llt-review 0.0.1

data/spec/lib/llt/review/alignment/parser_spec.rb

@@ -0,0 +1,72 @@
+require 'spec_helper'
+
+describe LLT::Review::Alignment::Parser do
+  let(:parser) { LLT::Review::Alignment::Parser.new }
+
+  # Orgetorix rex fuit. => Orgetorix was a king. - Actually, he wasn't.
+  let(:data) { <<-EOF }
+    <aligned-text xmlns="http://alpheios.net/namespaces/aligned-text">
+      <language lnum="L1" xml:lang="lat"/>
+      <language lnum="L2" xml:lang="eng"/>
+      <sentence id="1">
+          <wds lnum="L1">
+              <w n="1-1">
+                  <text>Orgetorix</text>
+                  <refs nrefs="1-1"/>
+              </w>
+              <w n="1-2">
+                  <text>rex</text>
+                  <refs nrefs="1-3 1-4"/>
+              </w>
+              <w n="1-3">
+                  <text>fuit</text>
+                  <refs nrefs="1-2"/>
+              </w>
+              <w n="1-4">
+                  <text>.</text>
+                  <refs nrefs="1-5"/>
+              </w>
+          </wds>
+          <wds lnum="L2">
+              <w n="1-1">
+                  <text>Orgetorix</text>
+                  <refs nrefs="1-1"/>
+              </w>
+              <w n="1-2">
+                  <text>was</text>
+                  <refs nrefs="1-3"/>
+              </w>
+              <w n="1-3">
+                  <text>a</text>
+                  <refs nrefs="1-2"/>
+              </w>
+              <w n="1-4">
+                  <text>king</text>
+                  <refs nrefs="1-2"/>
+              </w>
+              <w n="1-5">
+                  <text>.</text>
+                  <refs nrefs="1-4"/>
+              </w>
+          </wds>
+      </sentence>
+    </aligned-text>
+  EOF
+
+  describe "#parse" do
+    it "returns parsed sentences" do
+      result = parser.parse(data)
+      s1= result[1]
+      s1.should be_true
+      s1.lang1.should == 'lat'
+      s1.lang2.should == 'eng'
+      s1.should have(4).items
+      rex = s1[2]
+      rex.should have(2).items
+      rex.translation.should == 'a king'
+      rex[3].to_s.should == 'a'
+      rex[4].to_s.should == 'king'
+    end
+  end
+end
+

