shalmaneser-fred 1.2.0.rc4 → 1.2.rc5

metadata

@@ -1,15 +1,29 @@
 --- !ruby/object:Gem::Specification
 name: shalmaneser-fred
 version: !ruby/object:Gem::Version
-  version: 1.2.0.rc4
+  version: 1.2.rc5
 platform: ruby
 authors:
 - Andrei Beliankou
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-09-04 00:00:00.000000000 Z
+date: 2016-01-13 00:00:00.000000000 Z
 dependencies:
+- !ruby/object:Gem::Dependency
+  name: shalmaneser-lib
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - '='
+      - !ruby/object:Gem::Version
+        version: 1.2.rc5
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - '='
+      - !ruby/object:Gem::Version
+        version: 1.2.rc5
 - !ruby/object:Gem::Dependency
   name: mysql
   requirement: !ruby/object:Gem::Requirement
@@ -39,31 +53,49 @@ files:
 - LICENSE.md
 - README.md
 - bin/fred
-- lib/fred/Baseline.rb
-- lib/fred/FileZipped.rb
-- lib/fred/FredBOWContext.rb
 - lib/fred/FredConventions.rb
-- lib/fred/FredDetermineTargets.rb
-- lib/fred/FredEval.rb
-- lib/fred/FredFeatureExtractors.rb
-- lib/fred/FredFeatures.rb
-- lib/fred/FredFeaturize.rb
-- lib/fred/FredNumTrainingSenses.rb
-- lib/fred/FredParameters.rb
-- lib/fred/FredSplit.rb
-- lib/fred/FredSplitPkg.rb
-- lib/fred/FredTest.rb
-- lib/fred/FredTrain.rb
-- lib/fred/PlotAndREval.rb
+- lib/fred/abstract_context_provider.rb
+- lib/fred/abstract_fred_feature_access.rb
+- lib/fred/answer_key_access.rb
+- lib/fred/aux_keep_writers.rb
+- lib/fred/baseline.rb
+- lib/fred/context_provider.rb
+- lib/fred/feature_extractors.rb
+- lib/fred/feature_extractors/fred_context_feature_extractor.rb
+- lib/fred/feature_extractors/fred_context_pos_feature_extractor.rb
+- lib/fred/feature_extractors/fred_feature_extractor.rb
+- lib/fred/feature_extractors/fred_ngram_feature_extractor.rb
+- lib/fred/feature_extractors/fred_syn_feature_extractor.rb
+- lib/fred/feature_extractors/fred_synsem_feature_extractor.rb
+- lib/fred/file_zipped.rb
+- lib/fred/find_all_targets.rb
+- lib/fred/find_targets_from_frames.rb
 - lib/fred/fred.rb
-- lib/fred/fred_config_data.rb
+- lib/fred/fred_error.rb
+- lib/fred/fred_eval.rb
+- lib/fred/fred_feature_access.rb
+- lib/fred/fred_feature_info.rb
+- lib/fred/fred_featurize.rb
+- lib/fred/fred_parameters.rb
+- lib/fred/fred_split.rb
+- lib/fred/fred_split_pkg.rb
+- lib/fred/fred_test.rb
+- lib/fred/fred_train.rb
+- lib/fred/grammatical_function_access.rb
 - lib/fred/md5.rb
+- lib/fred/meta_feature_access.rb
+- lib/fred/non_contiguous_context_provider.rb
 - lib/fred/opt_parser.rb
-- test/frprep/test_opt_parser.rb
-- test/functional/functional_test_helper.rb
-- test/functional/test_fred.rb
-- test/functional/test_frprep.rb
-- test/functional/test_rosy.rb
+- lib/fred/plot_and_r_eval.rb
+- lib/fred/single_sent_context_provider.rb
+- lib/fred/slide_var.rb
+- lib/fred/targets.rb
+- lib/fred/toggle_var.rb
+- lib/fred/word_lemma_pos_ne.rb
+- lib/fred/write_features_binary.rb
+- lib/fred/write_features_nary.rb
+- lib/fred/write_features_nary_or_binary.rb
+- lib/shalmaneser/fred.rb
 homepage: https://github.com/arbox/shalmaneser
 licenses:
 - GPL-2.0
@@ -86,14 +118,9 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: 1.3.1
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.4.5
+rubygems_version: 2.5.1
 signing_key: 
 specification_version: 4
 summary: FRED
-test_files:
-- test/frprep/test_opt_parser.rb
-- test/functional/functional_test_helper.rb
-- test/functional/test_fred.rb
-- test/functional/test_frprep.rb
-- test/functional/test_rosy.rb
+test_files: []
 has_rdoc: 

