traject_sequel_writer 0.11.0 → 0.12.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 81efee56a46a47e26c17eff759d4cc0a7dc36726
-  data.tar.gz: 72a8923c02d9e6052342c4c953fbc9cc34857ca2
+  metadata.gz: e73f97cb0e4558e71b2960c0d42accad74041758
+  data.tar.gz: 1e0e5f6ddacfef11a408e503ee987a7dc8f0ba35
 SHA512:
-  metadata.gz: 7e5a7a0710a5e8fa437bf4a8976119ef8f4a8a8995a7eb4633364eb17ef666e6472d44f5bef965c434d208ed9c3cd101638276f74572164bd53c3c03ae1629c3
-  data.tar.gz: ba50260b19700091db4d1827bcf80446595b40a7b53933312028c4524da1143af425c65f7a6bd481fb41b7322bb87355661effbc7cf25c1ddf2b691ed75cb616
+  metadata.gz: cdc2945a7cd9457e4a287e00948bd069bc6c9b081aac8686addb3ebeef6d8076ddf2bbb9b717385c575b0098f092f6c87735545abdd25b2b2c6cd9c33d37dc0f
+  data.tar.gz: 0ef683cf6ee481e62ad600d088919fd537817459793d49d3645e954291e546a95145dc32133e377bd50cfa52e0b2e15595aa24d18228d354005b50224b5fce82

data/README.md

@@ -9,6 +9,7 @@ The writer can be used as standard, as the destination of your indexing pipeline
 
 It was actually written for a use case where it's used as a "side effect" in a traject `each_record`, writing different data out to an rdbms on the side, while the main indexing is to Solr. This ends up a bit hacky at present but works.  
 
+Currently has a pre-1.0 release number, as it has not seen wide use, and may have some oddities. 
 
 ## Installation
 

data/lib/traject/sequel_writer.rb

@@ -140,11 +140,33 @@ module Traject
 
     def hash_to_array(column_names, hash)
       column_names.collect do |c| 
-        v = hash[c.to_s]
-        v.kind_of?(Array) ? v.join(@internal_delimiter) : v
+        output_value_to_column_value(hash[c.to_s])
       end
     end
 
+    # Traject context.output_hash values are arrays.
+    # turn them into good column values, joining strings if needed. 
+    #
+    # Single values also accepted, even though not traject standard, they
+    # will be passed through unchanged. 
+    def output_value_to_column_value(v)
+      if v.kind_of?(Array)
+        if v.length == 0
+          nil
+        elsif v.length == 1
+          v.first
+        elsif v.first.kind_of?(String)
+          v.join(@internal_delimiter)
+        else
+          # Not a string? Um, raise for now?
+          raise ArgumentError.new("Traject::SequelWriter, multiple non-String values provided: #{v}")
+        end
+      else
+        v
+      end
+    end
+
+
     def after_send_batch(&block)
       @after_send_batch_callbacks << block
     end

data/lib/traject_sequel_writer/version.rb

@@ -1,3 +1,3 @@
 module TrajectSequelWriter
-  VERSION = "0.11.0"
+  VERSION = "0.12.0"
 end

data/test/test_helper.rb

@@ -1,10 +1,9 @@
+require 'minitest'
 require 'minitest/autorun'
 require 'minitest/spec'
 
 require 'traject_sequel_writer'
 
-require 'minitest/autorun'
-
 require 'sequel'
 
 # Create a temporary sqlite db for tests, remove it when we're done

data/test/test_traject_sequel_writer.rb

@@ -41,6 +41,19 @@ describe "Traject::SequelWriter" do
       assert @writer.db_table.where(:string_a => "String_One,String_Two", :string_b => "String_B_One").count == 1, "Expected written row with expected values"
     end
 
+    it "complains about multiple non-string values" do
+      @writer = self.writer
+
+      context = Traject::Indexer::Context.new
+      context.output_hash.merge!(
+        "int_a" => [1001, 1002]
+      )
+      assert_raises(ArgumentError) do
+        @writer.put context  
+        @writer.close
+      end
+    end
+
     after do
       @writer.db_table.delete
     end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: traject_sequel_writer
 version: !ruby/object:Gem::Version
-  version: 0.11.0
+  version: 0.12.0
 platform: ruby
 authors:
 - Jonathan Rochkind