llt-review 0.0.2 → 0.0.3
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/lib/llt/review/api.rb +1 -1
- data/lib/llt/review/common/sentence.rb +9 -1
- data/lib/llt/review/treebank/postag.rb +9 -1
- data/lib/llt/review/treebank/sentence.rb +7 -0
- data/lib/llt/review/treebank/word.rb +5 -1
- data/lib/llt/review/version.rb +1 -1
- data/spec/lib/llt/review/treebank_spec.rb +111 -0
- metadata +3 -3
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA1:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 9b91983dc42ac4d0aece4d4d745e2576442f91fc
|
4
|
+
data.tar.gz: faa768dab56010e1bc570506e7ae718ca503a11b
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: e2f9e3f7d61e8ac1493035d8789eb414190cbbe3f218a3c229ab66af09d55f0f7df1d7ba2ae0c761e195caa6a103a9bd77081a8938adb88e0626bcf849a65ff3
|
7
|
+
data.tar.gz: 65f0e95aa16565761a0025f2832e91b2e8e22b4277ad35d625bc8ab63ad2072cad9d0c1acb8b321edddd7d0bb4b89e19d224129a243b7d329dbd4113b5e0d520
|
data/lib/llt/review/api.rb
CHANGED
@@ -44,7 +44,7 @@ class Api < Sinatra::Base
|
|
44
44
|
if backend = params[:backend]
|
45
45
|
p = case backend
|
46
46
|
when 'perseids' then 'sosol.perseids.org/sosol'
|
47
|
-
when 'perseids-dev' then '
|
47
|
+
when 'perseids-dev' then 'sosol-rep.perseids.org'
|
48
48
|
end
|
49
49
|
|
50
50
|
expand_perseids_urls(params, :gold, p)
|
@@ -13,7 +13,11 @@ module LLT
|
|
13
13
|
|
14
14
|
def compare(other, comparables = nil)
|
15
15
|
diff = new_sentence_diff
|
16
|
-
words.each
|
16
|
+
words.each do |id, word|
|
17
|
+
other_word = other[id] || dummy_word(id)
|
18
|
+
|
19
|
+
word.compare(other_word, diff, comparables)
|
20
|
+
end
|
17
21
|
diff
|
18
22
|
end
|
19
23
|
|
@@ -25,6 +29,10 @@ module LLT
|
|
25
29
|
|
26
30
|
private
|
27
31
|
|
32
|
+
def dummy_word(id)
|
33
|
+
# implemented by subclasses
|
34
|
+
end
|
35
|
+
|
28
36
|
def new_sentence_diff
|
29
37
|
diff_namespace.const_get(:Sentence).new(self)
|
30
38
|
end
|
@@ -5,7 +5,7 @@ module LLT
|
|
5
5
|
|
6
6
|
def initialize(postag)
|
7
7
|
@id = :postag
|
8
|
-
@postag = postag
|
8
|
+
@postag = new_postag(postag)
|
9
9
|
end
|
10
10
|
|
11
11
|
def to_s
|
@@ -58,6 +58,14 @@ module LLT
|
|
58
58
|
data.add(Report::Generic.new(pl, 0, sg))
|
59
59
|
end
|
60
60
|
end
|
61
|
+
|
62
|
+
def new_postag(postag)
|
63
|
+
if !postag || postag.empty? || postag == '-'
|
64
|
+
'-' * POSTAG_SCHEMA.length
|
65
|
+
else
|
66
|
+
postag
|
67
|
+
end
|
68
|
+
end
|
61
69
|
end
|
62
70
|
end
|
63
71
|
end
|
@@ -3,7 +3,7 @@ module LLT
|
|
3
3
|
class Word
|
4
4
|
include Core::Structures::HashContainable
|
5
5
|
|
6
|
-
attr_accessor :form, :lemma, :head, :relation
|
6
|
+
attr_accessor :form, :lemma, :head, :relation, :inexistant
|
7
7
|
attr_reader :postag
|
8
8
|
|
9
9
|
Attr = Struct.new(:id, :attribute) do
|
@@ -20,6 +20,10 @@ module LLT
|
|
20
20
|
define_method("#{type}=") { |val| add(Attr.new(type, val)) }
|
21
21
|
end
|
22
22
|
|
23
|
+
def [](key)
|
24
|
+
super(key) || send("#{key}=", '-')
|
25
|
+
end
|
26
|
+
|
23
27
|
def postag=(tag)
|
24
28
|
add(Postag.new(tag))
|
25
29
|
end
|
data/lib/llt/review/version.rb
CHANGED
@@ -141,6 +141,117 @@ describe LLT::Review::Treebank do
|
|
141
141
|
diff[3].should be_nil # has only lemma and postag differences
