation of annotations
|
14
14
|
to text.
|
@@ -33,14 +33,14 @@ required_ruby_version: !ruby/object:Gem::Requirement
|
|
33
33
|
requirements:
|
34
34
|
- - ">="
|
35
35
|
- !ruby/object:Gem::Version
|
36
|
-
version:
|
36
|
+
version: 3.0.0
|
37
37
|
required_rubygems_version: !ruby/object:Gem::Requirement
|
38
38
|
requirements:
|
39
39
|
- - ">="
|
40
40
|
- !ruby/object:Gem::Version
|
41
41
|
version: '0'
|
42
42
|
requirements: []
|
43
|
-
rubygems_version: 3.
|
43
|
+
rubygems_version: 3.5.11
|
44
44
|
signing_key:
|
45
45
|
specification_version: 4
|
46
46
|
summary: A RDF statement generator for annotations in the PubAnnotation JSON format.
|