namor 0.4.0 → 0.4.1

data/lib/namor/namor.rb

@@ -1,12 +1,25 @@
 class Namor::Namor
   def initialize(opts = {})
     config(opts)
+    @re_cache = {}
   end
 
   def config(opts)
     @config = opts
   end
 
+
+  def suppression_re(supp_list)
+    suppression_list = (@config[:suppress] || []) + (supp_list || [])
+
+    re = '\b(' + suppression_list.compact.map{|s| s.chomp('.')}.map(&:upcase).join('|') + ')\b'
+    Regexp.new(re)
+    # bits = suppression_list.compact.map do |s|
+    #   '\b' + s.upcase.chomp('.') + '\b'
+    # end
+    # Regexp.new(bits.join('|'))
+  end
+
   # clean up a single name component
   # * output all converted to uppercase
   # * strip leading ZZ+ or XX+ (frequently used as invalid-account prefixes)
@@ -16,10 +29,9 @@ class Namor::Namor
   # * remove punctuation
   # * squeeze whitespace & trim spaces from ends
   def scrub(name, opts = {})
-    suppression_list = @config[:suppress] || []
-    suppression_re = Regexp.new('(\s|^)' + (suppression_list + (opts[:suppress]||[])).compact.map(&:upcase).join('|') + '(\s|\.|$)')
+    @re_cache[opts[:suppress]] ||= suppression_re(opts[:suppress])
 
-    name && name.upcase.gsub(/^[ZX]{2,}/, '').gsub(suppression_re, '').gsub(/\b(JR|SR|II|III|IV)\b/i, '').gsub(/\([^\(]*\)/, '').gsub(/\./, ' ').gsub(/[_'\&]/, '').gsub(/,\s*$/, '').gsub(/ +/, ' ').strip
+    name && name.upcase.gsub(/^[ZX]{2,}/, '').gsub(@re_cache[opts[:suppress]], '').gsub(/\b(JR|SR|II|III|IV)\b/i, '').gsub(/\([^\(]*\)/, '').gsub(/\./, ' ').gsub(/[_'\&]/, '').gsub(/,\s*$/, '').gsub(/ +/, ' ').strip
   end
 
   def fullscrub(name, opts = {})

data/lib/namor/version.rb

@@ -1,3 +1,3 @@
 module Namor
-  VERSION = "0.4.0"
+  VERSION = "0.4.1"
 end

data/spec/lib/namor_spec.rb

@@ -5,7 +5,7 @@ require "spec_helper"
 describe "name extract" do
   before(:all) do
     @namor = Namor::Namor.new
-    @namor.config(:suppress => ['MD', 'dds', 'M\.D\.'])
+    @namor.config(:suppress => ['MD', 'dds', 'M\.D'])
   end
 
   it "should handle 2-part names without commas" do
@@ -145,10 +145,13 @@ describe "title suppression" do
   it "should only suppress isolated terms" do
     @namor.scrub("Smith, Mary RN", :suppress => ['RN']).should == 'SMITH, MARY'
     @namor.scrub("Smith, Marnie", :suppress => ['RN']).should == 'SMITH, MARNIE'
+
+    @namor.scrub("Gonzalez, Lourdes", :suppress => ['RN', 'RD', 'DO']).should == 'GONZALEZ, LOURDES'
   end
 
   it "should scrub words with periods" do
-    @namor.scrub("Smith, Mary M.D.", :suppress => ['M.D.']).should == 'SMITH, MARY'
+    @namor.scrub("Smith, Mary M.D.", :suppress => ['M.D']).should == 'SMITH, MARY'
+    @namor.scrub("Smith, Mary M.D.", :suppress => ['RN', 'M.D.', 'DDS']).should == 'SMITH, MARY'
   end
 
   it "should scrub individual name components of punctuation and titles" do

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: namor
 version: !ruby/object:Gem::Version
-  version: 0.4.0
+  version: 0.4.1
   prerelease: 
 platform: ruby
 authors:
@@ -108,7 +108,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: -2390200260815377507
+      hash: -2504802179826978420
 required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
   requirements:
@@ -117,7 +117,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: -2390200260815377507
+      hash: -2504802179826978420
 requirements: []
 rubyforge_project: 
 rubygems_version: 1.8.24