data/spec/lib/llt/review/alignment_spec.rb

@@ -0,0 +1,242 @@
+require 'spec_helper'
+
+describe LLT::Review::Alignment do
+  let(:alignment) { LLT::Review::Alignment.new }
+
+  let(:g1) { <<-EOF }
+    <aligned-text xmlns="http://alpheios.net/namespaces/aligned-text">
+      <language lnum="L1" xml:lang="lat"/>
+      <language lnum="L2" xml:lang="eng"/>
+      <sentence id="1">
+          <wds lnum="L1">
+              <w n="1-1">
+                  <text>Orgetorix</text>
+                  <refs nrefs="1-1"/>
+              </w>
+              <w n="1-2">
+                  <text>rex</text>
+                  <refs nrefs="1-3 1-4"/>
+              </w>
+              <w n="1-3">
+                  <text>fuit</text>
+                  <refs nrefs="1-2"/>
+              </w>
+              <w n="1-4">
+                  <text>.</text>
+                  <refs nrefs="1-5"/>
+              </w>
+          </wds>
+          <wds lnum="L2">
+              <w n="1-1">
+                  <text>Orgetorix</text>
+                  <refs nrefs="1-1"/>
+              </w>
+              <w n="1-2">
+                  <text>was</text>
+                  <refs nrefs="1-3"/>
+              </w>
+              <w n="1-3">
+                  <text>a</text>
+                  <refs nrefs="1-2"/>
+              </w>
+              <w n="1-4">
+                  <text>king</text>
+                  <refs nrefs="1-2"/>
+              </w>
+              <w n="1-5">
+                  <text>.</text>
+                  <refs nrefs="1-4"/>
+              </w>
+          </wds>
+      </sentence>
+    </aligned-text>
+  EOF
+
+  let(:r1) { <<-EOF }
+    <aligned-text xmlns="http://alpheios.net/namespaces/aligned-text">
+      <language lnum="L1" xml:lang="lat"/>
+      <language lnum="L2" xml:lang="eng"/>
+      <sentence id="1">
+          <wds lnum="L1">
+              <w n="1-1">
+                  <text>Orgetorix</text>
+                  <refs nrefs="1-1"/>
+              </w>
+              <w n="1-2">
+                  <text>rex</text>
+                  <refs nrefs="1-3 1-4"/>
+              </w>
+              <w n="1-3">
+                  <text>fuit</text>
+                  <refs nrefs="1-2"/>
+              </w>
+              <w n="1-4">
+                  <text>.</text>
+                  <refs nrefs="1-5"/>
+              </w>
+          </wds>
+          <wds lnum="L2">
+              <w n="1-1">
+                  <text>Orgetorix</text>
+                  <refs nrefs="1-1"/>
+              </w>
+              <w n="1-2">
+                  <text>was</text>
+                  <refs nrefs="1-2"/>
+              </w>
+              <w n="1-3">
+                  <text>a</text>
+                  <refs nrefs="1-3"/>
+              </w>
+              <w n="1-4">
+                  <text>king</text>
+                  <refs nrefs="1-3"/>
+              </w>
+              <w n="1-5">
+                  <text>.</text>
+                  <refs nrefs="1-4"/>
+              </w>
+          </wds>
+      </sentence>
+    </aligned-text>
+  EOF
+
+  describe "#diff" do
+    describe "creates a diff report of a gold and review annotation" do
+      it "contains all differences in detail" do
+        allow(alignment).to receive(:get_from_uri).with(:uri_for_g1) { g1 }
+        allow(alignment).to receive(:get_from_uri).with(:uri_for_r1) { r1 }
+
+        result = alignment.diff([:uri_for_g1], [:uri_for_r1])
+        result.should have(1).item         # we had one reviewable annotation
+        result[0].should have(2).item # one sentence with differences & the report
+        result[0][1].should have(2).items # and 2 words with differences
+        wrong_words = result[0][1]
+        w2, w3 = wrong_words.take(2, 3)
+
+        w2.original.should == 'a king'
+        w2.new.should == 'was'
+
+        w3.original.should == 'was'
+        w3.new.should == 'a king'
+      end
+
+      it "contains a full report section" do
+        allow(alignment).to receive(:get_from_uri).with(:uri_for_g1) { g1 }
+        allow(alignment).to receive(:get_from_uri).with(:uri_for_r1) { r1 }
+
+        result = alignment.diff([:uri_for_g1], [:uri_for_r1])
+        report = result.first.report
+        report.should_not be_empty
+
+        sentences = report[:sentences]
+        sentences.total.should == 1
+        sentences.right.should == 0
+        sentences.wrong.should == 1
+        sentences.unique.should == 1
+
+        words = report[:words]
+        words.total.should == 4
+        words.right.should == 2
+        words.wrong.should == 2
+        words.unique.should == 2
+
+        rex = words['rex']
+        rex.total.should == 1
+        rex.right.should == 0
+        rex.wrong.should == 1
+        rex.unique.should == 1
+
+        fuit = words['fuit']
+        fuit.total.should == 1
+        fuit.right.should == 0
+        fuit.wrong.should == 1
+        fuit.unique.should == 1
+      end
+    end
+  end
+
+  let(:rep) { <<-EOF }
+    <aligned-text xmlns="http://alpheios.net/namespaces/aligned-text">
+        <language lnum="L1" xml:lang="lat"/>
+        <language lnum="L2" xml:lang="eng"/>
+        <sentence id="1">
+            <wds lnum="L1">
+                <w n="1-1">
+                    <text>Orgetorix</text>
+                    <refs nrefs="1-1"/>
+                </w>
+                <w n="1-2">
+                    <text>rex</text>
+                    <refs nrefs="1-3 1-4"/>
+                </w>
+                <w n="1-3">
+                    <text>fuit</text>
+                    <refs nrefs="1-2"/>
+                </w>
+                <w n="1-4">
+                    <text>,</text>
+                    <refs nrefs="1-5"/>
+                </w>
+                <w n="1-5">
+                    <text>rex</text>
+                    <refs nrefs="1-6 1-7"/>
+                </w>
+                <w n="1-6">
+                    <text>.</text>
+                    <refs nrefs="1-8"/>
+                </w>
+            </wds>
+            <wds lnum="L2">
+                <w n="1-1">
+                    <text>Orgetorix</text>
+                    <refs nrefs="1-1"/>
+                </w>
+                <w n="1-2">
+                    <text>was</text>
+                    <refs nrefs="1-3"/>
+                </w>
+                <w n="1-3">
+                    <text>a</text>
+                    <refs nrefs="1-2"/>
+                </w>
+                <w n="1-4">
+                    <text>king</text>
+                    <refs nrefs="1-2"/>
+                </w>
+                <w n="1-5">
+                    <text>,</text>
+                    <refs nrefs="1-4"/>
+                </w>
+                <w n="1-6">
+                    <text>a</text>
+                    <refs nrefs="1-5"/>
+                </w>
+                <w n="1-7">
+                    <text>king</text>
+                    <refs nrefs="1-5"/>
+                </w>
+                <w n="1-8">
+                    <text>.</text>
+                    <refs nrefs="1-6"/>
+                </w>
+            </wds>
+        </sentence>
+    </aligned-text>
+  EOF
+
+  describe "#report" do
+    it "reports about how every word was translated" do
+      allow(alignment).to receive(:get_from_uri).with(:uri_for_rep) { rep }
+      result = alignment.report(:uri_for_rep)
+      result.should have(1).item
+      report = result.first
+      report[:sentences].total.should == 1
+      words = report[:words]
+      words.total.should == 6
+      rex = words['rex']
+      rex.total.should == 2
+      rex['a king'].total.should == 2
+    end
+  end
+end

