geohydra 0.3

data/lib/geohydra/gazetteer.csv

@@ -0,0 +1,36 @@
+"geonames_kw","geonames_id","lc_kw","lc_id"
+"Earth",6295630,"Earth (Planet)",lcsh:85040427
+"United States",6252001,"United States",
+"India",1269750,"India",
+"State of Arunāchal Pradesh",1278341,,
+"State of Assam",1278253,"Assam (India)",
+"State of  Bihār",1275715,"Bihar (India)",
+"Bilāspur",1275638,,
+"Mumbai",1275339,"Mumbai (India)",
+"Union Territory of Chandīgarh",1274744,,
+"Union Territory of Damān and Diu",1271155,"Daman and Diu (India)",
+"National Capital Territory of Delhi",1273293,"Delhi (India)",
+"Goa",1271157,"Goa (India : State)",
+"State of Gujarāt",1270770,"Gujarat (India)",
+"Haryana",1270260,"Haryana (India)",
+"State of Himāchal Pradesh",1270101,"Himachal Pradesh (India)",
+"State of Jammu and Kashmīr",1269320,"Jammu and Kashmir (India)",
+"State of Jharkhand",1444365,"Jharkhand (India)",
+"State of Karnātaka",1267701,"Karnataka (India)",
+"State of Kerala",1267254,"Kerala (India)",
+"Kachchh",1268731,"Kachchh (India)",
+"State of Madhya Pradesh",1264542,"Madhya Pradesh (India)",
+"State of Mahārāshtra",1264418,"Maharashtra (India)",
+"State of Manipur",1263706,"Manipur (India)",
+"Meghalaya",1263207,,
+"Mizoram",1262963,,
+"Nagaland",1262271,,
+"Orissa",1261029,,
+"Patiala",1260108,,
+"Punjab",1259223,,
+"Rajasthan",1258899,,
+"Sikkim",1256312,,
+"Tamil Nadu",1255053,,
+"Tripura",1254169,,
+"Uttar Pradesh",1253626,,
+"West Bengal",1252881,,

data/lib/geohydra/gazetteer.rb

@@ -0,0 +1,73 @@
+# encoding: UTF-8
+
+require 'csv'
+require 'awesome_print'
+
+module GeoHydra
+  class Gazetteer
+    
+    CSV_FN = File.join(File.dirname(__FILE__), 'gazetteer.csv')
+    
+    def initialize
+      @registry = {}
+      CSV.foreach(CSV_FN, :encoding => 'UTF-8') do |v|
+        @registry[v[0]] = {
+          :id => v[1].to_i,
+          :lc => v[2],
+          :lcid => v[3]
+        }
+      end
+    end
+    
+    def _get(k, i)
+      return nil unless @registry.include?(k)
+      @registry[k][i]
+    end
+    
+    # @return [Integer] geonames id
+    def find_id_by_keyword(k)
+      _get(k, :id)
+    end
+
+    # @return [String] library of congress name
+    def find_lc_by_keyword(k)
+      _get(k, :lc)
+    end
+    
+    # @return [String] library of congress valueURI
+    def find_lcuri_by_keyword(k)
+      lcid = _get(k, :lcid)
+      if lcid =~ /^lcsh:(\d+)$/
+        "http://id.loc.gov/authorities/subjects/sh#{$1}"
+      elsif lcid =~ /^lcnaf:(\d+)$/
+        "http://id.loc.gov/authorities/names/n#{$1}"
+      else
+        nil
+      end
+    end
+    
+    def find_lcauth_by_keyword(k)
+      lcid = _get(k, :lcid)
+      return $1 if lcid =~ /^(lcsh|lcnaf):/
+      return 'lcsh' unless find_lc_by_keyword(k).nil? # default to lcsh if present
+      nil
+    end
+    
+
+    # @see http://www.geonames.org/ontology/documentation.html
+    # @return [String] geonames uri (includes trailing / as specified)
+    def find_uri_by_keyword(k)
+      return nil if _get(k, :id).nil?
