hallelujah-cassandra 0.12.3

data/lib/cassandra/ordered_hash.rb

@@ -0,0 +1,192 @@
+# OrderedHash is namespaced to prevent conflicts with other implementations
+class Cassandra
+    class OrderedHashInt < Hash #:nodoc:
+      def initialize(*args, &block)
+        super
+        @keys = []
+      end
+
+      def self.[](*args)
+        ordered_hash = new
+
+        if (args.length == 1 && args.first.is_a?(Array))
+          args.first.each do |key_value_pair|
+            next unless (key_value_pair.is_a?(Array))
+            ordered_hash[key_value_pair[0]] = key_value_pair[1]
+          end
+
+          return ordered_hash
+        end
+
+        unless (args.size % 2 == 0)
+          raise ArgumentError.new("odd number of arguments for Hash")
+        end
+
+        args.each_with_index do |val, ind|
+          next if (ind % 2 != 0)
+          ordered_hash[val] = args[ind + 1]
+        end
+
+        ordered_hash
+      end
+
+      def initialize_copy(other)
+        super
+        # make a deep copy of keys
+        @keys = other.keys
+      end
+
+      def []=(key, value)
+        @keys << key if !has_key?(key)
+        super
+      end
+
+      def delete(key)
+        if has_key? key
+          index = @keys.index(key)
+          @keys.delete_at index
+        end
+        super
+      end
+
+      def delete_if
+        super
+        sync_keys!
+        self
+      end
+
+      def reject!
+        super
+        sync_keys!
+        self
+      end
+
+      def reject(&block)
+        dup.reject!(&block)
+      end
+
+      def keys
+        @keys.dup
+      end
+
+      def values
+        @keys.collect { |key| self[key] }
+      end
+
+      def to_hash
+        self
+      end
+
+      def to_a
+        @keys.map { |key| [ key, self[key] ] }
+      end
+
+      def each_key
+        @keys.each { |key| yield key }
+      end
+
+      def each_value
+        @keys.each { |key| yield self[key]}
+      end
+
+      def each
+        @keys.each {|key| yield [key, self[key]]}
+      end
+
+      alias_method :each_pair, :each
+
+      def clear
+        super
+        @keys.clear
+        self
+      end
+
+      def shift
+        k = @keys.first
+        v = delete(k)
+        [k, v]
+      end
+
+      def merge!(other_hash)
+        other_hash.each {|k,v| self[k] = v }
+        self
+      end
+
+      def merge(other_hash)
+        dup.merge!(other_hash)
+      end
+
+      # When replacing with another hash, the initial order of our keys must come from the other hash -ordered or not.
+      def replace(other)
+        super
+        @keys = other.keys
+        self
+      end
+      
+      def reverse
+        OrderedHashInt[self.to_a.reverse]
+      end
+
+    private
+
+      def sync_keys!
+        @keys.delete_if {|k| !has_key?(k)}
+      end
+    end
+
+  class OrderedHash < OrderedHashInt #:nodoc:
+    def initialize(*args, &block)
+      @timestamps = OrderedHashInt.new
+      super
+    end
+
+    def initialize_copy(other)
+      @timestamps = other.timestamps
+      super
+    end
+
+    def []=(key, value, timestamp = nil)
+      @timestamps[key] = timestamp
+      super(key, value)
+    end
+
+    def delete(key)
+      @timestamps.delete(key)
+      super
+    end
+
+    def delete_if(&block)
+      @timestamps.delete_if(&block)
+      super
+    end
+
+    def reject!(&block)
+      @timestamps.reject!(&block)
+      super
+    end
+
+    def timestamps
+      @timestamps.dup
+    end
+
+    def clear
+      @timestamps.clear
+      super
+    end
+
+    def shift
+      k, v = super
+      @timestamps.delete(k)
+      [k, v]
+    end
+
+    def replace(other)
+      @timestamps = other.timestamps
+      super
+    end
+
+    def inspect
+      "#<OrderedHash #{super}\n#{@timestamps.inspect}>"
+    end
+  end
+end

