rbbt-text 1.1.3 → 1.1.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 8e24d9496396b5679cdea2d9e4659fdc86bf9baf
-  data.tar.gz: 97b1cafa958ee4e60bd66e64fc3803d2fa2bf972
+  metadata.gz: 571b05a78ab3a1bb51d37604cab69773555fd331
+  data.tar.gz: 61983fbd5d98c8412bb2fd36d0639eb7086f9287
 SHA512:
-  metadata.gz: 21c91bf76cfb78e0c2a674a62c5c8e7971e981c631e74091524610f37351992d3b9ca5b197fcfccfe7d5315df5125cd1d6b5dfbc3b70566d9df7f9afe90d0a32
-  data.tar.gz: dbfd6442800434cf7073591f27007392e003d9511f2e00d0bd6c12c6ad5fd22bdde4b9e4b729334ddf6957f64aba360a0b407f926fa0e8c7bbe84e2bb9318bcf
+  metadata.gz: 25f718d2733cbd5bb2bb521bdc4d50e7386ac39a573c2424897bad98b1f8674beeec56c7162bec91c5afb555163eea995bbd740f9d855d13fc084d6248eff4cd
+  data.tar.gz: 184f47a53ffcb220ba68faf03343f869eb65b0f12022561355c9cf417245781ebfc46efa81a49c1a9d9d746b9e42434bbe4e9dc49cc6d70eb216ecc6f0517b48

data/lib/rbbt/ner/finder.rb

@@ -1,4 +1,5 @@
 require 'rbbt/ner/rnorm'
+require 'rbbt/ner/regexpNER'
 
 class Finder
   
@@ -20,6 +21,10 @@ class Finder
         @namespace = path.namespace 
         @format = path.key_field
         @normalizer = Normalizer.new(path)
+      when Hash
+        @namespace = open_options[:namespace]
+        @format = nil
+        @normalizer = RegExpNER.new(path)
       else
         open_options = Misc.add_defaults open_options, :type => :flat
         tsv = TSV.open(path, open_options)

data/lib/rbbt/ner/regexpNER.rb

@@ -56,8 +56,8 @@ class RegExpNER < NER
       chunks = Segment.split(text, matches)
       chunks.each do |chunk|
         chunk_offset = chunk.offset
-        match_regexp_list(chunk, regexp_list, type).collect do |match| 
-          match.offset += chunk_offset; 
+        match_regexp_list(chunk, regexp_list, type).each do |match| 
+          match.offset = match.offset + chunk_offset; 
           matches << match 
         end
       end
@@ -68,9 +68,12 @@ class RegExpNER < NER
 
   attr_accessor :regexps
   def initialize(regexps = {})
-    @regexps = regexps.collect
+    @regexps = regexps.collect{|p| p }
   end
 
+  def token_score(*args)
+    1
+  end
 
   def __define_regexp_hook(name, regexp, *args)
     @regexps << [name, regexp]
@@ -86,7 +89,9 @@ class RegExpNER < NER
 
   def match(text)
     matches = RegExpNER.match_regexp_hash(text, @regexps)
-    matches
+    matches.collect do |m|
+      NamedEntity.setup(m, :offset => m.offset, :type =>  m.type, :code => m)
+    end
   end
 
 end

data/lib/rbbt/ner/segment.rb

@@ -5,10 +5,6 @@ module Segment
   extend Annotation
   self.annotation :offset
 
-  def offset=(offset)
-    @offset = offset.nil? ? nil : offset.to_i
-  end
-
   def segment_length
     begin
       super()

data/test/rbbt/ner/test_finder.rb

@@ -10,7 +10,7 @@ class TestFinder < Test::Unit::TestCase
 
   def test_namespace_and_format
     f = Finder.new(CMD.cmd("head -n 1000", :in => Open.open(Organism.identifiers(Organism.default_code("Hsa")).produce.find)))
-    assert_equal "Hsa", f.instances.first.namespace
+    assert_equal Organism.default_code("Hsa"), f.instances.first.namespace
     assert_equal "Ensembl Gene ID", f.instances.first.format
   end
 
@@ -19,15 +19,24 @@ class TestFinder < Test::Unit::TestCase
 
     assert_equal "ENSG00000115524", f.find("SF3B1").first
     if defined? Entity
-      ddd f.find("SF3B1").first.info
       assert_equal "Ensembl Gene ID", f.find("SF3B1").first.format
     end
   end
 
-  def test_find
+  def test_find2
     f = Finder.new(Organism.lexicon(Organism.default_code("Hsa")), :grep => ["RASGRF2"])
 
+    m = f.find("RAS").first
+
     assert f.find("RAS").include? "ENSG00000113319"
   end
 
+  def test_find_mutation
+    f = Finder.new(Organism.lexicon(Organism.default_code("Hsa")), :grep => ["RASGRF2"])
+    f.add_instance({"Genomic Mutation" => /\w+:\w+:\w+/})
+
+    m =  f.find("The mutation 1:1234:A is intergenic").first
+    assert_equal "1:1234:A", m.info[:code]
+  end
+
 end

data/test/rbbt/ner/test_regexpNER.rb

@@ -79,6 +79,7 @@ class TestRegExpNER < Test::Unit::TestCase
     assert_equal :should, matches.select{|m| m.type == :should}[0].type
   end
 
+
   def test_regexp_order
     text =<<-EOF
   * Human AUC 0-24h= 7591 ng.h/ml at 30 mg/day    In mice, dietary administration of aripiprazole at doses of 1, 3, and 10 asdf mg/kg/day for 104 weeks was

data/test/test_helper.rb

@@ -20,8 +20,8 @@ class Test::Unit::TestCase
 
   def teardown
     FileUtils.rm_rf Rbbt.tmp.test.find :user
-    Persist::TC_CONNECTIONS.values.each do |c| c.close end
-    Persist::TC_CONNECTIONS.clear
+    Persist::CONNECTIONS.values.each do |c| c.close end
+    Persist::CONNECTIONS.clear
     DocumentRepo::TC_CONNECTIONS.values.each do |c| c.close end
     DocumentRepo::TC_CONNECTIONS.clear
   end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: rbbt-text
 version: !ruby/object:Gem::Version
-  version: 1.1.3
+  version: 1.1.4
 platform: ruby
 authors:
 - Miguel Vazquez
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-08-26 00:00:00.000000000 Z
+date: 2015-07-21 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rbbt-util
@@ -178,7 +178,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.2.2
+rubygems_version: 2.4.6
 signing_key: 
 specification_version: 4
 summary: Text mining tools for the Ruby Bioinformatics Toolkit (rbbt)