+      "http://sws.geonames.org/#{_get(k, :id)}/"
+    end
+  
+    # @return [String] The keyword
+    def find_keyword_by_id(id)
+      @registry.each do |k,v|
+        return k if v[:id] == id
+      end
+      nil
+    end
+  end
+end
+

data/lib/geohydra/geonetwork.rb

@@ -0,0 +1,156 @@
+require 'rubygems'
+require 'nokogiri'
+require 'rest_client'
+require 'awesome_print'
+
+module GeoHydra
+  # Provides client interface to GeoNetwork's REST API
+  # @see http://geonetwork-opensource.org/manuals/2.8.0/eng/developer/xml_services XML Services documentation
+  class GeoNetwork  
+    # @see http://geonetwork-opensource.org/manuals/2.8.0/eng/developer/xml_services/services_site_info_forwarding.html#status Valid status descriptions
+    GEONETWORK_STATUS_CODES = %w{unknown draft approved retired submitted rejected}
+    
+    # @see  http://geonetwork-opensource.org/manuals/2.8.0/eng/developer/xml_services/services_site_info_forwarding.html#xml-info Valid info tags
+    GEONETWORK_INFO_CODES = %w{site users groups sources schemas categories operations regions status}
+    
+    # @param [Hash] options provides `:service_root` URL
+    def initialize options = {}
+      @service_root = options[:service_root] || GeoHydra::Config.geonetwork.service_root
+      ap({:service_root => @service_root}) if $DEBUG
+    end
+    
+    # @see  http://geonetwork-opensource.org/manuals/2.8.0/eng/developer/xml_services/services_site_info_forwarding.html#site-information-xml-info Site Information: `xml.info`
+    def site_info
+      service("xml.info", { :type => 'site' }).xpath('/info/site')
+    end
+
+    # Fetches metrics
+    def metrics
+      service("../../monitor/metrics", { :pretty => 'true' })
+    end
+
+    # @yield the UUID (fileIdentifier) for all metadata in the GeoNetwork database
+    def each
+      xml = service("xml.search", { :remote => 'off', :hitsPerPage => -1 })
+      xml.xpath('//uuid/text()').each do |uuid|
+        yield uuid.to_s.strip
+      end
+    end
+    
+    # @see http://geonetwork-opensource.org/manuals/2.8.0/eng/developer/xml_services/metadata_xml_search_retrieve.html Metadata retrieval: `xml.metadata.get`
+    # @see http://geonetwork-opensource.org/manuals/2.8.0/eng/developer/xml_services/metadata_xml_status.html Metadata status: `xml.metadata.status.get`
+    #
+    # @param  [String] uuid the UUID (fileIdentifier) in the GeoNetwork database
+    # @param  [Boolean] include_status 
+    # @return [Struct] with the following fields:
+    #    * `:content` as the XML metadata
+    #    * `:status` as the status value, 
+    #    * `:druid` as the druid identifier
+    def fetch(uuid, include_status = false)
+      status = nil
+      if include_status
+        xml = service('xml.metadata.status.get', { :uuid => uuid })
+        xml.xpath('/response/record/statusid') do |id|
+          status = GEONETWORK_STATUS_CODES[id.to_i]
+        end
+      end
+      doc = service('xml.metadata.get', { :uuid => uuid })
+      if doc.xpath('//gmd:MD_Metadata').empty?
+        raise ArgumentError, "#{uuid} not in ISO 19139 format"
+      end
+      doc.xpath('//geonet:info').each { |x| x.remove } # GeoNetwork additions
+      
+      druid = nil
+      doc.xpath(
+        '//gmd:MD_Metadata/' + 
+        'gmd:dataSetURI/' + 
+        'gco:CharacterString[starts-with(text(),"http://purl.stanford.edu")]/' + 
+        'text()').each do |i|
+        druid = to_druid(i.to_s)
+      end
+      raise ArgumentError, "Cannot extract DRUID: #{uuid}" if druid.nil?