data/lib/cassandra/protocol.rb

@@ -0,0 +1,120 @@
+class Cassandra
+  # Inner methods for actually doing the Thrift calls
+  module Protocol #:nodoc:
+    private
+
+    def _mutate(mutation_map, consistency_level)
+      client.batch_mutate(mutation_map, consistency_level)
+    end
+
+    def _remove(key, column_path, timestamp, consistency_level)
+      client.remove(key, column_path, timestamp, consistency_level)
+    end
+
+    def _count_columns(column_family, key, super_column, start, stop, count, consistency)
+      client.get_count(key,
+        CassandraThrift::ColumnParent.new(:column_family => column_family, :super_column => super_column),
+        CassandraThrift::SlicePredicate.new(:slice_range =>
+                                            CassandraThrift::SliceRange.new(
+                                              :start  => start  || '',
+                                              :finish => stop   || '',
+                                              :count  => count  || 100
+                                            )),
+        consistency
+      )
+    end
+
+    # FIXME: Add support for start, stop, count
+    def _get_columns(column_family, key, columns, sub_columns, consistency)
+      result = if is_super(column_family)
+        if sub_columns
+          columns_to_hash(column_family, client.get_slice(key,
+            CassandraThrift::ColumnParent.new(:column_family => column_family, :super_column => columns),
+            CassandraThrift::SlicePredicate.new(:column_names => sub_columns),
+            consistency))
+        else
+          columns_to_hash(column_family, client.get_slice(key,
+            CassandraThrift::ColumnParent.new(:column_family => column_family),
+            CassandraThrift::SlicePredicate.new(:column_names => columns),
+            consistency))
+        end
+      else
+        columns_to_hash(column_family, client.get_slice(key,
+          CassandraThrift::ColumnParent.new(:column_family => column_family),
+          CassandraThrift::SlicePredicate.new(:column_names => columns),
+          consistency))
+      end
+
+      klass = column_name_class(column_family)
+      (sub_columns || columns).map { |name| result[klass.new(name)] }
+    end
+
+    def _multiget(column_family, keys, column, sub_column, count, start, finish, reversed, consistency)
+      # Single values; count and range parameters have no effect
+      if is_super(column_family) and sub_column
+        predicate = CassandraThrift::SlicePredicate.new(:column_names => [sub_column])
+        column_parent = CassandraThrift::ColumnParent.new(:column_family => column_family, :super_column => column)
+        column_hash = multi_sub_columns_to_hash!(column_family, client.multiget_slice(keys, column_parent, predicate, consistency))
+
+        klass = sub_column_name_class(column_family)
+        keys.inject({}){|hash, key| hash[key] = column_hash[key][klass.new(sub_column)]; hash}
+      elsif !is_super(column_family) and column
+        predicate = CassandraThrift::SlicePredicate.new(:column_names => [column])
+        column_parent = CassandraThrift::ColumnParent.new(:column_family => column_family)
+        column_hash  = multi_columns_to_hash!(column_family, client.multiget_slice(keys, column_parent, predicate, consistency))
+
+        klass = column_name_class(column_family)
+        keys.inject({}){|hash, key| hash[key] = column_hash[key][klass.new(column)]; hash}
+
+      # Slices
+      else
+        predicate = CassandraThrift::SlicePredicate.new(:slice_range =>
+          CassandraThrift::SliceRange.new(
+            :reversed => reversed,
+            :count => count,
+            :start => start,
+            :finish => finish))
+
+        if is_super(column_family) and column
+          column_parent = CassandraThrift::ColumnParent.new(:column_family => column_family, :super_column => column)
+          multi_sub_columns_to_hash!(column_family, client.multiget_slice(keys, column_parent, predicate, consistency))
+        else
+          column_parent = CassandraThrift::ColumnParent.new(:column_family => column_family)
+          multi_columns_to_hash!(column_family, client.multiget_slice(keys, column_parent, predicate, consistency))
+        end
+      end
+    end
+
+    def _get_range(column_family, start_key, finish_key, key_count, columns, start, finish, count, consistency, reversed=false)
+      column_parent = CassandraThrift::ColumnParent.new(:column_family => column_family)
+      predicate = if columns
+                    CassandraThrift::SlicePredicate.new(:column_names => columns)
+                  else
+                    CassandraThrift::SlicePredicate.new(:slice_range => 
+                      CassandraThrift::SliceRange.new(
+                        :start  => start, 
+                        :finish => finish,
+                        :count  => count,
+                        :reversed => reversed))
+                  end
+      range = CassandraThrift::KeyRange.new(:start_key => start_key, :end_key => finish_key, :count => key_count)
+      client.get_range_slices(column_parent, predicate, range, consistency)
+    end
+
+    # TODO: Supercolumn support
+    def _get_indexed_slices(column_family, index_clause, column, count, start, finish, reversed, consistency)
+      column_parent = CassandraThrift::ColumnParent.new(:column_family => column_family)
+      if column
+        predicate = CassandraThrift::SlicePredicate.new(:column_names => [column])
+      else
+        predicate = CassandraThrift::SlicePredicate.new(:slice_range =>
+          CassandraThrift::SliceRange.new(
+            :reversed => reversed,
+            :count => count,
+            :start => start,
+            :finish => finish))
+      end
+      client.get_indexed_slices(column_parent, index_clause, predicate, consistency)
+    end
+  end
+end

data/lib/cassandra/time.rb

@@ -0,0 +1,11 @@
+
+class Time
+  def self.stamp
+    Time.now.stamp
+  end
+  
+  def stamp
+    to_i * 1_000_000 + usec
+  end
+end
+        

data/test/cassandra_client_test.rb

@@ -0,0 +1,20 @@
+require File.expand_path(File.dirname(__FILE__) + '/test_helper')
+
+class CassandraClientTest < Test::Unit::TestCase
+  include Cassandra::Constants
+  
+  def setup
+    @twitter = Cassandra.new('Twitter', "127.0.0.1:9160", :retries => 2, :exception_classes => [])
+  end
+  
+  def test_client_method_is_called
+    assert_nil @twitter.instance_variable_get(:@client)
+    @twitter.insert(:Statuses, key, {'1' => 'v', '2' => 'v', '3' => 'v'})
+    assert_not_nil @twitter.instance_variable_get(:@client)
+  end
+  
+  def key
+    caller.first[/`(.*?)'/, 1]
+  end
+  
+end

