solr_ead 0.4.4 → 0.4.5

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    solr_ead (0.4.4)
+    solr_ead (0.4.5)
       om
       rsolr
       sanitize

data/History.txt

@@ -0,0 +1,6 @@
+0.4.5
+New Features:
+-  Ability to pass in a custom component and the simple param to the rake task
+
+Fixes:
+- fixed greedy regex to allow for "messier" ead xml files

data/README.md

@@ -39,6 +39,8 @@ Or install it yourself:
 ## Usage
 
     $ rake solr_ead:index FILE=/path/to/your/ead.xml
+    $ rake solr_ead:index_dir DIR=/path/to/your/eads SIMPLE=true
+    $ rake solr_ead:index_dir DIR=/path/to/your/eads SOLR_URL=http://127.0.0.1:8983
 
 You can also do this via the command line:
 
@@ -121,9 +123,29 @@ following content:
 From the console, index you ead document using your new definition.
 
     > file = "path/to/ead.xml"
-    > indexer = SolrEad::Indexer.new(:document=>"CustomDocument")
+    > indexer = SolrEad::Indexer.new(:document=>CustomDocument)
     > indexer.create(file)
 
+Or index from the rake task 
+
+    $ rake solr_ead:index FILE=path/to/file.xml CUSTOM_DOCUMENT=path/to/custom_document.rb
+
+### Writing a custom component definition
+
+Similar to the custom document definition, you can create a custom component definition for component indexing:
+
+    class CustomComponent < SolrEad::Component
+      ...
+    end
+
+Call this from the console
+
+    > indexer = SolrEad::Indexer.new(:document=>CustomDocument, :component=>CustomComponent)
+
+Or from the rake task
+
+    $ rake solr_ead:index FILE=path/to/file.xml CUSTOM_DOCUMENT=path/to/custom_document.rb CUSTOM_COMPONENT=path/to/custom_component.rb
+
 ### Adding custom methods
 
 Suppose you want to add some custom methods that perform additional manipulations of

data/lib/solr_ead/behaviors.rb

@@ -7,7 +7,7 @@ module SolrEad::Behaviors
   # It'll make an attempt at substituting numbered component levels for non-numbered
   # ones.
   def components(file)
-    raw = File.read(file).gsub!(/xmlns=".*"/, '')
+    raw = File.read(file).gsub!(/xmlns="(.*?)"/, '')
     raw.gsub!(/c[0-9]{2,2}/,"c")
     xml = Nokogiri::XML(raw)
     return xml.xpath("//c")

data/lib/solr_ead/version.rb

@@ -1,3 +1,3 @@
 module SolrEad
-  VERSION = "0.4.4"
+  VERSION = "0.4.5"
 end

data/lib/tasks/solr_ead.rake

@@ -29,16 +29,20 @@ end
 
 # Set up a new indexer object
 # 
-# If CUSTOM_DOCUMENT is present, require the file and instantiate the indexer with it
-# Otherwise instantiate a default indexer
+# Instantiate a new indexer object with a custom document, component and/or simple switch if present
 def load_indexer
+  options = {}
   if ENV['CUSTOM_DOCUMENT']
     raise "Please specify a valid file for your custom document." unless File.exists? ENV['CUSTOM_DOCUMENT']
     require File.join(Rails.root, ENV['CUSTOM_DOCUMENT'])
-    custom_document = File.basename(ENV['CUSTOM_DOCUMENT']).split(".").first.classify.constantize
-    indexer = SolrEad::Indexer.new(:document=>custom_document)
-  else
-    indexer = SolrEad::Indexer.new
+    options[:document] = File.basename(ENV['CUSTOM_DOCUMENT']).split(".").first.classify.constantize
+  end
+  if ENV['CUSTOM_COMPONENT']
+    raise "Please specify a valid file for your custom component." unless File.exists? ENV['CUSTOM_COMPONENT']
+    require File.join(Rails.root, ENV['CUSTOM_COMPONENT'])
+    options[:component] = File.basename(ENV['CUSTOM_COMPONENT']).split(".").first.classify.constantize
   end
+  options[:simple] = (ENV['SIMPLE'] or ENV['SIMPLE']=="true") ? true : false
+  indexer = SolrEad::Indexer.new(options)
   return indexer
 end