data/lib/fred/Baseline.rb

@@ -1,150 +0,0 @@
-# Baseline
-# Katrin Erk April 05
-#
-# baseline for WSD:
-# always assign most frequent sense
-# The baseline doesn't do binary classifiers.
-
-require "fred/FredConventions"
-require "fred/FredSplitPkg"
-require "fred/FredFeatures"
-require "fred/FredDetermineTargets"
-
-class Baseline
-  ###
-  # new
-  #
-  # get splitlog dir (if any) along with everything else
-  # because we are only evaluating the training data
-  # at test time
-  #
-  def initialize(exp, # FredConfigData object 
-		 split_id = nil) # string: split ID
-    @exp = exp
-    @split_id = split_id
-
-    # for each lemma: remember prevalent sense
-    @lemma_to_sense = Hash.new()
-
-    if @split_id
-      split_obj = FredSplitPkg.new(@exp)
-    end
-
-    lemma_done = Hash.new()
-
-    # iterate through lemmas
-    @target_obj = Targets.new(@exp, nil, "r")
-    unless @target_obj.targets_okay
-      # error during initialization
-      $stderr.puts "Error: Could not read list of known targets, bailing out."
-      exit 1
-    end
-
-    @target_obj.get_lemmas().each { |lemmapos|
-
-      if @split_id
-        # read training split of answer keys
-        answer_obj = AnswerKeyAccess.new(@exp, "train", lemmapos, "r", @split_id, "train")
-      else
-        # read full answer key file of training data
-        answer_obj = AnswerKeyAccess.new(@exp, "train", lemmapos, "r")
-      end      
-
-      count_senses = Hash.new(0)
-
-      answer_obj.each { |lemma, pos, ids, sid, senses_all, senses_this|
-        # senses_this may include more than one sense for multi-label assignment
-        senses_this.each { |sense|
-          count_senses[sense] += 1
-        }
-      }
-
-      @lemma_to_sense[lemmapos] = count_senses.keys().max { |a, b| 
-        count_senses[a] <=> count_senses[b] 
-      }
-    }
-
-
-    @lemma = nil
-  end
-
-  ###
-  def train(infilename)
-    # no training here
-  end
-
-  ###
-  def write(classifier_file)
-    # no classifiers to write
-  end
-
-  def exists?(classifier_file)
-    return true
-  end
-
-  def read(classifier_file)
-    values = deconstruct_fred_classifier_filename(File.basename(classifier_file))
-    @lemma = values["lemma"]
-    if @lemma
-      return true
-    else
-      $stderr.puts "Warning: couldn't determine lemma name in #{classifier_file}, skipping"
-      return false
-    end
-  end
-
-
-  def read_resultfile(filename)
-    retv = Array.new()
-    begin
-      f = File.new(filename)
-    rescue
-      raise "Could not read baseline result file #{filename}"
-    end
-
-    f.each { |line|
-      retv << [[ line.chomp(), 1.0 ]]
-    }
-
-    return retv
-  end
-
-  def apply(infilename, outfilename)
-    # open input and output file
-    begin
-      out_f = File.new(outfilename, "w")
-    rescue
-      $stderr.puts "Error: cannot write to classification output file #{outfilename}."
-      exit 1
-    end
-    begin
-      f = File.new(infilename)
-    rescue
-      $stderr.puts "Error: cannot read feature file #{infilename}."
-      exit 1
-    end
-
-    # deconstruct input filename to determine lemma
-    unless @lemma
-      # something went wrong in read()
-      return false
-    end
- 
-    # do we have a sense for this?
-    unless (sense = @lemma_to_sense[@lemma])
-      # nope: assign "NONE" (or whatever the null label is here)
-      sense = @exp.get("negsense")
-      unless sense
-        sense = "NONE"
-      end
-    end
-
-    f.each { |line|
-      out_f.puts sense
-    }
-    out_f.close()
-    f.close()
-
-    return true
-  end
-end

data/lib/fred/FileZipped.rb

@@ -1,31 +0,0 @@
-class FileZipped
-
-  def FileZipped.new(filename,
-                     mode = "r")
-
-    # escape characters in the filename that
-    # would make the shell hiccup on the command
-    filename = filename.gsub(/([();:!?'`])/, 'XXSLASHXX\1')
-    filename = filename.gsub(/XXSLASHXX/, "\\")
-
-    begin
-      case mode
-      when "r"
-        unless File.exists? filename
-          raise "catchme"
-        end
-        return IO.popen("gunzip -c #{filename}")
-      when "w"
-        return IO.popen("gzip > #{filename}", "w")
-      when "a"
-        return IO.popen("gzip >> #{filename}", "w")
-      else
-        $stderr.puts "FileZipped error: only modes r, w, a are implemented. I got: #{mode}."
-        exit 1
-      end
-    rescue
-      raise "Error opening file #{filename}."
-    end
-  end
-
-end