data/spec/lib/llt/review/helpers/reportable_spec.rb

@@ -0,0 +1,26 @@
+require 'spec_helper'
+
+describe LLT::Review::Helpers::Reportable do
+  let(:dummy) { Class.new { include LLT::Review::Helpers::Reportable } }
+
+  describe "#clone" do
+    it "performs a deep copy including all container contents" do
+      d = dummy.new('id')
+      5.times { |i| d.add(dummy.new("id#{i}", 1)) }
+      d['id1'].add(dummy.new('id5'))
+
+      cloned = d.clone
+      cloned.increment
+      cloned_id5 = cloned['id1']['id5']
+      d_id5 = d['id1']['id5']
+      cloned_id5.increment
+      cloned_id5.add(dummy.new('id6'))
+
+      cloned.should_not == d
+      cloned.total.should == 2
+      d.total.should == 1
+      d_id5.total.should_not == cloned_id5.total
+      d_id5['id6'].should_not == cloned_id5['id6']
+    end
+  end
+end

data/spec/lib/llt/review/treebank/parser_spec.rb

@@ -0,0 +1,21 @@
+require 'spec_helper'
+
+describe LLT::Review::Treebank::Parser do
+  let(:parser) { LLT::Review::Treebank::Parser.new() }
+  let(:data) { <<-EOF }
+    <sentence id="21" document_id="Perseus:text:1999.02.0002" subdoc="Book=2:chapter=5" span="In3:erat0">
+      <word id="1" form="In" lemma="in1" postag="r--------" head="5" relation="AuxP"/>
+      <word id="2" form="eo" lemma="is1" postag="p-s---nb-" head="3" relation="ATR"/>
+      <word id="3" form="flumine" lemma="flumen1" postag="n-s---nb-" head="1" relation="ADV"/>
+      <word id="4" form="pons" lemma="pons1" postag="n-s---mn-" head="5" relation="SBJ"/>
+      <word id="5" form="erat" lemma="sum1" postag="v3siia---" head="0" relation="PRED"/>
+    </sentence>
+  EOF
+
+  describe "#parse" do
+    it "returns parsed sentences" do
+      result = parser.parse(data)
+      result[21].should be_true
+    end
+  end
+end