data/spec/behaviors_spec.rb

@@ -5,6 +5,7 @@ describe SolrEad::Behaviors do
   before :all do
     @not_numbered = fixture "ARC-0005.xml"
     @numbered     = fixture "pp002010.xml"
+    @messy        = fixture "ead_messy_format.xml"
     class TestClass
       include SolrEad::Behaviors
     end
@@ -16,6 +17,7 @@ describe SolrEad::Behaviors do
     before :all do
       @non_numbered_nodeset = @test.components(@not_numbered)
       @numbered_nodeset     = @test.components(@numbered)
+      @messy_nodeset        = @test.components(@messy)
     end
 
     it "should return a nodeset" do
@@ -27,6 +29,11 @@ describe SolrEad::Behaviors do
       @non_numbered_nodeset.count.should == 135
       @numbered_nodeset.count.should == 83
     end
+    
+    it "should find some components even if ead is messily formatted" do
+      @messy_nodeset.count.should > 0
+    end
+    
   end
 
   describe "#prep" do

data/spec/fixtures/ead_messy_format.xml

@@ -0,0 +1,27 @@
+<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
+<ead xsi:schemaLocation="urn:isbn:1-931666-22-9 http://www.loc.gov/ead/ead.xsd" xmlns:ns2="http://www.w3.org/1999/xlink" xmlns="urn:isbn:1-931666-22-9" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">  <eadheader findaidstatus="Complete" repositoryencoding="iso15511" countryencoding="iso3166-1" dateencoding="iso8601" langencoding="iso639-2b">
+    <eadid>sample_ead2</eadid>
+    <filedesc></filedesc>
+    <profiledesc></profiledesc>
+    <revisiondesc></revisiondesc>
+  </eadheader>
+  <archdesc>
+    <dsc>
+      <c id="series1" level="series">
+        <did>
+          <unittitle>sample series</unittitle>
+          <unitdate>sample series date</unitdate>
+        </did>
+        <scopecontent>
+          <head>scopecontent heading</head>
+          <p>Sample scopecontent text</p>
+        </scopecontent>
+      </c>
+    </dsc>
+  </archdesc>
+  <!-- Not found in AT EAD -->
+  <frontmatter></frontmatter>
+</ead>
+
+
+

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: solr_ead
 version: !ruby/object:Gem::Version
-  version: 0.4.4
+  version: 0.4.5
   prerelease: 
 platform: ruby
 authors:
@@ -9,7 +9,7 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-04-22 00:00:00.000000000 Z
+date: 2013-05-07 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: om
@@ -184,6 +184,7 @@ files:
 - .rvmrc
 - Gemfile
 - Gemfile.lock
+- History.txt
 - LICENSE
 - README.md
 - Rakefile
@@ -205,6 +206,7 @@ files:
 - spec/fixtures/ARC-0005.xml
 - spec/fixtures/ARC-0161.xml
 - spec/fixtures/component_template.xml
+- spec/fixtures/ead_messy_format.xml
 - spec/fixtures/ead_sample.xml
 - spec/fixtures/ead_template.xml
 - spec/fixtures/pp002010.xml
@@ -224,7 +226,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: 2881959855902156324
+      hash: 3066311826305966167
 required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
   requirements:
@@ -233,7 +235,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: 2881959855902156324
+      hash: 3066311826305966167
 requirements: []
 rubyforge_project: 
 rubygems_version: 1.8.23
@@ -247,6 +249,7 @@ test_files:
 - spec/fixtures/ARC-0005.xml
 - spec/fixtures/ARC-0161.xml
 - spec/fixtures/component_template.xml
+- spec/fixtures/ead_messy_format.xml
 - spec/fixtures/ead_sample.xml
 - spec/fixtures/ead_template.xml
 - spec/fixtures/pp002010.xml