+      Struct.new(:content, :status, :druid).new(doc, status, druid)
+    end
+    
+    # @see http://geonetwork-opensource.org/manuals/2.8.0/eng/developer/xml_services/services_mef.html#mef-services  MEF Service
+    # @see http://geonetwork-opensource.org/manuals/2.8.0/eng/developer/xml_services/csw_services.html CSW service
+    # @param [String] uuid
+    # @param [String] dir directory into which method will save export file(s)
+    # @param [Symbol] format -- Either `:mef` or `:csw`
+    def export(uuid, dir = ".", format = :mef)
+      case format
+      when :mef then
+        export_mef(uuid, dir)
+      when :csw then
+        export_csw(uuid, dir)
+      else
+        raise ArgumentError, "Unsupported export format #{format}"        
+      end
+    end
+    
+    # @see  http://geonetwork-opensource.org/manuals/2.8.0/eng/developer/xml_services/system_configuration.html#system-configuration 
+    #   System Configuration
+    # @param types [Array] see {GeoHydra::GeoNetwork::GEONETWORK_INFO_CODES}
+    def info(types = GEONETWORK_INFO_CODES)
+      r = {}
+      types.each do |t|
+        if GEONETWORK_INFO_CODES.include?(t)
+          r[t] = service("xml.info", { :type => t })
+        else
+          raise ArgumentError, "#{t} is not a supported type for xml.info REST service"
+        end
+      end
+      r
+    end
+    
+  private
+    
+    def service(name, params, format = :default)
+      if format == :default and name.start_with?('xml.')
+        format = :xml
+      end
+      uri = "#{@service_root}/srv/eng/#{name}"
+      
+      ap({ :uri => uri, :params => params, :format => format }) if $DEBUG
+      
+      r = RestClient.get uri, :params => params
+      if format == :xml
+        Nokogiri::XML(r)
+      elsif format == :default
+        r
+      else
+        raise ArgumentError, "service requires format valid parameter: #{format}"
+      end
+    end
+    
+    def export_mef(uuid, dir = ".")
+      res = service("mef.export", { 
+        :uuid => uuid, 
+        :version => 'true',
+        :relation => 'false'
+      })
+      fn = "#{dir}/#{uuid}.mef"
+      File.open(fn, 'wb') {|f| f.puts(res.body) }
+      raise ArgumentError, "MEF #{fn} is missing" unless File.exist? fn
+      fn
+    end
+  
+    def export_csw(uuid, dir = ".")
+      res = service("csw", { 
+        :request => 'GetRecordById',
+        :service => 'CSW',
+        :version => '2.0.2',
+        :elementSetName => 'full',
+        :id => uuid 
+      })
+      fn = "#{dir}/#{uuid}.csw"
+      File.open(fn, 'wb') {|f| f.write(res.body) }
+      raise ArgumentError, "CSW #{fn} is missing" unless File.exist? fn
+      fn
+    end
+    
+    def to_druid(purl)
+      purl.to_s.downcase.gsub(%r{^(http://purl.stanford.edu.*)/([a-z0-9]{11})$}, "\\2")
+    end
+  end  
+end

data/lib/geohydra/mods2ogp.xsl

@@ -0,0 +1,198 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!--
+     mods2ogp.xsl - Transforms MODS with GML extensions into an OGP Solr document
+
+     Copyright 2013, Stanford University Libraries.
+
+     Created by Darren Hardy.
