dataoperations-aggregate 0.0.6 → 0.0.8

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: e2cf76bae5e748e46672c8d59f8859358452f2fca813499b148e85dfa61f38c9
-  data.tar.gz: 7a9fbbd0115edefa597b9d374407576f4dfedb133da781dffe066c3e4ba6f73a
+  metadata.gz: 9e69958cbf826f7524a31c8621afd7c554c61ceeead22b62460c464d0d584e79
+  data.tar.gz: c8451ad56f5cae4c5b4b4ff26d02bf9445f3fc925c02a2cf843a551f3006620a
 SHA512:
-  metadata.gz: 403435362bd15744b154604e6322dcbba89ea9952e36d82ffe3ac3299f3ad28cd5c29187843e780cf149dabf2abdb4fbb6700f1e8ddce8e792ab033adf174255
-  data.tar.gz: f883f091f074ebbe48e18c46790ceb37c5ba98d7a879ec8259f69c88a19225130864897ab8bb63e47c7ab31dbcc5a400ae6c7fe68f3b8a27abec476c6e34123d
+  metadata.gz: a85e773496690078a0ad56971ea0a17b15c0656abc2f9d28eefbfbeec86b867f378b4bc3175eb84462d75b4035a5249e356fc969f80302ed1b691f00be569e20
+  data.tar.gz: 0f1dd3bccb61205b3decbce1a2321ba5b2326978d9584e200cd3077b5f9eebe077015a8460f2261fc4583ec9850f3178f79171dc1f90a0ca622e47961dc2edb4

data/lib/dataoperations-aggregate.rb

@@ -12,7 +12,10 @@ module DataOperations
     DEFAULT_TIME_STARTED_MODE = :first_message
     DEFAULT_FIELD_NO_DATA_VALUE = 'no_data'.freeze
     DEFAULT_AGGREGATIONS = %w[sum min max mean median variance standard_deviation].freeze
-    VALID_AGGREGATIONS = %w[sum min max mean median variance standard_deviation].freeze
+    VALID_AGGREGATIONS = %w[sum min max mean median variance standard_deviation histogram].freeze
+    DEFAULT_HISTOGRAM_CUMULATIVE = false
+    DEFAULT_HISTOGRAM_BUCKET_INFINITE_ENABLED = true
+    DEFAULT_HISTOGRAM_BUCKET_COMPARATION = :less_or_equal
     DEFAULT_HASH_TIME_FORMAT = '%Y-%m-%dT%H'.freeze
     DEFAULT_INERVAL_SECONDS = 3600
 
@@ -30,7 +33,12 @@ module DataOperations
                    log: Logger.new(STDOUT),
                    aggregation_names:,
                    group_field_names:,
-                   aggregate_field_names:
+                   aggregate_field_names:,
+                   histogram_cumulative: DEFAULT_HISTOGRAM_CUMULATIVE,
+                   histogram_bucket_infinite_enabled: DEFAULT_HISTOGRAM_BUCKET_INFINITE_ENABLED,
+                   histogram_bucket_comparation: DEFAULT_HISTOGRAM_BUCKET_COMPARATION,
+                   histogram_buckets:[],
+                   histogram_fields:[]
                 )
       @aggregator = aggregator
       @time_format = time_format
@@ -43,6 +51,11 @@ module DataOperations
       @processing_mode = processing_mode
       @time_started_mode = time_started_mode
       @aggregator_name = aggregator_name
+      @histogram_cumulative = histogram_cumulative
+      @histogram_bucket_infinite_enabled = histogram_bucket_infinite_enabled
+      @histogram_bucket_comparation = histogram_bucket_comparation
+      @histogram_buckets = histogram_buckets
+      @histogram_fields = histogram_fields
 