|
142
142
|
diff[4].should be_true
|
143
143
|
end
|
144
|
+
|
145
|
+
context "with slightly out of sync documents" do
|
146
|
+
let(:ge1) do
|
147
|
+
<<-EOF
|
148
|
+
<treebank>
|
149
|
+
<sentence id="21" document_id="Perseus:text:1999.02.0002" subdoc="Book=2:chapter=5" span="In3:erat0">
|
150
|
+
<word id="1" form="In" lemma="in1" postag="r--------" head="5" relation="AuxP"/>
|
151
|
+
<word id="2" form="eo" lemma="is1" postag="p-s---nb-" head="3" relation="ATR"/>
|
152
|
+
<word id="3" form="flumine" lemma="flumen1" postag="n-s---nb-" head="1" relation="ADV"/>
|
153
|
+
<word id="4" form="pons" lemma="pons1" postag="n-s---mn-" head="5" relation="SBJ"/>
|
154
|
+
<word id="5" form="erat" lemma="sum1" postag="v3siia---" head="0" relation="PRED"/>
|
155
|
+
<word id="6" form="erat" lemma="sum1" postag="v3siia---" head="0" relation="PRED" elliptic="true"/>
|
156
|
+
</sentence>
|
157
|
+
</treebank>
|
158
|
+
EOF
|
159
|
+
end
|
160
|
+
|
161
|
+
let(:re1) do
|
162
|
+
<<-EOF
|
163
|
+
<treebank>
|
164
|
+
<sentence id="21" document_id="Perseus:text:1999.02.0002" subdoc="Book=2:chapter=5" span="In3:erat0">
|
165
|
+
<word id="1" form="In" lemma="in1" postag="r--------" head="4" relation="AuxP"/>
|
166
|
+
<word id="2" form="eo" lemma="is1" postag="p-s---nd-" head="3" relation="ATR"/>
|
167
|
+
<word id="3" form="flumine" lemma="flumen2" postag="n-s---nd-" head="1" relation="ADV"/>
|
168
|
+
<word id="4" form="pons" lemma="pons1" postag="n-s---mn-" head="5" relation="OBJ"/>
|
169
|
+
<word id="5" form="erat" lemma="sum1" postag="v3siia---" head="0" relation="PRED"/>
|
170
|
+
</sentence>
|
171
|
+
</treebank>
|
172
|
+
EOF
|
173
|
+
end
|
174
|
+
|
175
|
+
it "does not fall for additional (such as elliptic) nodes in the gold file" do
|
176
|
+
allow(differ).to receive(:get_from_uri).with(:uri_for_ge1) { ge1 }
|
177
|
+
allow(differ).to receive(:get_from_uri).with(:uri_for_re1) { re1 }
|
178
|
+
|
179
|
+
expect { differ.diff([:uri_for_ge1], [:uri_for_re1]) }.to_not raise_error
|
180
|
+
end
|
181
|
+
|
182
|
+
it "does not fall for additional (such as elliptic) nodes in the review file" do
|
183
|
+
allow(differ).to receive(:get_from_uri).with(:uri_for_re1) { re1 }
|
184
|
+
allow(differ).to receive(:get_from_uri).with(:uri_for_ge1) { ge1 }
|
185
|
+
|
186
|
+
expect { differ.diff([:uri_for_re1], [:uri_for_ge1]) }.to_not raise_error
|
187
|
+
end
|
188
|
+
end
|
189
|
+
|
190
|
+
context "with partially unannotated files" do
|
191
|
+
let(:gi1) do
|
192
|
+
<<-EOF
|
193
|
+
<treebank>
|
194
|
+
<sentence id="21" document_id="Perseus:text:1999.02.0002" subdoc="Book=2:chapter=5" span="In3:erat0">
|
195
|
+
<word id="1" form="In" lemma="in1" postag="r--------" head="5" relation="AuxP"/>
|
196
|
+
<word id="2" form="eo" lemma="is1" postag="p-s---nb-" head="3" relation="ATR"/>
|
197
|
+
<word id="3" form="flumine" lemma="flumen1" postag="n-s---nb-" head="1" relation="ADV"/>
|
198
|
+
<word id="4" form="pons" lemma="pons1" postag="n-s---mn-" head="5" relation="SBJ"/>
|
199
|
+
<word id="5" form="erat" lemma="sum1" postag="v3siia---" head="0" relation="PRED"/>
|
200
|
+
</sentence>
|
201
|
+
</treebank>
|
202
|
+
EOF
|
203
|
+
end