data/spec/lib/llt/review/treebank/postag_spec.rb

@@ -0,0 +1,22 @@
+require 'spec_helper'
+
+describe LLT::Review::Treebank::Postag do
+  let(:postag) { LLT::Review::Treebank::Postag.new('v3siia---') }
+
+  describe "#analysis" do
+    it "decodes the postag into a hash" do
+      res = {
+        part_of_speech: 'v',
+        person: '3',
+        number: 's',
+        tense: 'i',
+        mood: 'i',
+        voice: 'a',
+        gender: '-',
+        case: '-',
+        degree: '-',
+      }
+      postag.analysis.should == res
+    end
+  end
+end

data/spec/lib/llt/review/treebank_spec.rb

@@ -0,0 +1,222 @@
+require 'spec_helper'
+
+describe LLT::Review::Treebank do
+  let(:differ) { LLT::Review::Treebank.new }
+
+  let(:g1) do
+    <<-EOF
+      <treebank>
+        <sentence id="21" document_id="Perseus:text:1999.02.0002" subdoc="Book=2:chapter=5" span="In3:erat0">
+          <word id="1" form="In" lemma="in1" postag="r--------" head="5" relation="AuxP"/>
+          <word id="2" form="eo" lemma="is1" postag="p-s---nb-" head="3" relation="ATR"/>
+          <word id="3" form="flumine" lemma="flumen1" postag="n-s---nb-" head="1" relation="ADV"/>
+          <word id="4" form="pons" lemma="pons1" postag="n-s---mn-" head="5" relation="SBJ"/>
+          <word id="5" form="erat" lemma="sum1" postag="v3siia---" head="0" relation="PRED"/>
+        </sentence>
+      </treebank>
+    EOF
+  end
+
+  let(:r1) do
+    <<-EOF
+      <treebank>
+        <sentence id="21" document_id="Perseus:text:1999.02.0002" subdoc="Book=2:chapter=5" span="In3:erat0">
+          <word id="1" form="In" lemma="in1" postag="r--------" head="4" relation="AuxP"/>
+          <word id="2" form="eo" lemma="is1" postag="p-s---nd-" head="3" relation="ATR"/>
+          <word id="3" form="flumine" lemma="flumen2" postag="n-s---nd-" head="1" relation="ADV"/>
+          <word id="4" form="pons" lemma="pons1" postag="n-s---mn-" head="5" relation="OBJ"/>
+          <word id="5" form="erat" lemma="sum1" postag="v3siia---" head="0" relation="PRED"/>
+        </sentence>
+      </treebank>
+    EOF
+  end
+
+  let(:g2) do
+    <<-EOF
+      <treebank>
+        <sentence id="21" document_id="Perseus:text:1999.02.0002" subdoc="Book=2:chapter=5" span="In3:erat0">
+          <word id="1" form="In" lemma="in1" postag="r--------" head="5" relation="AuxP"/>
+          <word id="2" form="eo" lemma="is1" postag="p-s---nb-" head="3" relation="ATR"/>
+          <word id="3" form="flumine" lemma="flumen1" postag="n-s---nb-" head="1" relation="ADV"/>
+          <word id="4" form="pons" lemma="pons1" postag="n-s---mn-" head="5" relation="SBJ"/>
+          <word id="5" form="erat" lemma="sum1" postag="v3siia---" head="0" relation="PRED"/>
+        </sentence>
+        <sentence id="22" document_id="Perseus:text:1999.02.0002" subdoc="Book=2:chapter=5" span="In3:erat0">
+          <word id="1" form="In" lemma="in1" postag="r--------" head="5" relation="AuxP"/>
+          <word id="2" form="eo" lemma="is1" postag="p-s---nb-" head="3" relation="ATR"/>
+          <word id="3" form="flumine" lemma="flumen1" postag="n-s---nb-" head="1" relation="ADV"/>
+          <word id="4" form="pons" lemma="pons1" postag="n-s---mn-" head="5" relation="SBJ"/>
+          <word id="5" form="erat" lemma="sum1" postag="v3siia---" head="0" relation="PRED"/>
+        </sentence>
+      </treebank>
+    EOF
+  end
+
+  describe "#diff" do
+    describe "creates a diff report of a gold and review annotation" do
