scalastic 0.1.0 → 0.2.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 8d4a081eb653e2378d6a39f4f77cd51f44b553b0
-  data.tar.gz: c38641e0fe62a596f2d631a7edbfc6c4276785b7
+  metadata.gz: fd3dbca7672a601bba49bc18531c66f5d4427710
+  data.tar.gz: 3669c3673f0ded567a2acec6993028195a3b2c44
 SHA512:
-  metadata.gz: a46a54dd509265a69b621b5596bd21c23b3d766f1d4c479faa28436e69b5650bb133dba236316860eb8ae48cca864fdf9b1f06b1d20f28bfb898162287cefb03
-  data.tar.gz: 252794ac6ea2ccb301636a90953e8a31ed2a3186ea079ba1c4894ffbc4ea86ddf41075e89f3e37d49a527ef7f5c05992c40bd46e86f4cc9e694d06a5920b330d
+  metadata.gz: d5ea03aa00e7513b8c33359480d1d3e49edeaf62e5408362ed4d156666adc4573dbb89872fa3824c838570407768f65ddd0dba52cf723621d969e6ad6602aeb3
+  data.tar.gz: 46cfbf9227dc75f94e6e2ea0defa7979e8ddf55cd38f31817b047e043b979983badf578d10e098ed784f317186a8ea4162ee69e40f8a5b6743d7270a355d2b68

data/README.md

@@ -135,6 +135,67 @@ count = partition1.search(search_type: 'count', body: {query: {match_all: {}}})[
 raise "Expected 1 document, got #{count}" unless count == 1
 ```
 
+### Bulk operations
+```ruby
+client = Elasticsearch::Client.new
+partitions = client.partitions
+
+client.indices.create(index: 'bulk_operations')
+partitions.prepare_index(index: 'bulk_operations')
+
+partition = partitions.create(index: 'bulk_operations', id: 1)
+
+partition.bulk(body: [
+  {index: {_type: 'test', _id: 1, data: {subject: 'test1'}}},
+  {create: {_type: 'test', _id: 2, data: {subject: 'test2'}}}
+])
+
+partition.bulk(body: [
+  {index: {_type: 'test', _id: 3}},
+  {subject: 'test3'},
+  {create: {_type: 'test', _id: 4}},
+  {subject: 'test4'}
+])
+
+partition.bulk(body: [
+  {update: {_type: 'test', _id: 1, data: {doc: {body: 'Document 1'}}}},
+  {update: {_type: 'test', _id: 2, data: {doc: {body: 'Document 2'}}}}
+])
+
+partition.bulk(body: [
+  {update: {_type: 'test', _id: 3}},
+  {doc: {body: 'Document 3'}},
+  {update: {_type: 'test', _id: 4}},
+  {doc: {body: 'Document 4'}}
+])
+
+client.indices.refresh    # Commit all pending writes
+
+hits = partition.search['hits']['hits'].sort{|h1, h2| h1['_id'].to_i <=> h2['_id'].to_i}
+raise 'Unexpected count' unless hits.size == 4
+
+expected_hits = [
+  {'_index' => 'bulk_operations', '_type' => 'test', '_id' => '1', '_score' => 1.0, '_source' => {'subject' => 'test1', 'body' => 'Document 1', 'scalastic_partition_id' => 1}},
+  {'_index' => 'bulk_operations', '_type' => 'test', '_id' => '2', '_score' => 1.0, '_source' => {'subject' => 'test2', 'body' => 'Document 2', 'scalastic_partition_id' => 1}},
+  {'_index' => 'bulk_operations', '_type' => 'test', '_id' => '3', '_score' => 1.0, '_source' => {'subject' => 'test3', 'body' => 'Document 3', 'scalastic_partition_id' => 1}},
+  {'_index' => 'bulk_operations', '_type' => 'test', '_id' => '4', '_score' => 1.0, '_source' => {'subject' => 'test4', 'body' => 'Document 4', 'scalastic_partition_id' => 1}},
+]
+
+raise 'Unexpected results' unless hits == expected_hits
+
+res = partition.bulk(body: [
+  {delete: {_type: 'test', _id: 1}},
+  {delete: {_type: 'test', _id: 2}},
+  {delete: {_type: 'test', _id: 3}},
+  {delete: {_type: 'test', _id: 4}},
+])
+
+client.indices.refresh    # Commit all pending writes
+
+count = partition.search(search_type: 'count')['hits']['total']
+raise 'Some documents were not removed' unless count == 0
+```
+
 ### Notes
 * Indices must be *prepared* before they can be used by Scalastic by calling "prepare" on the partitions client; doing so will create critical field mappings. Each index must be prepared only once. 
 * All hash keys in arguments must be symbols; using anything else may result in unexpected behavior.

data/lib/scalastic/config.rb

@@ -9,7 +9,7 @@ module Scalastic
 
     def initialize
       @partition_prefix = "scalastic"
-      @partition_selector = "scalastic_partition_id"
+      @partition_selector = :scalastic_partition_id
     end
 
     def index_endpoint(partition_id)

data/lib/scalastic/partition.rb

@@ -51,10 +51,52 @@ module Scalastic
       all_aliases.any?{|_index, data| data['aliases'].any?}
     end
 
-    #TODO: add bulk
+    def bulk(args)
+      body = args.clone[:body] || raise(ArgumentError, 'Missing required argument :body')
+      index = config.index_endpoint(id)
+      selector = config.partition_selector
+
+      new_ops = body.map{|entry| [operation_name(entry), entry]}.reduce([]){|acc, op| acc << [op.first, update_entry(acc, *op)]; acc}
+      args[:body] = new_ops.map{|_op, entry| entry}
+
+      es_client.bulk(args)
+    end
 
     def inspect
       "ES partition #{id}"
     end
+
+    private
+
+    def operation_name(entry)
+      [:create, :index, :update, :delete].find{|name| entry.has_key?(name)}
+    end
+
+    def update_entry(acc, operation, entry)
+      if (operation)
+        op_data = entry[operation]
+        op_data[:_index] = config.index_endpoint(id)
+        if (op_data[:data])
+          if (operation == :update)
+            op_data[:data][:doc] ||= {}
+            op_data[:data][:doc][config.partition_selector] = id
+          else
+            op_data[:data][config.partition_selector] = id
+          end
+        end
+      else
+        parent = acc.last
+        # A previous record must be create/index/update/delete
+        raise(ArgumentError, "Unexpected entry: #{entry}") unless parent && parent.first
+
+        if (parent.first == :update)
+          entry[:doc] ||= {}
+          entry[:doc][config.partition_selector] = id
+        else
+          entry[config.partition_selector] = id
+        end
+      end
+      entry
+    end
   end
 end

data/lib/scalastic/version.rb

@@ -1,3 +1,3 @@
 module Scalastic
-  VERSION = "0.1.0"
+  VERSION = "0.2.0"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: scalastic
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.2.0
 platform: ruby
 authors:
 - Aliaksei Baturytski
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2016-01-12 00:00:00.000000000 Z
+date: 2016-01-18 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler