Merge pull request #128 from OpenGeoMetadata/121-migrate-to-aardvark

Add initial class for aardvark migration

lib/geo_combine.rb

@@ -72,6 +72,9 @@ def to_html
 # Require harvesting/indexing files
 require 'geo_combine/geo_blacklight_harvester'
 
+# Migrators
+require 'geo_combine/migrators/v1_aardvark_migrator'
+
 # Require gem files
 require 'geo_combine/version'
 require 'geo_combine/railtie' if defined?(Rails)

lib/geo_combine/migrators/v1_aardvark_migrator.rb

@@ -0,0 +1,52 @@
+# frozen_string_literal: true
+
+module GeoCombine
+  module Migrators
+    # TODO: WARNING! This class is not fully implemented and should not be used in
+    # production. See https://github.com/OpenGeoMetadata/GeoCombine/issues/121
+    # for remaining work.
+    #
+    # migrates the v1 schema to the aardvark schema
+    class V1AardvarkMigrator
+      attr_reader :v1_hash
+
+      # @param v1_hash [Hash] parsed json in the v1 schema
+      def initialize(v1_hash:)
+        @v1_hash = v1_hash
+      end
+
+      def run
+        v2_hash = convert_keys
+        v2_hash['gbl_mdVersion_s'] = 'Aardvark'
+        v2_hash
+      end
+
+      def convert_keys
+        v1_hash.transform_keys do |k|
+          SCHEMA_FIELD_MAP[k] || k
+        end
+      end
+
+      SCHEMA_FIELD_MAP = {
+        'dc_title_s' => 'dct_title_s', # new namespace
+        'dc_description_s' => 'dct_description_sm', # new namespace; single to multi-valued
+        'dc_language_s' => 'dct_language_sm', # new namespace; single to multi-valued
+        'dc_language_sm' => 'dct_language_sm', # new namespace; single to multi-valued
+        'dc_creator_sm' => 'dct_creator_sm', # new namespace
+        'dc_publisher_s' => 'dct_publisher_sm', # new namespace; single to multi-valued
+        'dct_provenance_s' => 'schema_provider_s', # new URI name
+        'dc_subject_sm' => 'dct_subject_sm', # new namespace
+        'solr_year_i' => 'gbl_indexYear_im', # new URI name; single to multi-valued
+        'dc_source_sm' => 'dct_source_sm', # new namespace
+        'dc_rights_s' => 'dct_accessRights_s', # new URI name
+        'dc_format_s' => 'dct_format_s', # new namespace
+        'layer_id_s' => 'gbl_wxsIdentifier_s', # new URI name
+        'layer_slug_s' => 'id', # new URI name
+        'dc_identifier_s' => 'dct_identifier_sm', # new namespace; single to multi-valued
+        'layer_modified_dt' => 'gbl_mdModified_dt', # new URI name
+        'geoblacklight_version' => 'gbl_mdVersion_s', # new URI name
+        'suppressed_b' => 'gbl_suppressed_b' # new namespace
+      }.freeze
+    end
+  end
+end

spec/fixtures/docs/full_geoblacklight_aardvark.json

@@ -0,0 +1,33 @@
+{
+  "gbl_mdVersion_s":"Aardvark",
+  "dct_identifier_sm":"http://purl.stanford.edu/cz128vq0535",
+  "dct_title_s":"2005 Rural Poverty GIS Database: Uganda",
+  "dct_description_sm":"This polygon shapefile contains 2005 poverty data for 855 rural subcounties in Uganda. These data are intended for researchers, students, policy makers and the general public for reference and mapping purposes, and may be used for basic applications such as viewing, querying, and map output production.",
+  "dct_accessRights_s":"Public",
+  "schema_provider_s":"Stanford",
+  "dct_references_s":"{\"http://schema.org/url\":\"http://purl.stanford.edu/cz128vq0535\",\"http://schema.org/downloadUrl\":\"http://stacks.stanford.edu/file/druid:cz128vq0535/data.zip\",\"http://www.loc.gov/mods/v3\":\"http://purl.stanford.edu/cz128vq0535.mods\",\"http://www.isotc211.org/schemas/2005/gmd/\":\"http://opengeometadata.stanford.edu/metadata/edu.stanford.purl/druid:cz128vq0535/iso19139.xml\",\"http://www.w3.org/1999/xhtml\":\"http://opengeometadata.stanford.edu/metadata/edu.stanford.purl/druid:cz128vq0535/default.html\",\"http://www.opengis.net/def/serviceType/ogc/wfs\":\"https://geowebservices.stanford.edu/geoserver/wfs\",\"http://www.opengis.net/def/serviceType/ogc/wms\":\"https://geowebservices.stanford.edu/geoserver/wms\"}",
+  "gbl_wxsIdentifier_s":"druid:cz128vq0535",
+  "id":"stanford-cz128vq0535",
+  "layer_geom_type_s":"Polygon",
+  "gbl_mdModified_dt":"2015-01-13T18:46:38Z",
+  "dct_format_s":"Shapefile",
+  "dct_language_sm":"English",
+  "dc_type_s":"Dataset",
+  "dct_publisher_sm":"Uganda Bureau of Statistics",
+  "dct_creator_sm":[
+    "Uganda Bureau of Statistics"
+  ],
+  "dct_subject_sm":[
+    "Poverty",
+    "Statistics"
+  ],
+  "dct_issued_s":"2005",
+  "dct_temporal_sm":[
+    "2005"
+  ],
+  "dct_spatial_sm":[
+    "Uganda"
+  ],
+  "solr_geom":"ENVELOPE(29.572742, 35.000308, 4.234077, -1.478794)",
+  "gbl_indexYear_im":2005
+}

spec/fixtures/json_docs.rb

@@ -15,6 +15,12 @@ def full_geoblacklight
     File.read(File.join(File.dirname(__FILE__), './docs/full_geoblacklight.json'))
   end
 
+  ##
+  #  full_geoblacklight fixture converted to the aardvark schema
+  def full_geoblacklight_aardvark
+    File.read(File.join(File.dirname(__FILE__), './docs/full_geoblacklight_aardvark.json'))
+  end
+
   ##
   # A sample Esri OpenData metadata record
   def esri_opendata_metadata

spec/lib/geo_combine/migrators/v1_aardvark_migrator_spec.rb

@@ -0,0 +1,22 @@
+# frozen_string_literal: true
+
+require 'spec_helper'
+
+RSpec.describe GeoCombine::Migrators::V1AardvarkMigrator do
+  include JsonDocs
+
+  describe '#run' do
+    it 'migrates keys' do
+      input_hash = JSON.parse(full_geoblacklight)
+      # TODO: Note that this fixture has not yet been fully converted to
+      # aardvark. See https://github.com/OpenGeoMetadata/GeoCombine/issues/121
+      # for remaining work.
+      expected_output = JSON.parse(full_geoblacklight_aardvark)
+      expect(described_class.new(v1_hash: input_hash).run).to eq(expected_output)
+    end
+
+    context 'when the given record is already in aardvark schema' do
+      xit 'returns the record unchanged'
+    end
+  end
+end