Merge branch 'release-1.6.0'

.travis.yml

@@ -1,6 +1,7 @@
 sudo: false
 language: ruby
 rvm:
+  - 2.3.1
   - 2.2
   - 2.1
   - 2.0.0

VERSION

@@ -1 +1 @@
-1.5.0
+1.6.0

lib/ezid/batch_download.rb

@@ -40,6 +40,7 @@ class BatchDownload < Hashie::Dash
 
     # Parameters
     property :format, required: true # {anvl|csv|xml}
+    property :compression            # {gzip|zip}
     property :column                 # repeatable
     property :notify                 # repeatable
     property :convertTimestamps      # {yes|no}
@@ -87,36 +88,45 @@ def download_file(path: nil)
       path ||= Dir.getwd
       fullpath = File.directory?(path) ? File.join(path, download_filename) : path
       tries = 0
-      begin
-        tries += 1
-        download = Net::HTTP.get_response(download_uri)
-        download.value
-      rescue Net::HTTPServerException => e
-        if download.is_a?(Net::HTTPNotFound)
-          if tries < MAX_DOWNLOAD_TRIES
-            print "Download file not yet available (attempt #{tries} of #{MAX_DOWNLOAD_TRIES})."
-            puts " Trying again in #{DOWNLOAD_RETRY_INTERVAL} second(s) ..."
-            sleep DOWNLOAD_RETRY_INTERVAL
-            retry
-          else
-            raise BatchDownloadError,
-                  "Maximum download attempts (#{MAX_DOWNLOAD_TRIES}) reached unsuccessfully."
+      ready = false
+
+      print "Checking for download "
+      Net::HTTP.start(download_uri.host, download_uri.port) do |http|
+        while tries < MAX_DOWNLOAD_TRIES
+          tries += 1
+          sleep DOWNLOAD_RETRY_INTERVAL
+          print "."
+          response = http.head(download_uri.path)
+          if response.code == '200'
+            ready = true
+            break
           end
-        else
-          raise
         end
-      else
-        File.open(fullpath, "wb") do |f|
-          f.write(download.body)
+      end
+      puts
+
+      unless ready
+        raise BatchDownloadError,
+              "Download not ready after checking #{MAX_DOWNLOAD_TRIES} times."
+      end
+
+      File.open(fullpath, "wb") do |f|
+        Net::HTTP.start(download_uri.host, download_uri.port) do |http|
+          http.request_get(download_uri.path) do |response|
+            response.read_body do |chunk|
+              f.write(chunk)
+            end
+          end
         end
-        puts "File successfully download to #{fullpath}."
       end
+
+      fullpath
     end
 
     private
 
     def download_uri
-      URI(download_url)
+      @download_uri ||= URI(download_url)
     end
 
     def download_filename

lib/ezid/batch_enumerator.rb

@@ -0,0 +1,42 @@
+module Ezid
+  class BatchEnumerator
+    include Enumerable
+
+    attr_reader :format, :batch_file
+
+    def initialize(format, batch_file)
+      @format = format
+      @batch_file = batch_file
+    end
+
+    def each(&block)
+      case format
+      when :anvl
+        each_anvl(&block)
+      when :xml
+        each_xml(&block)
+      when :csv
+        each_csv(&block)
+      end
+    end
+
+    def each_anvl(&block)
+      File.open(batch_file, "rb") do |f|
+        while record = f.gets("")
+          head, metadata = record.split(/\n/, 2)
+          id = head.sub(/\A::/, "").strip
+          yield Ezid::Identifier.new(id, metadata: metadata)
+        end
+      end
+    end
+
+    def each_xml
+      raise NotImplementedError
+    end
+
+    def each_csv
+      raise NotImplementedError
+    end
+
+  end
+end

spec/fixtures/anvl_batch.txt

@@ -0,0 +1,39 @@
+:: ark:/99999/fk4086hs23
+_updated: 1488227717
+_target: http://example.com
+_profile: erc
+_ownergroup: apitest
+_owner: apitest
+_export: yes
+_created: 1488227717
+_status: public
+
+:: ark:/99999/fk4086hs23/123
+_updated: 1488227718
+_target: http://ezid.cdlib.org/id/ark:/99999/fk4086hs23/123
+_profile: erc
+_ownergroup: apitest
+_owner: apitest
+_export: yes
+_created: 1488227718
+_status: public
+
+:: ark:/99999/fk40p1bb85
+_updated: 1488303565
+_target: http://ezid.cdlib.org/id/ark:/99999/fk40p1bb85
+_profile: erc
+_ownergroup: apitest
+_owner: apitest
+_export: yes
+_created: 1488303565
+_status: public
+
+:: ark:/99999/fk40z7fh7x
+_updated: 1488232856
+_target: http://ezid.cdlib.org/id/ark:/99999/fk40z7fh7x
+_profile: erc
+_ownergroup: apitest
+_owner: apitest
+_export: yes
+_created: 1488232856
+_status: public

spec/integration/batch_download_spec.rb

@@ -0,0 +1,21 @@
+require 'tempfile'
+
+module Ezid
+  RSpec.describe BatchDownload do
+
+    subject do
+      a_week_ago = (Time.now - (7*24*60*60)).to_i
+      described_class.new(:anvl, compression: "zip", permanence: "test", status: "public", createdAfter: a_week_ago)
+    end
+
+    its(:download_url) { is_expected.to match(/\Ahttp:\/\/ezid\.cdlib\.org\/download\/\w+\.zip\z/) }
+
+    specify {
+      Dir.mktmpdir do |tmpdir|
+        expect(subject.download_file(path: tmpdir))
+          .to match(/\A#{tmpdir}\/\w+\.zip\z/)
+      end
+    }
+
+  end
+end

spec/unit/batch_enumerator_spec.rb

@@ -0,0 +1,34 @@
+require 'ezid/batch_enumerator'
+
+module Ezid
+  RSpec.describe BatchEnumerator do
+
+    let(:batch_file) { File.expand_path("../../fixtures/anvl_batch.txt", __FILE__) }
+
+    subject { described_class.new(:anvl, batch_file) }
+
+    its(:count) { is_expected.to eq 4 }
+
+    specify {
+      subject.each do |id|
+        expect(id).to be_a(Identifier)
+      end
+    }
+
+    specify {
+      batch_array = subject.to_a
+      expect(batch_array.length).to eq 4
+    }
+
+    specify {
+      ids = subject.map(&:id)
+      expect(ids).to eq ["ark:/99999/fk4086hs23", "ark:/99999/fk4086hs23/123", "ark:/99999/fk40p1bb85", "ark:/99999/fk40z7fh7x"]
+    }
+
+    specify {
+      id = subject.first
+      expect(id.target).to eq "http://example.com"
+    }
+
+  end
+end