Move exports/exports.csv to mskk_benchmark after each series of runs

benchmarks-files/algos.yaml

@@ -5,12 +5,12 @@ float:
       docker-tag: ann-benchmarks-ecp
       module: ann_benchmarks.algorithms.eCP
       constructor: eCP
-      base-args: ["@metric", 50] #percentage of dataset to load dynamically load
+      base-args: ["@metric", 0] #percentage of dataset to load dynamically load
       run-groups:
        eCP:
           # args: [[30, 300, 800, 1200], #sc
-          args: [[100], #sc
-                 [0.3], #span (0.0 < p < 1.0)
+          args: [[130], #sc
+                 [0.0], #span (0.0 < p < 1.0)
                  [1],   #cpol (1=avg, 2=abs)
                  [1]]   #npol
           query-args: [[1, 5, 10, 20]] #b  
@@ -19,12 +19,12 @@ float:
       docker-tag: ann-benchmarks-ecp
       module: ann_benchmarks.algorithms.eCP-2
       constructor: eCP
-      base-args: ["@metric", 50] #percentage of dataset to load dynamically load
+      base-args: ["@metric", 0] #percentage of dataset to load dynamically load
       run-groups:
        eCP:
           # args: [[30, 300, 800, 1200], #sc
-          args: [[100], #sc
-                 [0.3], #span (0.0 < p < 1.0)
+          args: [[130], #sc
+                 [0.0], #span (0.0 < p < 1.0)
                  [1],   #cpol (1=avg, 2=abs)
                  [1]]   #npol
           query-args: [[1, 5, 10, 20]] #b  
@@ -33,12 +33,12 @@ float:
       docker-tag: ann-benchmarks-ecp
       module: ann_benchmarks.algorithms.eCP-3
       constructor: eCP
-      base-args: ["@metric", 50] #percentage of dataset to load dynamically load
+      base-args: ["@metric", 0] #percentage of dataset to load dynamically load
       run-groups:
        eCP:
           # args: [[30, 300, 800, 1200], #sc
-          args: [[100], #sc
-                 [0.3], #span (0.0 < p < 1.0)
+          args: [[130], #sc
+                 [0.0], #span (0.0 < p < 1.0)
                  [1],   #cpol (1=avg, 2=abs)
                  [1]]   #npol
           query-args: [[1, 5, 10, 20]] #b  

scripts/ecp_install.sh

@@ -48,7 +48,7 @@ cd ann-benchmarks
 python3.6 -m venv env
 source env/bin/activate
 
-echo "Making export dir and export file..."
+echo "${NAME}: Making export dir and export file..."
 mkdir exports
 
 echo "${NAME}: Install dependencies"

scripts/ecp_run_local.sh

@@ -6,12 +6,10 @@
 NAME="ecp_run_local"  # Script name.
 
 # ALGO_NAME="eCP" # WARNING: Must match in algos.yaml + eCP.py.
-
-DATASET="random-xs-20-euclidean"
-
 ALGOS=(eCP eCP-2 eCP-3)
 
-#DATASET="glove-25-angular"
+# DATASET="random-xs-20-euclidean"
+DATASET="glove-25-angular"
 # DATASET="sift-128-euclidean"
 
 echo "${NAME}: Will run benchmarks on ${ALGOS[*]}. First run 'ecp_install.sh'. Dataset: $DATASET."
@@ -41,13 +39,13 @@ done
 echo "${NAME}: Creating plots and website"
 python plot.py --dataset ${DATASET}
 
-echo "Exporting data to ./exports"
+echo "${NAME}: Exporting data to ./exports"
 python data_export.py --output exports/exports.csv
 
-echo "Cleaning up data"
+echo "${NAME}: Averaging AB data samples"
 python cleanup.py
 
-echo "Moving samples into separate directory"
+echo "${NAME}: Moving samples into separate directory"
 mkdir mskk_benchmark
 mv -v ecp_clusters_bulk.csv   mskk_benchmark/ecp_clusters_bulk.csv
 mv -v ecp_clusters_incr.csv   mskk_benchmark/ecp_clusters_incr.csv 
@@ -56,6 +54,7 @@ mv -v ecp_general_stats.csv   mskk_benchmark/ecp_general_stats.csv
 mv -v ecp_maintenance.csv     mskk_benchmark/ecp_maintenance.csv 
 mv -v ecp_nodes_bulk.csv      mskk_benchmark/ecp_nodes_bulk.csv 
 mv -v ecp_nodes_incr.csv      mskk_benchmark/ecp_nodes_incr.csv
+mv -v exports/exports.csv     mskk_benchmark/exports.csv
 
 # mkdir -p website
 # python create_website.py --outputdir website --scatter --latex