biodiversity19 0.7.3 → 1.0.0

data/README.rdoc

@@ -67,3 +67,5 @@ You can use it as a library
   # to resolve lsid and get back RDF file
   LsidResolver.resolve("urn:lsid:ubio.org:classificationbank:2232671")
 
+Copyright (c) 2009-2011 Marine Biological Laboratory. See LICENSE.txt for
+further details.

data/Rakefile

@@ -2,14 +2,15 @@ dir = File.dirname(__FILE__)
 require 'rubygems'
 require 'rake'
 #$LOAD_PATH.unshift(File.join(dir, 'vendor', 'rspec', 'lib'))
-require 'spec/rake/spectask'
+require 'rspec/core'
+require 'rspec/core/rake_task'
 
 #Gem::manage_gems
 #require 'rake/gempackagetask'
 
 task :default => :spec
 
-Spec::Rake::SpecTask.new do |t|
+RSpec::Core::RakeTask.new do |t|
   t.pattern = 'spec/**/*spec.rb'
 end
 
@@ -50,7 +51,7 @@ task :tt do
     f.each_with_index do |l, i|
       skip_head = l.match(/^# Autogenerated/) if i == 0
       if skip_head && (l.strip == '' || l.match(/^# Autogenerated/))
-        next  
+        next
       else
         skip_head = false
         rfn.write(l)

data/VERSION

@@ -1 +1 @@
-0.7.3
+1.0.0

data/bin/parserver

@@ -1,18 +1,95 @@
 #!/usr/bin/env ruby
 require 'rubygems'
+require 'optparse'
 require 'socket'
 require 'biodiversity'          # Get sockets from stdlib
-puts "Running parser service on port 4334"
+
+DEFAULT_PORT = 4334
+RUBY_VERSION_INT = RUBY_VERSION.split(".")[0..1].join('').to_i
+OPTIONS = {
+  :output => "json",
+  :port => DEFAULT_PORT
+}
+
+ARGV.options do |opts|
+  script_name = File.basename($0)
+  opts.banner = "Usage: ruby #{script_name} [options]"
+
+  opts.separator ""
+
+  opts.on("-o", "--output=output", String,
+          "Specifies the type of the output:
+    json - parsed results in json
+    canonical - canonical version
+    canonical_with_rank - canonical with rank",
+          "Default: json") { |OPTIONS[:output]| }
+
+  opts.separator ""
+
+  opts.on("-p", "--port=port", String,
+          "Specifies the port number",
+          "Default: #{DEFAULT_PORT}") { |OPTIONS[:port]| }
+
+  opts.separator ""
+
+  opts.on("-h", "--help",
+          "Show this help message.") { puts opts; exit }
+
+  opts.parse!
+end
+
+OPTIONS[:output] = "json" unless ['canonical', 'canonical_with_rank'].include?(OPTIONS[:output])
+OPTIONS[:port] = (OPTIONS[:port].to_i > 0 ? OPTIONS[:port].to_i : DEFAULT_PORT)
+
+def parser_error(name_string)
+  {:scientificName => {:parsed => false, :verbatim => name_string,  :error => 'Parser error'}}
+end
+
+def get_output(name_string, parser)
+  begin
+    if RUBY_VERSION_INT < 19
+      old_kcode = $KCODE
+      $KCODE = 'NONE'
+    end
+    parsed = parser.parse(name_string)
+    if RUBY_VERSION_INT < 19
+      $KCODE = old_kcode
+    end
+  rescue
+    parsed = parser_error(name_string)
+  end
+  output =  OPTIONS[:output]
+  return parsed.to_json if output == 'json'
+  canonical = parsed[:scientificName][:canonical]
+  return canonical.to_s if output == 'canonical' || canonical == nil || parsed[:scientificName][:hybrid] || !parsed[:scientificName][:parsed]
+  parts = parsed[:scientificName][:canonical].split(" ")
+
+  if parts.size > 2 && parsed[:scientificName][:details][0][:infraspecies]
+    name_ary = parts[0..1]
+    parsed[:scientificName][:details][0][:infraspecies].each do |data|
+      name_ary << (data[:rank] && data[:rank] != 'n/a'? "#{data[:rank]} #{data[:string]}" : data[:string])
+    end
+    canonical = name_ary.join(" ")
+  end
+  canonical
+end
+
+puts "Running parser service on port #{OPTIONS[:port]}, output type is '#{OPTIONS[:output]}'"
 parser = ScientificNameParser.new
-server = TCPServer.open(4334)  # Socket to listen on port 4334
+server = TCPServer.open(OPTIONS[:port])  # Socket to listen on a port
 loop do                         # Servers run forever
   client = server.accept       # Wait for a client to connect
-  while a = client.readline
-    if ['end','exit','q', '.'].include? a.strip
+  while true
+    begin
+      a = client.readline
+      if ['end','exit','q', '.'].include? a.strip
+        client.close
+        break
+      end
+      client.puts get_output(a, parser)
+    rescue EOFError
       client.close
       break
     end
-    client.puts parser.parse(a).to_json
   end
 end
-

