Sha256: 0199c782e6e8cef28bda104dca2a9275e13aefa698074209482f9e52bac64631
Contents?: true
Size: 848 Bytes
Versions: 25
Compression:
Stored size: 848 Bytes
Contents
#!/usr/bin/env ruby
# encoding: utf-8
require 'biodiversity'
require 'json'
def parser_error(name)
{ 'scientificName' =>
{ 'parsed' => false,
'verbatim' => name,
'error' => 'Parser error' } }.to_json
end
if ARGV.empty?
puts "Usage:\n\n" +
"nnparse file_with_scientific_names [output_file]\n\n" +
"default output_file is parsed.json\n\n"
exit
end
input = ARGV[0]
output = ARGV[1] || 'parsed.json'
p = ScientificNameParser.new
o = open(output, 'w')
count = 0
puts 'Parsing...'
f = open(input, 'r:utf-8')
f.each do |line|
count += 1
puts("%s lines parsed" % count) if count % 10000 == 0
name = line.gsub(/^[\d]*\s*/, '').strip
begin
p.parse(name)
parsed_data = p.parsed.all_json rescue parser_error(name)
rescue
parsed_data = parser_error(name)
end
o.write parsed_data + "\n"
end
Version data entries
25 entries across 25 versions & 2 rubygems