def test_KMeans_covtype_cols_fvec(self):
h2o.beta_features = True
# just do the import folder once
# make the timeout variable per dataset. it can be 10 secs for covtype 20x (col key creation)
# so probably 10x that for covtype200
csvFilenameList = [
("covtype.binary.svm", "cC", 30, 1),
# normal csv
]
### csvFilenameList = random.sample(csvFilenameAll,1)
# h2b.browseTheCloud()
lenNodes = len(h2o.nodes)
firstDone = False
importFolderPath = "libsvm"
for (csvFilename, hex_key, timeoutSecs, resultMult) in csvFilenameList:
# have to import each time, because h2o deletes source after parse
csvPathname = importFolderPath + "/" + csvFilename
# PARSE******************************************
# creates csvFilename.hex from file in importFolder dir
parseResult = h2i.import_parse(bucket='home-0xdiag-datasets', path=csvPathname,
hex_key=hex_key, timeoutSecs=2000)
print "Parse result['destination_key']:", parseResult['destination_key']
# INSPECT******************************************
start = time.time()
inspect = h2o_cmd.runInspect(None, parseResult['destination_key'], timeoutSecs=360)
print "Inspect:", parseResult['destination_key'], "took", time.time() - start, "seconds"
h2o_cmd.infoFromInspect(inspect, csvFilename)
numRows = inspect['numRows']
numCols = inspect['numCols']
# KMEANS******************************************
for trial in range(1):
kwargs = {
'k': 3,
'initialization': 'Furthest',
'ignored_cols': range(11, numCols),
'max_iter': 10,
# 'normalize': 0,
# reuse the same seed, to get deterministic results (otherwise sometimes fails
'seed': 265211114317615310,
}
# fails if I put this in kwargs..i.e. source = dest
# 'destination_key': parseResult['destination_key'],
for trial2 in range(3):
timeoutSecs = 600
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult, timeoutSecs=timeoutSecs, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
# this does an inspect of the model and prints the clusters
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
(centers, tupleResultList) = h2o_kmeans.bigCheckResults(self, kmeans, csvPathname, parseResult, 'd', **kwargs)
def kmeans_doit(self, csvFilename, csvPathname, timeoutSecs=30):
print "\nStarting KMeans of", csvFilename
parseKey = h2o_cmd.parseFile(csvPathname=csvPathname, key2=csvFilename + ".hex", timeoutSecs=10)
# hastie has two values, 1 and -1.
# we could not specify cols, but this is more fun
cols = ",".join(map(str,range(11)))
kwargs = {
'k': 1,
'epsilon': 1e-6,
'cols': cols,
'destination_key': 'KMeansModel.hex'
}
start = time.time()
kmeans = h2o_cmd.runKMeansOnly(parseKey=parseKey, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
inspect = h2o_cmd.runInspect(None, key=kmeans['destination_key'])
### print h2o.dump_json(inspect)
# compare this kmeans to the first one. since the files are replications, the results
# should be similar?
KMeansModel = inspect['KMeansModel']
clusters = KMeansModel['clusters'][0]
print "clusters:", h2o.dump_json(clusters)
if self.clusters1:
h2o_kmeans.compareToFirstKMeans(self, clusters, self.clusters1)
else:
self.clusters1 = copy.deepcopy(clusters)
def kmeans_doit(self, csvFilename, csvPathname, timeoutSecs=30):
print "\nStarting KMeans of", csvFilename
parseKey = h2o_cmd.parseFile(csvPathname=csvPathname,
key2=csvFilename + ".hex",
timeoutSecs=10)
# hastie has two values, 1 and -1.
# we could not specify cols, but this is more fun
cols = ",".join(map(str, range(11)))
kwargs = {
'k': 1,
'epsilon': 1e-6,
'cols': cols,
'destination_key': 'KMeansModel.hex'
}
start = time.time()
kmeans = h2o_cmd.runKMeansOnly(parseKey=parseKey, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
inspect = h2o_cmd.runInspect(None, key=kmeans['destination_key'])