|
204
|
+
|
205
|
+
let(:ri1) do
|
206
|
+
<<-EOF
|
207
|
+
<treebank>
|
208
|
+
<sentence id="21" document_id="Perseus:text:1999.02.0002" subdoc="Book=2:chapter=5" span="In3:erat0">
|
209
|
+
<word id="1" form="In" postag="r--------" head="4" relation="AuxP"/>
|
210
|
+
<word id="2" form="eo" postag="p-s---nd-" head="3" relation="ATR"/>
|
211
|
+
<word id="3" form="flumine" lemma="flumen2" postag="n-s---nd-" relation="ADV"/>
|
212
|
+
<word id="4" form="pons" lemma="pons1" postag="n-s---mn-" head="5"/>
|
213
|
+
<word id="5" form="erat" lemma="sum1" postag="v3siia---" head="0" relation="PRED"/>
|
214
|
+
</sentence>
|
215
|
+
</treebank>
|
216
|
+
EOF
|
217
|
+
end
|
218
|
+
|
219
|
+
let(:ri2) do
|
220
|
+
<<-EOF
|
221
|
+
<treebank>
|
222
|
+
<sentence id="21" document_id="Perseus:text:1999.02.0002" subdoc="Book=2:chapter=5" span="In3:erat0">
|
223
|
+
<word id="1" form="In" postag="" head="4" relation="AuxP"/>
|
224
|
+
<word id="2" form="eo" head="3" relation="ATR"/>
|
225
|
+
<word id="3" form="flumine" lemma="flumen2" postag="n-s---nd-" relation="ADV"/>
|
226
|
+
<word id="4" form="pons" lemma="pons1" postag="n-s---mn-" head="5"/>
|
227
|
+
<word id="5" form="erat" lemma="sum1" postag="v3siia---" head="0" relation="PRED"/>
|
228
|
+
</sentence>
|
229
|
+
</treebank>
|
230
|
+
EOF
|
231
|
+
end
|
232
|
+
|
233
|
+
it "does not fall with less content in the review file" do
|
234
|
+
allow(differ).to receive(:get_from_uri).with(:uri_for_gi1) { gi1 }
|
235
|
+
allow(differ).to receive(:get_from_uri).with(:uri_for_ri1) { ri1 }
|
236
|
+
|
237
|
+
expect { differ.diff([:uri_for_gi1], [:uri_for_ri1]) }.to_not raise_error
|
238
|
+
end
|
239
|
+
|
240
|
+
it "does not fall with less content in the gold file" do
|
241
|
+
allow(differ).to receive(:get_from_uri).with(:uri_for_gi1) { gi1 }
|
242
|
+
allow(differ).to receive(:get_from_uri).with(:uri_for_ri1) { ri1 }
|
243
|
+
|
244
|
+
expect { differ.diff([:uri_for_ri1], [:uri_for_gi1]) }.to_not raise_error
|
245
|
+
end
|
246
|
+
|
247
|
+
it "does not fall with missing postags" do
|
248
|
+
allow(differ).to receive(:get_from_uri).with(:uri_for_gi1) { gi1 }
|
249
|
+
allow(differ).to receive(:get_from_uri).with(:uri_for_ri2) { ri2 }
|
250
|
+
|
251
|
+
expect { differ.diff([:uri_for_ri2], [:uri_for_gi1]) }.to_not raise_error
|
252
|
+
expect { differ.diff([:uri_for_gi1], [:uri_for_ri2]) }.to_not raise_error
|
253
|
+
end
|
254
|
+
end
|
144
255
|
end
|
145
256
|
|
146
257
|
describe "#report" do
|
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: llt-review
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.0.
|
4
|
+
version: 0.0.3
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- LFDM
|
8
8
|
autorequire:
|
9
9
|
bindir: bin
|
10
10
|
cert_chain: []
|
11
|
-
date: 2014-08-
|
11
|
+
date: 2014-08-14 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
14
|
name: bundler
|
@@ -223,7 +223,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
|
|
223
223
|
version: '0'
|
224
224
|
requirements: []
|
225
225
|
rubyforge_project:
|
226
|
-
rubygems_version: 2.2.
|
226
|
+
rubygems_version: 2.2.2
|
227
227
|
signing_key:
|
228
228
|
specification_version: 4
|
229
229
|
summary: Reviews annotations
|