 
       if aggregation_names.nil? || !aggregation_names.is_a?(Array)
@@ -219,16 +232,52 @@ module DataOperations
       # Aggregate data
       if aggregate_field_value.is_a?(Array)
         @aggregation_names.each do |operation|
-          data = aggregate_field_value.method(operation).call
-          aggregator_data["#{aggregate_field_key}_#{operation}"] = data
-
-          # Add aggregated data to interval
-          group_item_value['intervals'].keys[1..-1].each do |interval_secs|
-            interval_aggregator_item_key = (aggregator_item_key / interval_secs.to_i) * interval_secs.to_i
-            interval_aggregator_item_value = group_item_value['intervals'][interval_secs][interval_aggregator_item_key]
-            interval_aggregator_item_value['aggregate_fields'][aggregate_field_key][operation] << data
+          
+          #If histogram, calculate bucket for metric
+          if operation == 'histogram'
+            if @histogram_bucket_comparation == :less_or_equal
+              bucket_comparation = "le"
+            else
+              bucket_comparation = "ge"
+            end
+            #If set buckets and set metrics to calculate (bucket values depends of ranges based in metric activity)
+            if !@histogram_buckets.nil? && !@histogram_fields.nil? && @histogram_fields.include?(aggregate_field_key)
+              data_bucket = calculate_histogram_buckets(aggregate_field_value, @histogram_buckets)
+             
+              data_bucket.each {|bucket,bucket_count|
+                #@log.info("#{aggregate_field_key}_#{bucket} = #{bucket_count}")
+                aggregator_data["#{aggregate_field_key}_bucket_#{bucket_comparation}_#{bucket}"] = bucket_count
+              }
+
+              # Add aggregated data to interval
+              group_item_value['intervals'].keys[1..-1].each do |interval_secs|
+
+                interval_aggregator_item_key = (aggregator_item_key / interval_secs.to_i) * interval_secs.to_i
+                interval_aggregator_item_value = group_item_value['intervals'][interval_secs][interval_aggregator_item_key]
+                data_bucket.each {|bucket,bucket_count|
+                  #@log.info("#{aggregate_field_key}_#{bucket} = #{bucket_count}")
+                  interval_aggregator_item_value['aggregate_fields'][aggregate_field_key]["bucket_#{bucket_comparation}_#{bucket}"] = [] if interval_aggregator_item_value['aggregate_fields'][aggregate_field_key]["bucket#{bucket}"].nil?
+                  interval_aggregator_item_value['aggregate_fields'][aggregate_field_key]["bucket_#{bucket_comparation}_#{bucket}"] << bucket_count
+                }
+                  
+              end
+
+            end
+          else
+            data = aggregate_field_value.method(operation).call
+            aggregator_data["#{aggregate_field_key}_#{operation}"] = data
+
+            # Add aggregated data to interval
+            group_item_value['intervals'].keys[1..-1].each do |interval_secs|
+              interval_aggregator_item_key = (aggregator_item_key / interval_secs.to_i) * interval_secs.to_i
+              interval_aggregator_item_value = group_item_value['intervals'][interval_secs][interval_aggregator_item_key]
+              interval_aggregator_item_value['aggregate_fields'][aggregate_field_key][operation] << data
+            end
           end
+
+
         end
+
       end
     end
 
@@ -244,6 +293,7 @@ module DataOperations
         @aggregation_names.each do |operation|
           interval_aggregator_item_value['aggregate_fields'][aggregate_field_key][operation] = []
         end
+
       end
     end
 
@@ -297,10 +347,17 @@ module DataOperations
             case operation
             when 'max', 'min', 'mean', 'median'
               data = vector.method(operation).call
+            when 'histogram'
+              #Bucket operation generate bucket[\d]+
+              data = nil
+            when /^bucket.+/
+              #For buckets sum accumulations for internvals
+              data = vector.method('sum').call
             else
               data = vector.median
             end
-            aggregator_data["#{field_name}_#{operation}"] = data
+            #Nil data is avoid (for example for 'bucket' name operation)
+            aggregator_data["#{field_name}_#{operation}"] = data unless data.nil?
           end
         end
         # @log.debug aggregator_item_value
@@ -310,5 +367,52 @@ module DataOperations
         aggregate_data[s_interval] << aggregator_data
       end
     end
+    
+    #Return Array with count by each bucket
+    def calculate_histogram_buckets(data, buckets_config)
+
+      if @histogram_bucket_comparation == :less_or_equal
+        buckets_config.sort!.uniq!
+      else
+        buckets_config.reverse!.uniq!
+      end
+      
+      buckets = {}
+    
+      buckets_config.each {|bucket| buckets[bucket] = 0}
+    
+      #By default add Inf bucket
+      buckets["Inf"] = 0 if @histogram_bucket_infinite_enabled
+
+      data.each do|item|
+        item_in_bucket = false
+        buckets_config.each do|bucket|
+          if @histogram_bucket_comparation == :less_or_equal
+            if item <= bucket
+              item_in_bucket = true
+            end
+          elsif item >= bucket
+              item_in_bucket = true
+          end
+
+          if item_in_bucket
+            buckets[bucket] += 1
+            #Not accumule if histogram cumulative is false
+            break if !@histogram_cumulative
+          end
+        end
+
+        #Only count Inf bucket if enabled
+        if @histogram_bucket_infinite_enabled
+          if !item_in_bucket
+            buckets["Inf"] += 1
+          elsif @histogram_cumulative
+            buckets["Inf"] += 1
+          end
+        end
+      end
+      return buckets
+    end
+
   end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: dataoperations-aggregate
 version: !ruby/object:Gem::Version
-  version: 0.0.6
+  version: 0.0.8
 platform: ruby
 authors:
 - Victor Guillen
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2020-05-19 00:00:00.000000000 Z
+date: 2023-04-09 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: descriptive_statistics
@@ -50,7 +50,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.0.3
+rubygems_version: 3.0.3.1
 signing_key: 
 specification_version: 4
 summary: Aggregate data