### print h2o.dump_json(inspect)
# compare this kmeans to the first one. since the files are replications, the results
# should be similar?
KMeansModel = inspect['KMeansModel']
clusters = KMeansModel['clusters'][0]
print "clusters:", h2o.dump_json(clusters)
if self.clusters1:
h2o_kmeans.compareToFirstKMeans(self, clusters, self.clusters1)
else:
self.clusters1 = copy.deepcopy(clusters)
def test_libsvm(self):
SYNDATASETS_DIR = h2o.make_syn_dir()
for trial in range(2):
csvFilename = "syn_ints.csv"
hex_key = "1.hex"
csvPathname = SYNDATASETS_DIR + '/' + csvFilename
write_syn_dataset(csvPathname, trial)
timeoutSecs = 10
# have to import each time, because h2o deletes source after parse
# PARSE******************************************
# creates csvFilename.hex from file in importFolder dir
# parseResult = h2i.import_parse(path=csvPathname, parser_type='SVMLight', hex_key=hex_key, timeoutSecs=2000)
parseResult = h2i.import_parse(parser_type=PARSER_TYPE,
path=csvPathname,
hex_key=hex_key,
timeoutSecs=2000)
# INSPECT******************************************
start = time.time()
inspect = h2o_cmd.runInspect(key=hex_key, timeoutSecs=360)
print "Inspect:", hex_key, "took", time.time() - start, "seconds"
h2o_cmd.infoFromInspect(inspect, csvFilename)
numRows = inspect['numRows']
numCols = inspect['numCols']
summaryResult = h2o_cmd.runSummary(key=hex_key)
h2o_cmd.infoFromSummary(summaryResult)
if DO_KMEANS:
# KMEANS******************************************
kwargs = {
'k': 3,
'initialization': 'Furthest',
'ignored_cols':
None, #range(11, numCols), # THIS BREAKS THE REST API
'max_iter': 10,
# 'normalize': 0,
# reuse the same seed, to get deterministic results (otherwise sometimes fails
'seed': 265211114317615310,
}
# fails if I put this in kwargs..i.e. source = dest
# 'destination_key': parseResult['destination_key'],
timeoutSecs = 600
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult,
timeoutSecs=timeoutSecs,
**kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
# this does an inspect of the model and prints the clusters
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
(centers, tupleResultList) = h2o_kmeans.bigCheckResults(
self, kmeans, csvPathname, parseResult, 'd', **kwargs)
def test_KMeans2_winesPCA(self):
h2o.beta_features = True
csvPathname = 'winesPCA.csv'
start = time.time()
parseResult = h2i.import_parse(bucket='smalldata', path=csvPathname, schema='put', timeoutSecs=10)
print "parse end on ", csvPathname, 'took', time.time() - start, 'seconds'
h2o.check_sandbox_for_errors()
inspect = h2o_cmd.runInspect(None, parseResult['destination_key'])
print "\n" + csvPathname, \
" numRows:", "{:,}".format(inspect['numRows']), \
" numCols:", "{:,}".format(inspect['numCols'])
kwargs = {
'initialization': 'Furthest',
# 'initialization': '',
# 'initialization': 'PlusPlus',
'max_iter': 50,
'k': 3,
'seed': '265211114317615310',
}
timeoutSecs = 480
# try the same thing 5 times
for trial in range (10):
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans #", trial, "end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
(centers, tupleResultList) = \
h2o_kmeans.bigCheckResults(self, kmeans, csvPathname, parseResult, 'd', **kwargs)
# tupleResultList has tuples = center, rows_per_cluster, sqr_error_per_cluster
# now compare expected vs actual. By sorting on center, we should be able to compare
# since the centers should be separated enough to have the order be consistent
if OLD_KMEANS:
expected = [
([-2.25977535371875, -0.8631572635625001], 64, 83.77800617624794),
([0.16232721958461543, 1.7626161107230771], 65, 111.64440134649745),
([2.7362112930204074, -1.2107751495102044], 49, 62.6290553489474),
]
else:
# error: 258.051462872
expected = [
([-2.23406681758209, -0.7729819755373136], 67, 96.85372611195429),
([0.25174392601612905, 1.792222172419355], 62, 99.21823733913352),
([2.7362112930204074, -1.2107751495102044], 49, 62.6290553489474),
]
# multipliers on the expected values for allowed
# within 2% of best with random seeds?
allowedDelta = (0.01, 0.01, 0.01)
h2o_kmeans.compareResultsToExpected(self, tupleResultList, expected, allowedDelta, trial)
def test_kmeans_sphere5(self):
SYNDATASETS_DIR = h2o.make_syn_dir()
CLUSTERS = 5
SPHERE_PTS = 10000
csvFilename = 'syn_spheres100.csv'
csvPathname = SYNDATASETS_DIR + '/' + csvFilename
centersList = write_spheres_dataset(csvPathname, CLUSTERS, SPHERE_PTS)
print "\nStarting", csvFilename
parseResult = h2i.import_parse(path=csvPathname, schema='put', hex_key=csvFilename + ".hex")
# try 5 times, to see if all inits by h2o are good
for trial in range(5):
# pass SEED so it's repeatable
kwargs = {
'k': CLUSTERS,
'max_iter': 10,
'initialization': 'Furthest',
'cols': None,
'destination_key': 'syn_spheres100.hex',
'seed': SEED
}
timeoutSecs = 30
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult, timeoutSecs=timeoutSecs, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.',\
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
kmeansResult = h2o_cmd.runInspect(key='syn_spheres100.hex')
### print h2o.dump_json(kmeans)
### print h2o.dump_json(kmeansResult)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
# cluster centers can return in any order
clusters = kmeansResult['KMeansModel']['clusters']
clustersSorted = sorted(clusters, key=itemgetter(0))
### print clustersSorted
print "\nh2o result, centers sorted"
print clustersSorted
print "\ngenerated centers"
print centersList
for i,center in enumerate(centersList):
a = center
b = clustersSorted[i]
print "\nexpected:", a
print "h2o:", b # h2o result
aStr = ",".join(map(str,a))
bStr = ",".join(map(str,b))
iStr = str(i)
self.assertAlmostEqual(a[0], b[0], delta=1, msg=aStr+"!="+bStr+". Sorted cluster center "+iStr+" x not correct.")