data/examples/socket_client.rb

@@ -0,0 +1,25 @@
+#!/usr/bin/env ruby
+require 'socket'
+
+host = 'localhost'
+port = 4334
+
+f = open('10000_names.txt')
+w = open('output.txt', 'w')
+s = TCPSocket.open(host, port)
+
+f.each_with_index do |line, i|
+  puts i if i % 1000 == 0
+  line = line.strip
+  s.puts(line.strip)
+  res = s.gets
+  if res && res.split(" ").size > 3
+    res = res.strip
+    w.write(line + "\n")
+    w.write(res + "\n")
+    w.write("\n")
+  end
+end
+
+s.close
+

data/spec/guid/lsid.spec.rb

@@ -1,10 +1,9 @@
 dir = File.dirname("__FILE__")
 require 'rubygems'
-require 'spec'
 require File.expand_path(dir + "../../conf/environment")
 require File.expand_path(dir + "../../lib/biodiversity/guid")
 
-describe LsidResolver do   
+describe LsidResolver do
   it "should return RFD document from lsid" do
     lsid = "urn:lsid:ubio.org:classificationbank:2232671"
     LsidResolver.resolve(lsid).class.should == "".class

data/spec/parser/spec_helper.rb

@@ -1,6 +1,5 @@
 dir = File.dirname("__FILE__")
 require 'rubygems'
-require 'spec'
 require 'yaml'
 require 'treetop'
 require 'json'

metadata

@@ -1,12 +1,8 @@
 --- !ruby/object:Gem::Specification 
 name: biodiversity19
 version: !ruby/object:Gem::Version 
-  prerelease: false
-  segments: 
-  - 0
-  - 7
-  - 3
-  version: 0.7.3
+  prerelease: 
+  version: 1.0.0
 platform: ruby
 authors: 
 - Dmitry Mozzherin
@@ -14,8 +10,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-02-07 00:00:00 -05:00
-default_executable: 
+date: 2011-07-10 00:00:00 Z
 dependencies: 
 - !ruby/object:Gem::Dependency 
   name: treetop
@@ -25,8 +20,6 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        segments: 
-        - 0
         version: "0"
   type: :runtime
   version_requirements: *id001
@@ -38,8 +31,6 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        segments: 
-        - 0
         version: "0"
   type: :development
   version_requirements: *id002
@@ -55,7 +46,6 @@ extra_rdoc_files:
 - README.rdoc
 files: 
 - .document
-- .gitignore
 - LICENSE
 - README.rdoc
 - Rakefile
@@ -63,6 +53,7 @@ files:
 - bin/nnparse
 - bin/parserver
 - conf/environment.rb
+- examples/socket_client.rb
 - lib/biodiversity.rb
 - lib/biodiversity/guid.rb
 - lib/biodiversity/guid/lsid.rb
@@ -80,13 +71,12 @@ files:
 - spec/parser/spec_helper.rb
 - spec/parser/test_data.txt
 - spec/spec_helper.rb
-has_rdoc: true
 homepage: http://github.com/GlobalNamesArchitecture/biodiversity
 licenses: []
 
 post_install_message: 
-rdoc_options: 
-- --charset=UTF-8
+rdoc_options: []
+
 require_paths: 
 - lib
 required_ruby_version: !ruby/object:Gem::Requirement 
@@ -94,30 +84,19 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements: 
   - - ">="
     - !ruby/object:Gem::Version 
-      segments: 
-      - 0
       version: "0"
 required_rubygems_version: !ruby/object:Gem::Requirement 
   none: false
   requirements: 
   - - ">="
     - !ruby/object:Gem::Version 
-      segments: 
-      - 0
       version: "0"
 requirements: []
 
 rubyforge_project: 
-rubygems_version: 1.3.7
+rubygems_version: 1.8.5
 signing_key: 
 specification_version: 3
 summary: Parser of scientific names
-test_files: 
-- spec/biodiversity_spec.rb
-- spec/guid/lsid.spec.rb
-- spec/parser/scientific_name.spec.rb
-- spec/parser/scientific_name_canonical.spec.rb
-- spec/parser/scientific_name_clean.spec.rb
-- spec/parser/scientific_name_dirty.spec.rb
-- spec/parser/spec_helper.rb
-- spec/spec_helper.rb
+test_files: []
+

data/.gitignore

@@ -1,16 +0,0 @@
-lib/biodiversity/parser/*rb
-*.gemspec
-*.sw?
-.DS_Store
-coverage
-rdoc
-pkg
-*.swp
-*.swo
-biodiversity*.gem
-*json
-*xml
-tmp
-.DS_Store
-spec/parser/test_data_new.txt
-t