+
+     For OGP Solr schema, see:
+
+       https://github.com/OpenGeoportal/ogpSolrConfig/blob/master/ogpSolrConfig/SolrConfig/schema.xml
+
+     Example usage:
+
+     xsltproc -stringparam geoserver_root 'http://kurma-podd1.stanford.edu/geoserver'
+              -stringparam purl 'http://purl-dev.stanford.edu/fw920bc5473'
+              -output '/var/geomdtk/current/workspace/fw/920/bc/5473/fw920bc5473/temp/ogpSolr.xml'
+              '/home/geostaff/geomdtk/current/lib/geomdtk/mods2ogp.xsl'
+              '/var/geomdtk/current/workspace/fw/920/bc/5473/fw920bc5473/metadata/descMetadata.xml'
+
+     Requires parameters:
+
+       - geoserver_root - URL prefix to the geoserver
+       - purl - complete URL with aa111bb1111 (len = 11)
+
+     -->
+<xsl:stylesheet xmlns="http://lucene.apache.org/solr/4/document" xmlns:gco="http://www.isotc211.org/2005/gco" xmlns:gmd="http://www.isotc211.org/2005/gmd" xmlns:gml="http://www.opengis.net/gml/3.2" xmlns:mods="http://www.loc.gov/mods/v3" xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#" xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0" exclude-result-prefixes="gmd gco gml mods rdf xsl">
+  <xsl:output method="xml" version="1.0" encoding="UTF-8" indent="yes"/>
+  <xsl:strip-space elements="*"/>
+  <xsl:template match="/mods:mods">
+    <xsl:variable name="druid" select="substring($purl, string-length($purl)-10)"/>
+    <xsl:variable name="datatype" select="substring-after(mods:extension/rdf:RDF[@rdf:type='geo']/rdf:Description[@rdf:type='geo#geometryType']/text(), 'gml:')"/>
+    <add>
+      <doc>
+        <field name="LayerId">
+          <!-- REQUIRED -->
+          <xsl:value-of select="$druid"/>
+        </field>
+        <field name="Name">
+          <!-- REQUIRED -->
+          <xsl:value-of select="$druid"/>
+        </field>
+        <field name="ExternalLayerId">
+          <xsl:value-of select="$purl"/>
+        </field>
+        <field name="CollectionId">
+          <xsl:value-of select="mods:relatedItem/mods:titleInfo/mods:title"/>
+        </field>
+        <field name="Access">
+          <!-- REQUIRED -->
+          <xsl:text>Restricted</xsl:text>
+        </field>
+        <field name="Institution">
+          <!-- REQUIRED -->
+          <xsl:text>Stanford</xsl:text>
+        </field>
+        <field name="WorkspaceName">
+          <xsl:text>druid</xsl:text>
+        </field>
+        <field name="GeoReferenced">
+          <xsl:text>true</xsl:text>
+        </field>
+        <field name="Availability">
+          <xsl:text>Online</xsl:text>
+        </field>
+        <xsl:comment>XXX - MODS data are year only but OGP's solr schema requires full date</xsl:comment>
+        <field name="ContentDate">
+          <xsl:choose>
+            <xsl:when test="mods:subject/mods:temporal">
+              <xsl:value-of select="substring(mods:subject/mods:temporal, 1, 4)"/>
+              <xsl:text>-01-01T00:00:00Z</xsl:text>
+            </xsl:when>
+            <xsl:when test="mods:originInfo/mods:dateIssued">
+              <xsl:value-of select="substring(mods:originInfo/mods:dateIssued, 1, 4)"/>
+              <xsl:text>-01-01T00:00:00Z</xsl:text>
+            </xsl:when>
+          </xsl:choose>
+        </field>
+        <field name="LayerDisplayName">
+          <!-- REQUIRED -->
+          <xsl:value-of select="mods:titleInfo/mods:title[not(@type)]"/>
+        </field>
+        <!-- <xsl:value-of select="'XXX'"/> -->
+        <xsl:for-each select="mods:extension[@displayLabel='geo']/rdf:RDF/rdf:Description/dc:type">
+          <field name="DataType">
+            <xsl:choose>
+              <xsl:when test="substring(., 9)='LineString'">
+                <xsl:text>Line</xsl:text>
+              </xsl:when>
+              <xsl:otherwise>
+                <xsl:value-of select='substring(., 9)'/><!-- strip Dataset# prefix -->
+              </xsl:otherwise>
+            </xsl:choose>
+          </field>
+        </xsl:for-each>
+        <field name="Publisher">
+          <xsl:value-of select="mods:originInfo/mods:publisher"/>
+        </field>
