adapted examples

examples/benchmark.py

@@ -31,7 +31,7 @@
 n_jobs = 8
 
 # parameters
-ofilename = "results.json"
+ofilename = "benchmark.json"
 n_test_range = [1000000, 2500000, 5000000, 7500000, 10000000]
 algorithms = ["brute", "kd_tree", "buffer_kd_tree"]
 
@@ -103,10 +103,6 @@ def run_algorithm(n_test, tree_depth=None, algorithm="buffer_kd_tree"):
     end_time = time.time()
     test_time = (end_time - start_time)
     print("Testing time: %f" % test_time)
-
-    # store results 
-    if algorithm not in results.keys():
-        results[algorithm] = {}
 
     return train_time, test_time
 
@@ -116,6 +112,8 @@ def run_algorithm(n_test, tree_depth=None, algorithm="buffer_kd_tree"):
 for i in xrange(len(algorithms)):
 
     algorithm = algorithms[i]
+    results[algorithm] = {}
+
     print("----------------------------------------------------------------------")
     print("\n\nRunning %s ...\n" % (algorithm))
     print("----------------------------------------------------------------------")

examples/skcompare.py

@@ -12,15 +12,17 @@
 """
 print(__doc__)
 
+import json
 import time
 import generate
 from bufferkdtree.neighbors import NearestNeighbors
 from sklearn.neighbors import NearestNeighbors as NearestNeighborsSKLEARN
 
 # parameters
-n_jobs = 2
+n_jobs = 4
 leaf_size = 30
-n_test = 100000
+ofilename = "skcomparison.json"
+n_test_range = [100000, 250000, 500000, 750000, 1000000]
 algorithms = ["kd_tree_sklearn", "kd_tree"]
 
 verbose = 0
@@ -34,6 +36,8 @@
 print("Dimensionality of patterns:\t%i" % Xtrain.shape[1])
 print("----------------------------------------------------------------------")
 
+results = {}
+
 def run_algorithm(n_test_local, leaf_size=30, algorithm="kd_tree"):
 
     print("----------------------------------------------------------------------")
@@ -75,8 +79,19 @@ def run_algorithm(n_test_local, leaf_size=30, algorithm="kd_tree"):
 for i in xrange(len(algorithms)):
 
     algorithm = algorithms[i]
+    results[algorithm] = {}
     print("----------------------------------------------------------------------")
     print("\n\nRunning %s ...\n" % (algorithm))
     print("----------------------------------------------------------------------")
-    run_algorithm(n_test, algorithm=algorithm, leaf_size=leaf_size)
-
+    for n_test in n_test_range:
+        train_time, test_time = run_algorithm(n_test, 
+                                              algorithm=algorithm, 
+                                              leaf_size=leaf_size)
+        results[algorithm][n_test] = {'train':train_time,
+                                      'test':test_time,
+                                      }
+
+# write results after each step
+print("Writing results to %s ..." % ofilename)
+with open(ofilename, 'w') as f:
+    json.dump(results, f)