+      it "contains all differences in detail" do
+        allow(differ).to receive(:get_from_uri).with(:uri_for_g1) { g1 }
+        allow(differ).to receive(:get_from_uri).with(:uri_for_r1) { r1 }
+
+        result = differ.diff([:uri_for_g1], [:uri_for_r1])
+        result.should have(1).item         # we had one reviewable annotation
+        result[0].should have(2).items     # one sentence with differences, report also sitting here
+        result[0][21].should have(4).items # and 4 words with differences
+        diff = result[0][21]
+        w1, w2,  w3, w4 = diff.take(1, 2, 3, 4).map(&:diff)
+
+        w1[:head].original.should == '5'
+        w1[:head].new.should == '4'
+
+        w2[:postag].original.should == 'p-s---nb-'
+        w2[:postag].new.should == 'p-s---nd-'
+        w2[:postag].unique.should == 1
+        w2[:postag][:case].original.should == 'b'
+        w2[:postag][:case].new.should == 'd'
+        w2[:postag][:case].unique.should == 1 # first occurence of this difference
+
+        w3[:lemma].original.should == 'flumen1'
+        w3[:lemma].new.should == 'flumen2'
+        w3[:postag].original.should == 'n-s---nb-'
+        w3[:postag].new.should == 'n-s---nd-'
+        w3[:postag][:case].original.should == 'b'
+        w3[:postag][:case].new.should == 'd'
+        w3[:postag][:case].unique.should == 0 # second occurence of this difference
+
+        w4[:relation].original.should == 'SBJ'
+        w4[:relation].new.should == 'OBJ'
+        w4[:relation].unique.should == 1
+      end
+
+      it "contains a full report section" do
+        allow(differ).to receive(:get_from_uri).with(:uri_for_g1) { g1 }
+        allow(differ).to receive(:get_from_uri).with(:uri_for_r1) { r1 }
+
+        result = differ.diff([:uri_for_g1], [:uri_for_r1])
+        report = result.first.report
+        report.should_not be_empty
+
+        sentences = report[:sentences]
+        sentences.total.should == 1
+        sentences.right.should == 0
+        sentences.wrong.should == 1
+        sentences.unique.should == 1
+
+        # TODO
+        # Add a couple of more assertions just to be safe
+
+        postags = report[:postags]
+        datapoints = postags[:datapoints]
+        cases = datapoints[:cases]
+        ablative = cases['b']
+        ablative.total.should == 2
+        ablative.right.should == 0
+        ablative.wrong.should == 2
+        ablative.unique.should == 1
+      end
+    end
+
+    it "takes multiple gold and review files" do
+      allow(differ).to receive(:get_from_uri).with(:uri_for_g1) { g1 }
+      allow(differ).to receive(:get_from_uri).with(:uri_for_g2) { g1 }
+      allow(differ).to receive(:get_from_uri).with(:uri_for_r1) { r1 }
+      allow(differ).to receive(:get_from_uri).with(:uri_for_r2) { r1 }
+
+      result = differ.diff(%i{ uri_for_g1 uri_for_g2 }, %i{ uri_for_r1 uri_for_r2 })
+      result.should have(4).items # we have two times two reviewable annotations now
+    end
+
+    it "takes an optional array of elements to compare with each other" do
+      allow(differ).to receive(:get_from_uri).with(:uri_for_g1) { g1 }
+      allow(differ).to receive(:get_from_uri).with(:uri_for_r1) { r1 }
+
+      result = differ.diff([:uri_for_g1], [:uri_for_r1], [:head, :relation])
+      result.should have(1).item         # we had one reviewable annotation
+      result[0].should have(2).items     # one sentence with differences, report also sitting here
+      result[0][21].should have(2).items # and 2 words with differences