data/test/cassandra_mock_test.rb

@@ -0,0 +1,122 @@
+require File.expand_path(File.dirname(__FILE__) + '/test_helper')
+require File.expand_path(File.dirname(__FILE__) + '/cassandra_test')
+require 'cassandra/mock'
+require 'json'
+
+class CassandraMockTest < CassandraTest
+  include Cassandra::Constants
+
+  def setup
+    @test_schema = JSON.parse(File.read(File.join(File.expand_path(File.dirname(__FILE__)), '..','conf', CASSANDRA_VERSION, 'schema.json')))
+    @twitter = Cassandra::Mock.new('Twitter', @test_schema)
+    @twitter.clear_keyspace!
+
+    @blogs = Cassandra::Mock.new('Multiblog', @test_schema)
+    @blogs.clear_keyspace!
+
+    @blogs_long = Cassandra::Mock.new('MultiblogLong', @test_schema)
+    @blogs_long.clear_keyspace!
+
+    @type_conversions = Cassandra::Mock.new('TypeConversions', @test_schema)
+    @type_conversions.clear_keyspace!
+
+    @uuids = (0..6).map {|i| SimpleUUID::UUID.new(Time.at(2**(24+i))) }
+    @longs = (0..6).map {|i| Long.new(Time.at(2**(24+i))) }
+    @composites = [
+      Cassandra::Composite.new([5].pack('N'), "zebra"),
+      Cassandra::Composite.new([5].pack('N'), "aardvark"),
+      Cassandra::Composite.new([1].pack('N'), "elephant"),
+      Cassandra::Composite.new([10].pack('N'), "kangaroo"),
+    ]
+  end
+
+  def test_setup
+    assert @twitter
+    assert @blogs
+    assert @blogs_long
+  end
+  
+  def test_schema_for_keyspace
+    data = @test_schema['Twitter']
+    stuff = @twitter.send(:schema_for_keyspace, 'Twitter')
+    data.keys.each do |k|
+      assert_equal data[k], stuff[k], k
+    end
+  end
+
+  def test_sorting_row_keys
+    @twitter.insert(:Statuses, 'b', {:text => 'foo'})
+    @twitter.insert(:Statuses, 'a', {:text => 'foo'})
+    assert_equal ['a'], @twitter.get_range(:Statuses, :key_count => 1).keys
+  end
+  
+  def test_get_range_reversed
+    data = 3.times.map { |i| ["body-#{i.to_s}", "v"] }
+    hash = Cassandra::OrderedHash[data]
+    reversed_hash = Cassandra::OrderedHash[data.reverse]
+    
+    @twitter.insert(:Statuses, "all-keys", hash)
+    
+    columns = @twitter.get_range(:Statuses, :reversed => true)["all-keys"]
+    columns.each do |column|
+      assert_equal reversed_hash.shift, column
+    end
+  end
+  
+  def test_get_range_reversed_slice
+    data = 4.times.map { |i| ["body-#{i.to_s}", "v"] }
+    hash = Cassandra::OrderedHash[data]
+    sliced_hash = Cassandra::OrderedHash[data.reverse[1..-1]]
+    
+    @twitter.insert(:Statuses, "all-keys", hash)
+    
+    columns = @twitter.get_range(
+      :Statuses,
+      :start => sliced_hash.keys.first,
+      :reversed => true
+    )["all-keys"]
+    
+    columns.each do |column|
+      assert_equal sliced_hash.shift, column
+    end
+  end
+  
+  def test_get_range_count
+    data = 3.times.map { |i| ["body-#{i.to_s}", "v"] }
+    hash = Cassandra::OrderedHash[data]
+    
+    @twitter.insert(:Statuses, "all-keys", hash)
+    
+    columns = @twitter.get_range(:Statuses, :count => 2)["all-keys"]
+    assert_equal 2, columns.count
+  end
+
+  def test_inserting_array_for_indices
+    @twitter.insert(:TimelinishThings, 'a', ['1','2'])
+    row = @twitter.get(:TimelinishThings, 'a')
+    assert_equal({'1' => nil, '2' => nil}, row)
+
+    assert_raises(ArgumentError) {
+      @twitter.insert(:UserRelationships, 'a', ['u1','u2'])
+    }
+  end
+  
+  def test_column_timestamps
+    base_time = Time.now
+    @twitter.insert(:Statuses, "time-key", { "body" => "value" })
+
+    results = @twitter.get(:Statuses, "time-key")
+    assert(results.timestamps["body"] / 1000000 >= base_time.to_i)
+  end
+  
+  def test_supercolumn_timestamps
+    base_time = Time.now
+    @twitter.insert(:StatusRelationships, "time-key", { "super" => { @uuids[1] => "value" }})
+
+    results = @twitter.get(:StatusRelationships, "time-key")
+    assert_nil(results.timestamps["super"])
+    
+    columns = results["super"]
+    assert(columns.timestamps[@uuids[1]] / 1000000 >= base_time.to_i)
+  end
+end