Experiment cleanup

run-cycle-experiments-detailed-sample.jl → Experiments/Scripts/run-cycle-experiments-without-join-table-cycle-stats.jl

@@ -1,9 +1,7 @@
 using Plots.PlotMeasures
-include("Experiments/Experiments.jl")
+include("../Experiments.jl")
 
-datasets::Vector{DATASET} = [aids]
-# datasets::Vector{DATASET} = [aids, human, yeast, wordnet, youtube, dblp, patents]
-# datasets::Vector{DATASET} = [aids, human, lubm80, yeast, hprd, wordnet, dblp, youtube, eu2005, patents]
+datasets::Vector{DATASET} = [aids, human, lubm80, yeast, hprd, wordnet, dblp, youtube, eu2005, patents]
 max_cycles = 6
 
 experiment_params_list::Vector{ExperimentParams} = [ExperimentParams(dataset=current_dataset, partitioner=QuasiStable, max_cycle_size=current_size) for current_dataset in datasets for current_size in 2:max_cycles]
@@ -16,10 +14,7 @@ for experiment_params in experiment_params_list
     summary_name = params_to_summary_filename(experiment_params)
     summary_file_location = "Experiments/SerializedSummaries/" * summary_name
     println("Building Color Summary: ", summary_name)
-    # normal_results = @timed generate_color_summary(data, summary_params; verbose=1)
-    # normal_results = @timed generate_color_summary(data, summary_params; verbose=0, detailed_cycles=false)
-    results= @timed generate_color_summary(data, summary_params; verbose=1, detailed_cycles=false)
-    # println("normal time: ", normal_results.time)
+    results= @timed generate_color_summary(data, summary_params; verbose=1, use_cycle_join_table=false)
     println("detailed time: ", results.time)
     summary_size = Base.summarysize(results.value)
     serialize(summary_file_location, results.value)
@@ -35,4 +30,4 @@ end
 println("started estimating")
 run_estimation_experiments(experiment_params_list)
 println("started graphing")
-graph_grouped_box_plot(experiment_params_list, x_type=dataset, y_type=estimate_error, grouping=cycle_size, filename="detailed-sample-experiment")
+graph_grouped_box_plot(experiment_params_list, x_type=dataset, y_type=estimate_error, grouping=cycle_size, filename="cycles-without-join-table-cycle-stats-experiment")

run-cycle-experiments.jl → Experiments/Scripts/run-cycle-experiments.jl

@@ -1,11 +1,10 @@
 using Plots.PlotMeasures
-include("Experiments/Experiments.jl")
+include("../Experiments.jl")
 
-datasets::Vector{DATASET} = [aids, human, yeast, wordnet, youtube, dblp, patents]
-# datasets::Vector{DATASET} = [aids, human, lubm80, yeast, hprd, wordnet, dblp, youtube, eu2005, patents]
+datasets::Vector{DATASET} = [aids, human, lubm80, yeast, hprd, wordnet, dblp, youtube, eu2005, patents]
 max_cycles = 6
-
 experiment_params_list::Vector{ExperimentParams} = [ExperimentParams(dataset=current_dataset, partitioner=QuasiStable, max_cycle_size=current_size) for current_dataset in datasets for current_size in 2:max_cycles]
+
 println("started building")
 build_experiments(experiment_params_list)
 println("started estimating")

Experiments/Scripts/run-inference-sampling-experiments.jl

@@ -0,0 +1,13 @@
+using Plots.PlotMeasures
+include("../Experiments.jl")
+
+datasets::Vector{DATASET} = [aids, human, yeast, wordnet, youtube, dblp, patents]
+max_paths = 60
+experiment_params_list::Vector{ExperimentParams} = [ExperimentParams(dataset=current_dataset, partitioner=QuasiStable, inference_max_paths=current_paths) for current_dataset in datasets for current_paths in 2:10:max_paths]
+
+println("started building")
+build_experiments(experiment_params_list)
+println("started estimating")
+run_estimation_experiments(experiment_params_list)
+println("started graphing")
+graph_grouped_box_plot(experiment_params_list, x_type=dataset, y_type=estimate_error, grouping=inference_paths, filename="inferencesampling")

