dimus-biodiversity 0.0.3 → 0.0.5

data/README.rdoc

@@ -11,11 +11,11 @@ To install gem you need RubyGems >= 1.2.0
 
 == Example usage
 
-* You can parse file with species names from command line. File should contain one scientific name per line
+You can parse file with species names from command line. File should contain one scientific name per line
 
   nnparser file_with_names
   
-* You can use it as a library
+You can use it as a library
 
   require 'biodiversity'
   

data/bin/nnparse

@@ -4,25 +4,45 @@ gem 'dimus-biodiversity'
 
 $LOAD_PATH.unshift(File.expand_path(File.dirname(__FILE__) + "/../lib"))
 require 'biodiversity'
+require 'yaml'
 
 if ARGV.empty?
-  puts "Usage:\n\nnnparse file_with_scientific_names\n\n"
+  puts "Usage:\n\nnnparse file_with_scientific_names [output_file]\n\ndefault output_file is parsed.yml\n\n"
   exit
 end
 
 
 parser = ScientificNameParser.new
-debug = true
+
+output = ARGV[1] || 'parsed.yml'
+o = File.open(output,'w')
+
 # parse a file with names
 count = count2 = 0
-IO.foreach(ARGV[-1]) do |n|
-  puts 'started' if count2 == 0 && debug
+names = []
+IO.foreach(ARGV[0]) do |n|
+  name_dict = {}
+  puts 'Parsing names' if count2 == 0
   count2 += 1
-  puts count2.to_s + "/" + count.to_s if count2 % 10000 == 0 && debug
+  p count2  if count2 % 5000 == 0
   n.strip!
-  unless parser.parse n
-    puts n
+  name_dict = {:input => n} 
+  parsed = parser.parse n
+  unless parsed
+    name_dict[:details] = {:parsed => false}
+    names << name_dict
     count += 1
-  end    
+  else
+    name_dict[:output] = parsed.value
+    name_dict[:caononical] = parsed.canonical
+    name_dict[:details] = parsed.details
+    name_dict[:parsed => true]
+    names << name_dict
+  end
 end
-puts count.to_s
+$KCODE = 'UTF8'
+puts "Converting results to YAML"
+results = YAML.dump(names)
+puts "Writing restuls to #{output} file"
+o.write(results)
+puts "Found #{count2} records, #{count} of them could not be parsed."

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification 
 name: dimus-biodiversity
 version: !ruby/object:Gem::Version 
-  version: 0.0.3
+  version: 0.0.5
 platform: ruby
 authors: 
 - Dmitry Mozzherin
@@ -23,8 +23,8 @@ dependencies:
     version: 
 description: Biodiversity library provides a parser tool for scientific species names
 email: dmozzherin {et} eol {dt} org
-executables: []
-
+executables: 
+- - nnparse
 extensions: []
 
 extra_rdoc_files: []