+      diff = result[0][21]
+
+      diff[1].should be_true
+      diff[2].should be_nil # has only postag differences
+      diff[3].should be_nil # has only lemma and postag differences
+      diff[4].should be_true
+    end
+  end
+
+  describe "#report" do
+    it "analyses occurences of lemmata, head, relation, postags... of passed uris" do
+      allow(differ).to receive(:get_from_uri).with(:uri_for_g1) { g2 }
+      allow(differ).to receive(:get_from_uri).with(:uri_for_g2) { g2 }
+      result = differ.report(:uri_for_g1)
+      result.should have(1).item
+      report = result.first
+      report[:sentences].total.should == 2
+      report[:words].total.should == 10
+      report[:heads].total.should == 10
+
+      relations = report[:relations]
+      relations.total.should == 10
+      relations['ADV'].total.should == 2
+      relations['ATR'].total.should == 2
+      relations['AuxP'].total.should == 2
+      relations['PRED'].total.should == 2
+      relations['SBJ'].total.should == 2
+
+      lemmata = report[:lemmata]
+      lemmata.total.should == 10
+      lemmata['flumen1'].total.should == 2
+      lemmata['in1'].total.should == 2
+      lemmata['is1'].total.should == 2
+      lemmata['pons1'].total.should == 2
+      lemmata['sum1'].total.should == 2
+
+      postags = report[:postags]
+      postags.total.should == 10
+      postags["r--------"].total.should == 2
+      postags["p-s---nb-"].total.should == 2
+      postags["n-s---nb-"].total.should == 2
+      postags["n-s---mn-"].total.should == 2
+      postags["v3siia---"].total.should == 2
+
+      datapoints = postags[:datapoints]
+      datapoints.total.should == 90
+
+      pos = datapoints[:parts_of_speech]
+      pos.total.should == 10
+      pos['r'].total.should == 2
+      pos['p'].total.should == 2
+      pos['n'].total.should == 4
+      pos['v'].total.should == 2
+
+      persons = datapoints[:persons]
+      persons.total.should == 2
+      persons['3'].total.should == 2
+
+      numbers = datapoints[:numbers]
+      numbers.total.should == 8
+      numbers['s'].total.should == 8
+
+      tenses = datapoints[:tenses]
+      tenses.total.should == 2
+      tenses['i'].total.should == 2
+
+      moods = datapoints[:moods]
+      moods.total.should == 2
+      moods['i'].total.should == 2
+
+      voices= datapoints[:voices]
+      voices.total.should == 2
+      voices['a'].total.should == 2
+
+      genders = datapoints[:genders]
+      genders.total.should == 6
+      genders['n'].total.should == 4
+      genders['m'].total.should == 2
+
+      cases = datapoints[:cases]
+      cases.total.should == 6
+      cases['b'].total.should == 4
+      cases['n'].total.should == 2
+    end
+  end
+end

data/spec/lib/llt/review_spec.rb

@@ -0,0 +1,7 @@
+require 'spec_helper'
+
+describe LLT::Review do
+  it 'should have a version number' do
+    LLT::Review::VERSION.should_not be_nil
+  end
+end

data/spec/spec_helper.rb

@@ -0,0 +1,22 @@
+require 'simplecov'
+require 'coveralls'
+
+Coveralls.wear!
+
+SimpleCov.formatter = SimpleCov::Formatter::MultiFormatter[
+  SimpleCov::Formatter::HTMLFormatter,
+  Coveralls::SimpleCov::Formatter
+]
+
+SimpleCov.start do
+  add_filter '/spec/'
+end
+
+$LOAD_PATH.unshift File.expand_path('../../lib', __FILE__)
+require 'llt/review'
+
+RSpec.configure do |config|
+  config.treat_symbols_as_metadata_keys_with_true_values = true
+  config.run_all_when_everything_filtered = true
+  config.filter_run :focus
+end