Experiments/Scripts/run-summary-sampling-experiments.jl

@@ -0,0 +1,14 @@
+using Plots.PlotMeasures
+include("../Experiments.jl")
+
+
+datasets::Vector{DATASET} = [aids, wordnet, lubm80, human]
+max_paths = 1000
+experiment_params_list::Vector{ExperimentParams} = [ExperimentParams(dataset=current_dataset, partitioner=QuasiStable, summary_max_paths=current_paths) for current_dataset in datasets for current_paths in 0:200:max_paths]
+
+println("started building")
+build_experiments(experiment_params_list)
+println("started estimating")
+run_estimation_experiments(experiment_params_list)
+println("started graphing")
+graph_grouped_box_plot(experiment_params_list, x_type=dataset, y_type=error, grouping=summary_paths, filename="summarysamples")

Experiments/Scripts/run-summary-sampling-with-query-type.jl

@@ -0,0 +1,13 @@
+using Plots.PlotMeasures
+include("../Experiments.jl")
+
+current_dataset = yeast
+max_paths = 300
+
+experiment_params_list::Vector{ExperimentParams} = [ExperimentParams(dataset=current_dataset, partitioner=QuasiStable, summary_max_paths=current_paths) for current_paths in 10:30:max_paths]
+println("started building")
+build_experiments(experiment_params_list)
+println("started estimating")
+run_estimation_experiments(experiment_params_list)
+println("started graphing")
+graph_grouped_box_plot(experiment_params_list, x_type=query_type, y_type=estimate_error, grouping=summary_paths, filename="summarysamplesquerytypesyeast")

Experiments/Scripts/update_experiments.jl

@@ -0,0 +1,21 @@
+using Plots.PlotMeasures
+using Graphs
+include("../Experiments.jl")
+
+datasets::Vector{DATASET} = [wordnet]
+# datasets::Vector{DATASET} = [aids, human, yeast, wordnet, youtube, dblp, patents]
+# datasets::Vector{DATASET} = [aids, human, lubm80, yeast, hprd, wordnet, dblp, youtube, eu2005, patents]
+max_cycles = 6
+proportions_not_updated = [0, 0.2, 0.4, 0.6, 0.8, 1]
+
+experiment_params_list::Vector{ExperimentParams} = [ExperimentParams(dataset=current_dataset, partitioner=QuasiStable, max_cycle_size=current_cycle, proportion_not_updated=current_proportion) 
+                                                    for current_dataset in datasets for current_cycle in 2:max_cycles for current_proportion in proportions_not_updated]
+println("started building")
+build_experiments(experiment_params_list)
+println("started estimating")
+run_estimation_experiments(experiment_params_list)
+println("started graphing")
+# compare how overall accuracy is affected by summary updates
+# graph_grouped_box_plot(experiment_params_list, x_type=dataset, y_type=estimate_error, grouping=proportion_not_updated, filename="overall-accuracy-and-updates")
+# compare how cycle stat accuracies are affected by summary updates
+graph_grouped_box_plot(experiment_params_list, x_type=proportion_not_updated, y_type=estimate_error, grouping=cycle_size, filename="cycle-stats-and-updates")

Experiments/build_color_summaries.jl

@@ -4,13 +4,34 @@ function build_experiments(experiment_params_list::Vector{ExperimentParams})
         dataset = experiment_params.dataset
         summary_params = experiment_params.summary_params
         data = load_dataset(dataset)
+        cloned_data = DataGraph(nv(data.graph))
+        remaining_edges = []
+        if (experiment_params.summary_params.proportion_not_updated < 1.0)
+            cloned_data.vertex_labels = data.vertex_labels
+            graph_edges = collect(edges(data.graph))
+            # edges_to_add = (length(graph_edges) * experiment_params.summary_params.proportion_not_updated)
+            for edge in graph_edges
+                if (rand() < experiment_params.summary_params.proportion_not_updated)
+                    add_labeled_edge!(cloned_data, (src(edge), dst(edge)), only(data.edge_labels[(src(edge), dst(edge))]))
+                    # edges_to_add -= 1
+                else
+                    push!(remaining_edges, edge)
+                end
+            end
+        end
         summary_name = params_to_summary_filename(experiment_params)
         summary_file_location = "Experiments/SerializedSummaries/" * summary_name
         println("Building Color Summary: ", summary_name)
         timing_vec = Float64[]
-        results = @timed generate_color_summary(data, summary_params; verbose=1, timing_vec=timing_vec)
-        summary_size = Base.summarysize(results.value)
-        serialize(summary_file_location, results.value)
+        results = @timed generate_color_summary((experiment_params.summary_params.proportion_not_updated < 1.0) ? cloned_data : data, summary_params; verbose=1, timing_vec=timing_vec)
+        current_summary = results.value
+        if (experiment_params.summary_params.proportion_not_updated < 1.0)
+            for edge in remaining_edges
+                add_summary_edge!(current_summary, src(edge), dst(edge), get(data.edge_labels, (src(edge), dst(edge)), []))
+            end
+        end
+        summary_size = Base.summarysize(current_summary)
+        serialize(summary_file_location, current_summary)
         push!(build_times, (string(dataset),
                              string(summary_params.partitioner),
                              string(summary_params.num_colors),

