removed unnecessary experiments and cleaned up existing ones

uwdb · Nov 13, 2023 · f9742e5 · f9742e5
1 parent b125408
commit f9742e5
Show file tree

Hide file tree

Showing 13 changed files with 69 additions and 276 deletions.
diff --git a/Experiments/Scripts/run-cycle-experiments-timed.jl b/Experiments/Scripts/run-cycle-experiments-timed.jl
diff --git a/.../run-cycle-experiments-detailed-sample.jl → ...riments-without-join-table-cycle-stats.jl b/.../run-cycle-experiments-detailed-sample.jl → ...riments-without-join-table-cycle-stats.jl
@@ -1,9 +1,7 @@
 using Plots.PlotMeasures
 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")
diff --git a/Experiments/Scripts/run-cycle-experiments.jl b/Experiments/Scripts/run-cycle-experiments.jl
@@ -1,12 +1,10 @@
 using Plots.PlotMeasures
 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]
+
 println("started building")
 build_experiments(experiment_params_list)
 println("started estimating")

diff --git a/Experiments/Scripts/run-inference-sampling-experiments.jl b/Experiments/Scripts/run-inference-sampling-experiments.jl
@@ -1,30 +1,10 @@
-# TODO:
-# - turn sampling up/down (build and inference)
-# - cycle length effects
-# - reverify ground truth wrt G-captures
-# - rerun initial g-care benchmarks and verify old results
-# - recreate initial G-Care Benchmark results
-
-# My tasks:
-# - turn sampling up/down
-# - cycle length effects
-
 using Plots.PlotMeasures
 include("../Experiments.jl")
 
-# datasets::Vector{DATASET} = [aids, wordnet, lubm80, human]
-# max_partial_paths = 10000
-
-# experiment_params_list::Vector{ExperimentParams} = [ExperimentParams(dataset=aids, partitioner=QuasiStable, inference_max_paths=2),
-#                                                     ExperimentParams(dataset=aids, partitioner=QuasiStable, inference_max_paths=10),
-#                                                     ExperimentParams(dataset=aids, partitioner=QuasiStable, inference_max_paths=50),
-#                                                     ExperimentParams(dataset=aids, partitioner=QuasiStable, inference_max_paths=250),
-#                                                     ExperimentParams(dataset=aids, partitioner=QuasiStable, inference_max_paths=1250)]
-
 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")

diff --git a/Experiments/Scripts/run-summary-sampling-experiments-full-range.jl b/Experiments/Scripts/run-summary-sampling-experiments-full-range.jl
diff --git a/Experiments/Scripts/run-summary-sampling-experiments.jl b/Experiments/Scripts/run-summary-sampling-experiments.jl
@@ -1,30 +1,14 @@
 using Plots.PlotMeasures
-include("../build_color_summaries.jl")
-include("../get_true_cardinalities.jl")
-include("../load_datasets.jl")
-include("../load_querysets.jl")
-include("../run_estimators.jl")
-include("../graph_results.jl")
-include("../utils.jl")
+include("../Experiments.jl")
 
-# datasets::Vector{DATASET} = [aids, wordnet, lubm80, human]
-# max_partial_paths = 10000
-
-# experiment_params_list::Vector{ExperimentParams} = [ExperimentParams(dataset=aids, partitioner=QuasiStable, inference_max_paths=2),
-#                                                     ExperimentParams(dataset=aids, partitioner=QuasiStable, inference_max_paths=10),
-#                                                     ExperimentParams(dataset=aids, partitioner=QuasiStable, inference_max_paths=50),
-#                                                     ExperimentParams(dataset=aids, partitioner=QuasiStable, inference_max_paths=250),
-#                                                     ExperimentParams(dataset=aids, partitioner=QuasiStable, inference_max_paths=1250)]
-
-# datasets::Vector{DATASET} = [aids, human, lubm80, yeast, hprd, wordnet, dblp, youtube, eu2005, patents]
-datasets::Vector{DATASET} = [aids]
 
+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]
 
-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 250:250: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")
diff --git a/Experiments/Scripts/run-summary-sampling-range-query-type.jl b/Experiments/Scripts/run-summary-sampling-range-query-type.jl
diff --git a/Experiments/Scripts/run-summary-sampling-with-query-type.jl b/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")