self.assertAlmostEqual(a[1], b[1], delta=1, msg=aStr+"!="+bStr+". Sorted cluster center "+iStr+" y not correct.")
self.assertAlmostEqual(a[2], b[2], delta=1, msg=aStr+"!="+bStr+". Sorted cluster center "+iStr+" z not correct.")
print "Trial #", trial, "completed"
def test_C_kmeans_prostate(self):
csvFilename = "prostate.csv"
print "\nStarting", csvFilename
csvPathname = h2o.find_file('smalldata/logreg' + '/' + csvFilename)
parseKey = h2o_cmd.parseFile(csvPathname=csvPathname, key2=csvFilename + ".hex")
kwargs = {'k': 1, 'epsilon': 1e-6, 'cols': None, 'destination_key': 'prostate_k.hex'}
kmeans = h2o_cmd.runKMeansOnly(parseKey=parseKey, timeoutSecs=5, **kwargs)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
def test_KMeans_params_rand2(self):
SEED = random.randint(0, sys.maxint)
# if you have to force to redo a test
# SEED =
random.seed(SEED)
print "\nUsing random seed:", SEED
if localhost:
csvFilenameList = [
# ('covtype.data', 60),
('covtype20x.data', 400),
]
else:
csvFilenameList = [
('covtype20x.data', 400),
('covtype200x.data', 2000),
]
importFolderPath = '/home/0xdiag/datasets/standard'
h2i.setupImportFolder(None, importFolderPath)
for csvFilename, timeoutSecs in csvFilenameList:
# creates csvFilename.hex from file in importFolder dir
parseKey = h2i.parseImportFolderFile(None, csvFilename, importFolderPath,
timeoutSecs=2000, pollTimeoutSecs=60)
inspect = h2o_cmd.runInspect(None, parseKey['destination_key'])
csvPathname = importFolderPath + "/" + csvFilename
print "\n" + csvPathname, \
" num_rows:", "{:,}".format(inspect['num_rows']), \
" num_cols:", "{:,}".format(inspect['num_cols'])
paramDict = define_params()
for trial in range(3):
randomV = paramDict['k']
k = random.choice(randomV)
randomV = paramDict['epsilon']
epsilon = random.choice(randomV)
randomV = paramDict['cols']
cols = random.choice(randomV)
kwargs = {'k': k, 'epsilon': epsilon, 'cols': cols,
'destination_key': csvFilename + "_" + str(trial) + '.hex'}
start = time.time()
kmeans = h2o_cmd.runKMeansOnly(parseKey=parseKey, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
### print h2o.dump_json(kmeans)
inspect = h2o_cmd.runInspect(None,key=kmeans['destination_key'])
print h2o.dump_json(inspect)
print "Trial #", trial, "completed\n"
def test_KMeans_winesPCA(self):
csvPathname = 'winesPCA.csv'
start = time.time()
parseResult = h2i.import_parse(bucket='smalldata',
path=csvPathname,
schema='put',
timeoutSecs=10)
print "parse end on ", csvPathname, 'took', time.time(
) - start, 'seconds'
h2o.check_sandbox_for_errors()
inspect = h2o_cmd.runInspect(None, parseResult['destination_key'])
print "\n" + csvPathname, \
" num_rows:", "{:,}".format(inspect['num_rows']), \
" num_cols:", "{:,}".format(inspect['num_cols'])
kwargs = {
#appears not to take 'cols'?