+        <field name="Originator">
+          <xsl:for-each select="mods:name">
+            <xsl:value-of select="mods:namePart"/>
+            <xsl:if test="position()!=last()">
+              <xsl:text>; </xsl:text>
+            </xsl:if>
+          </xsl:for-each>
+        </field>
+        <field name="Abstract">
+          <xsl:for-each select="mods:abstract[@displayLabel='Abstract' or @displayLabel='Purpose']/text()">
+            <xsl:value-of select="."/>
+            <xsl:text>; </xsl:text>
+          </xsl:for-each>
+        </field>
+        <field name="ThemeKeywords">
+          <xsl:for-each select="mods:subject/mods:topic">
+            <xsl:choose>
+              <xsl:when test="@authority='ISO19115topicCategory'">
+                <xsl:value-of select="@valueURI"/>
+              </xsl:when>
+              
+              <xsl:otherwise>
+                <xsl:value-of select="text()"/>
+              </xsl:otherwise>
+            </xsl:choose>
+            
+            <xsl:text>; </xsl:text>
+          </xsl:for-each>
+          
+        </field>
+        <field name="PlaceKeywords">
+          <xsl:for-each select="mods:subject/mods:geographic">
+            <xsl:value-of select="text()"/>
+            <xsl:if test="position()!=last()">
+              <xsl:text>; </xsl:text>
+            </xsl:if>
+          </xsl:for-each>
+        </field>
+        <xsl:for-each select="mods:extension/rdf:RDF[@rdf:type='geo']/rdf:Description[@rdf:type='geo#boundingBox']/gml:Envelope">
+          <xsl:variable name="x2" select="number(substring-before(gml:upperCorner/text(), ' '))"/>
+          <xsl:variable name="x1" select="number(substring-before(gml:lowerCorner/text(), ' '))"/>
+          <xsl:variable name="y2" select="number(substring-after(gml:upperCorner/text(), ' '))"/>
+          <xsl:variable name="y1" select="number(substring-after(gml:lowerCorner/text(), ' '))"/>
+          <field name="MinX">
+            <xsl:value-of select="$x1"/>
+          </field>
+          <field name="MinY">
+            <xsl:value-of select="$y1"/>
+          </field>
+          <field name="MaxX">
+            <xsl:value-of select="$x2"/>
+          </field>
+          <field name="MaxY">
+            <xsl:value-of select="$y2"/>
+          </field>
+          <xsl:comment> XXX: doesn't work across meridian </xsl:comment>
+          <field name="CenterX">
+            <xsl:value-of select="($x2 - $x1) div 2 + $x1"/>
+          </field>
+          <xsl:comment> XXX: doesn't work across meridian </xsl:comment>
+          <field name="CenterY">
+            <xsl:value-of select="($y2 - $y1) div 2 + $y1"/>
+          </field>
+          <xsl:comment> XXX: in degrees ??? </xsl:comment>
+          <field name="HalfWidth">
+            <xsl:value-of select="($x2 - $x1) div 2"/>
+          </field>
+          <xsl:comment> XXX: in degrees ??? </xsl:comment>
+          <field name="HalfHeight">
+            <xsl:value-of select="($y2 - $y1) div 2"/>
+          </field>
+          <xsl:comment> XXX: in degrees**2 ??? </xsl:comment>
+          <field name="Area">
+            <xsl:value-of select="round(($y2 - $y1) * ($x2 - $x1))"/>
+          </field>
+          <field name="SrsProjectionCode">
+            <xsl:value-of select="@srsName"/>
+          </field>
+        </xsl:for-each>
+        <field name="Location">
+          <!-- output is JSON so we wrap in CDATA -->
+          <xsl:text disable-output-escaping="yes">&lt;![CDATA[</xsl:text>
+          <xsl:text>{
+              "wms":       ["</xsl:text>
+          <xsl:value-of select="$geoserver_root"/>
+          <xsl:text>/wms"],
+              "wfs":       ["</xsl:text>
+          <xsl:value-of select="$geoserver_root"/>
+          <xsl:text>/wfs"],
+              "purl":      ["</xsl:text>
+          <xsl:value-of select="$purl"/>
+          <xsl:text>"] }</xsl:text>
+          <xsl:text disable-output-escaping="yes">]]&gt;</xsl:text>
+        </field>
+        <field name="FgdcText"/>
+      </doc>
+    </add>
+  </xsl:template>
+  <xsl:template match="*"/>
+</xsl:stylesheet>