traject_horizon 0.11.0 → 0.11.1

data/lib/traject/horizon_reader.rb

@@ -5,6 +5,7 @@ require 'traject/indexer/settings'
 require 'traject/horizon_bib_auth_merge'
 
 require 'marc'
+require 'marc/marc4j' # for marc4j jars
 
 module Traject
   #
@@ -133,7 +134,8 @@ module Traject
     def require_jars!
         Traject::Util.jruby_ensure_init!("Traject::HorizonReader")
 
-        Traject::Util.require_marc4j_jars(settings)
+        # ask marc-marc4j gem to load the marc4j jars
+        MARC::MARC4J.new(:jardir => settings['marc4j_reader.jar_dir'])
 
         # For some reason we seem to need to java_import it, and use
         # a string like this. can't just refer to it by full
@@ -228,7 +230,7 @@ module Traject
 
       # Turn Horizon's weird escaping into UTF8: <U+nnnn> where nnnn is a hex unicode
       # codepoint, turn it UTF8 for that codepoint
-      if settings["horizon.codepoint_translate"].to_s == "true" && settings["horizon.destination_encoding"] == "UTF8"
+      if false && settings["horizon.codepoint_translate"].to_s == "true" && settings["horizon.destination_encoding"] == "UTF8"
         text.gsub!(/\<U\+([0-9A-F]{4})\>/) do
           [$1.hex].pack("U")
         end
@@ -314,7 +316,10 @@ module Traject
         text     = rs.getBytes("longtext") || rs.getBytes("text")
 
         if tag == "000"
-          record.leader =  String.from_java_bytes text
+          # Horizon puts a \x1E marc field terminator on the end of hte
+          # leader in the db too, but that's not really part of it.
+          record.leader =  String.from_java_bytes(text).chomp("\x1E")
+
           fix_leader!(record.leader)
         elsif tag != "001"
           # we add an 001 ourselves with bib id in another part of code.
@@ -567,6 +572,14 @@ module Traject
       if settings['horizon.destination_encoding'] == "UTF8"
         leader[9] = 'a'
       end
+
+      # Do not understand why this voodoo that should be a no-op is neccesary,
+      # but get a mysterious and hard to isolate/reproduce encoding
+      # bug without it, but not with it. Think it may be the same
+      # but as this:
+      # https://github.com/jruby/jruby/issues/886
+      leader.force_encoding(leader.encoding)
+
     end
 
     def include_some_holdings?

data/lib/traject_horizon/version.rb

@@ -1,3 +1,3 @@
 module TrajectHorizon
-  VERSION = "0.11.0"
+  VERSION = "0.11.1"
 end

data/traject_horizon.gemspec

@@ -19,6 +19,8 @@ Gem::Specification.new do |spec|
 
   spec.add_dependency "traject"
 
+  spec.add_dependency "marc-marc4j" # for marc4j jar files
+
   spec.add_development_dependency "bundler", "~> 1.3"
   spec.add_development_dependency "rake"
   spec.add_development_dependency "minitest"

metadata

@@ -2,14 +2,14 @@
 name: traject_horizon
 version: !ruby/object:Gem::Version
   prerelease:
-  version: 0.11.0
+  version: 0.11.1
 platform: ruby
 authors:
 - Jonathan Rochkind
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2013-09-12 00:00:00.000000000 Z
+date: 2013-09-16 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: traject
@@ -27,6 +27,22 @@ dependencies:
     none: false
   prerelease: false
   type: :runtime
+- !ruby/object:Gem::Dependency
+  name: marc-marc4j
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+    none: false
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+    none: false
+  prerelease: false
+  type: :runtime
 - !ruby/object:Gem::Dependency
   name: bundler
   version_requirements: !ruby/object:Gem::Requirement