rsinger

get '/gacs/:id' do
  headers['Cache-Control'] = 'public, max-age=43200'
  area = GeographicArea.first(:code=>params[:id])
  halt 404, "Not found" unless area
  if area.superceded_by
    replacement = GeographicArea.get(area.superceded_by)
    redirect "http://purl.org/NET/marccodes/gacs/#{replacement.code}"
  end
  @title = area.pref_label
  @code = area.to_rdf_object

rsinger

<?xml version="1.0"?>
<?xml-stylesheet type="text/xsl" href="/viaf/xsl/searchRetrieveResponse_en.xsl"?>
<searchRetrieveResponse xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns="http://www.loc.gov/zing/srw/" xsi:schemaLocation="http://www.loc.gov/zing/srw/ http://www.loc.gov/standards/sru/sru1-1archive/xml-files/srw-types.xsd">
  <version>1.1</version>
  <numberOfRecords>4</numberOfRecords>
  <resultSetId>52whe1</resultSetId>
  <resultSetIdleTime>300</resultSetIdleTime>
  <records xmlns:ns1="http://www.loc.gov/zing/srw/">
    <record>
      <recordSchema>http://viaf.org/VIAFCluster</recordSchema>

rsinger

{
  'body': ['type', 'value'], 
  'first_sentence': ['type', 'value'], 
  'table_of_contents': ['type', 'class', 'title', 'value', 'label', 'pagenum'], 
  'description': ['type', 'value'], 
  'language': ['key'], 
  'created': ['type', 'value'], 
  '/type/edition': [
    'pagination', 
    'table_of_contents', 

rsinger

use strict;
use JSON::XS;
use Data::Dumper;
my $file = '/Volumes/External 7/shared/open-library/edition-2009-09-11.txt';
open(INFILE, $file) || die "Can't open '$file': $!";
my $i = 0;
my %elements;
 
 
sub amerge($$) {

rsinger

require 'rubygems'
require 'json'
def inspect_file(file)
  elements = {}
  while line = file.gets
    parts = line.split("\t")
    elements[parts[1]] ||=[]
    j = JSON.parse(parts.last)
    j.keys.each do | key |
      elements[parts[1]] << key unless elements[parts[1]].index(key)

rsinger

/b/OL1M /type/edition   {"pagination": "8, 304 p.", "table_of_contents": [{"type": {"key": "/type/toc_item"}, "class": "section"}], "last_modified": {"type": "/type/datetime", "value": "2009-09-03T02:29:15.166712"}, "title": "Kabit\u0101.", "lccn": ["sa 64009056"], "notes": "Bibliographical footnotes.\r\nIn Oriya.", "number_of_pages": 304, "created": {"type": "/type/datetime", "value": "2008-04-01T03:28:50.625462"}, "languages": [{"key": "/l/ori"}], "lc_classifications": ["PK2579.R255 K3"], "latest_revision": 38, "key": "/b/OL1M", "authors": [{"key": "/a/OL1A"}], "publish_date": "1962", "oclc_numbers": ["31249133"], "type": {"key": 
"/type/edition"}, "id": 96, "revision": 38} 
/b/OL2M /type/edition   {"pagination": "5, 193 p.", "key": "/b/OL2M", "title": "Bh\u0101sa n\u0101\u1e6dakacakra.", "lccn": ["sa 64009057"], "notes": {"type": "/type/text", "value": "Includes bibliographical references.\r\nIn Kannada."}, "number_of_pages": 193, "id": 98, "languages": [{"key": "/l/kan"}], "language_code": "193", "lc_clas

rsinger

python 2.5.1:
real	41m51.102s
user	39m39.701s
sys	1m4.934s

ruby 1.8.7
real	72m4.442s
user	70m8.836s
sys	1m11.959s

rsinger

import simplejson as json

def inspectJSON(fh):
    fields = {"/type/edition":[]}

    for line in fh:
        resourceId, resourceType, jsonData = line.split("\t")
	    resource = json.loads(jsonData)
        for k in resource:
            if not k in fields[resourceType]:

rsinger

{"collections"=>{"key"=>1},
 "body"=>{"value"=>1, "type"=>1},
 "/type/edition"=>
  {"personal_name"=>1,
   "key"=>23607186,
   "languages"=>21103552,
   "isbn_13"=>6624813,
   "url"=>559861,
   "displayname"=>1,
   "purchase_url"=>244,

rsinger

{
  "type":"feed",
    
  "request":"/connector/resources/160682/items/?format=dlfexpanded",
    
  "time":"2010-04-05T13:34:00-04:00",
    
  "offset":0,
    
  "totalResults":6,