Experiments/graph_results.jl

@@ -1,5 +1,4 @@
-@enum GROUP dataset technique cycle_size summary_paths inference_paths query_type sampling_type cycle_stats number_of_colors build_phase
-#todo: query type
+@enum GROUP dataset technique cycle_size summary_paths inference_paths query_type sampling_type cycle_stats number_of_colors build_phase proportion_not_updated
 
 @enum VALUE estimate_error runtime build_time memory_footprint
 
@@ -244,6 +243,8 @@ function get_value_from_param(experiment_param::ExperimentParams, value_type::GR
         return experiment_param.only_shortest_path_cycle
     elseif value_type == number_of_colors
         return experiment_param.summary_params.num_colors
+    elseif value_type == proportion_not_updated
+        return experiment_param.summary_params.proportion_not_updated
     else
         # default to grouping by technique
         return (experiment_param.summary_params.partitioner, experiment_param.summary_params.label_refining_rounds)

Experiments/utils.jl

@@ -16,13 +16,14 @@ struct ExperimentParams
     function ExperimentParams(;dataset::DATASET,  num_colors::Int=64, max_cycle_size=6,
         only_shortest_path_cycle=false, summary_max_paths=1000,
         partitioner::PARTITIONER = QuasiStable, weighting=true, inference_max_paths=500, use_partial_sums=true,
-        sampling_strategy=redistributive, label_refining_rounds = 0)
+        sampling_strategy=redistributive, label_refining_rounds = 0, proportion_not_updated=1.0)
         return new(dataset, ColorSummaryParams(num_colors=num_colors,
                                                        max_cycle_size=max_cycle_size,
                                                        max_partial_paths=summary_max_paths,
                                                        partitioner=partitioner,
                                                        weighting=weighting,
-                                                       label_refining_rounds=label_refining_rounds),
+                                                       label_refining_rounds=label_refining_rounds,
+                                                       proportion_not_updated=proportion_not_updated),
                     inference_max_paths,
                     only_shortest_path_cycle,
                     use_partial_sums,

Source/CardinalityWithColors.jl

@@ -33,10 +33,11 @@ struct ColorSummaryParams
     partitioner::PARTITIONER
     weighting::Bool
     label_refining_rounds::Int
+    proportion_not_updated::Float16
 
     function ColorSummaryParams(;num_colors::Int=64, max_cycle_size=4, max_partial_paths=1000,
-            partitioner::PARTITIONER = QuasiStable, weighting=true, label_refining_rounds = 0)
-        return new(num_colors, max_cycle_size, max_partial_paths, partitioner, weighting, label_refining_rounds)
+            partitioner::PARTITIONER = QuasiStable, weighting=true, label_refining_rounds = 0, proportion_not_updated = 1.0)
+        return new(num_colors, max_cycle_size, max_partial_paths, partitioner, weighting, label_refining_rounds, proportion_not_updated)
     end
 end
 
@@ -45,7 +46,8 @@ function params_to_string(params::ColorSummaryParams)
     summary_name *= string(params.num_colors) * "_"
     summary_name *= string(params.max_cycle_size) * "_"
     summary_name *= string(params.max_partial_paths)* "_"
-    summary_name *= string(params.label_refining_rounds)
+    summary_name *= string(params.label_refining_rounds)* "_"
+    summary_name *= string(params.proportion_not_updated)
     return summary_name
 end