'cols': None,
'initialization': 'Furthest',
'k': 3,
# reuse the same seed, to get deterministic results (otherwise sometimes fails
'seed': 265211114317615310,
}
timeoutSecs = 480
# try the same thing 5 times
for trial in range(10):
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans #", trial, "end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
(centers, tupleResultList) = \
h2o_kmeans.bigCheckResults(self, kmeans, csvPathname, parseResult, 'd', **kwargs)
# tupleResultList has tuples = center, rows_per_cluster, sqr_error_per_cluster
# now compare expected vs actual. By sorting on center, we should be able to compare
# since the centers should be separated enough to have the order be consistent
expected = [
([-2.25977535371875,
-0.8631572635625001], 64, 83.77800617624794),
([0.16232721958461543,
1.7626161107230771], 65, 111.64440134649745),
([2.7362112930204074,
-1.2107751495102044], 49, 62.6290553489474),
]
# multipliers on the expected values for allowed
allowedDelta = (0.01, 0.01, 0.01)
h2o_kmeans.compareResultsToExpected(self, tupleResultList,
expected, allowedDelta, trial)
def test_libsvm(self):
SYNDATASETS_DIR = h2o.make_syn_dir()
for trial in range(2):
csvFilename = "syn_ints.csv"
hex_key = "1.hex"
csvPathname = SYNDATASETS_DIR + '/' + csvFilename
write_syn_dataset(csvPathname, trial)
timeoutSecs = 10
# have to import each time, because h2o deletes source after parse
# PARSE******************************************
# creates csvFilename.hex from file in importFolder dir
# parseResult = h2i.import_parse(path=csvPathname, parser_type='SVMLight', hex_key=hex_key, timeoutSecs=2000)
parseResult = h2i.import_parse(parser_type=PARSER_TYPE, path=csvPathname, hex_key=hex_key, timeoutSecs=2000)
# INSPECT******************************************
start = time.time()
inspect = h2o_cmd.runInspect(key=hex_key, timeoutSecs=360)
print "Inspect:", hex_key, "took", time.time() - start, "seconds"
h2o_cmd.infoFromInspect(inspect, csvFilename)
numRows = inspect['numRows']
numCols = inspect['numCols']
summaryResult = h2o_cmd.runSummary(key=hex_key)
h2o_cmd.infoFromSummary(summaryResult)
if DO_KMEANS:
# KMEANS******************************************
kwargs = {
'k': 3,
'initialization': 'Furthest',
'ignored_cols': None, #range(11, numCols), # THIS BREAKS THE REST API
'max_iter': 10,
# 'normalize': 0,
# reuse the same seed, to get deterministic results (otherwise sometimes fails
'seed': 265211114317615310,
}
# fails if I put this in kwargs..i.e. source = dest
# 'destination_key': parseResult['destination_key'],
timeoutSecs = 600
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult, timeoutSecs=timeoutSecs, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
# this does an inspect of the model and prints the clusters
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
(centers, tupleResultList) = h2o_kmeans.bigCheckResults(self, kmeans, csvPathname, parseResult, 'd', **kwargs)
def test_KMeans_covtype20x_fvec(self):
h2o.beta_features = True
if localhost:
csvFilenameList = [
# 68 secs on my laptop?
('covtype20x.data', 1200, 'cA'),
]
else:
# None is okay for hex_key
csvFilenameList = [
('covtype20x.data', 1200, 'cA'),
# ('covtype200x.data', 1000,'cE'),
]
importFolderPath = "standard"
for csvFilename, timeoutSecs, hex_key in csvFilenameList:
csvPathname = importFolderPath + "/" + csvFilename
# creates csvFilename.hex from file in importFolder dir
start = time.time()
parseResult = h2i.import_parse(
bucket='home-0xdiag-datasets',
path=csvPathname,
timeoutSecs=2000,
hex_key=hex_key) # noise=('JStack', None)
print "parse end on ", csvPathname, 'took', time.time(
) - start, 'seconds'
h2o.check_sandbox_for_errors()
inspect = h2o_cmd.runInspect(None, parseResult['destination_key'])
print "\n" + csvPathname, \
" numRows:", "{:,}".format(inspect['numRows']), \
" numCols:", "{:,}".format(inspect['numCols'])
k = 2
kwargs = {
'max_iter': 25,
'initialization': 'Furthest',
'k': k,
# reuse the same seed, to get deterministic results (otherwise sometimes fails
'seed': 265211114317615310,
}
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
(centers, tupleResultList) = h2o_kmeans.bigCheckResults(
self, kmeans, csvPathname, parseResult, 'd', **kwargs)
def test_KMeans_winesPCA(self):
if localhost:
csvFilenameList = [
#with winesPCA2.csv speciy cols = "1,2"
('winesPCA.csv', 480, 'cA'),
]
else:
# None is okay for key2
csvFilenameList = [
('winesPCA.csv', 480,'cA'),
# ('covtype200x.data', 1000,'cE'),
]
importFolderPath = os.path.abspath(h2o.find_file('smalldata'))
h2i.setupImportFolder(None, importFolderPath)
for csvFilename, timeoutSecs, key2 in csvFilenameList:
csvPathname = importFolderPath + "/" + csvFilename
# creates csvFilename.hex from file in importFolder dir
start = time.time()
parseKey = h2i.parseImportFolderFile(None, 'winesPCA.csv', importFolderPath,
timeoutSecs=2000, key2=key2) # noise=('JStack', None)
print "parse end on ", csvPathname, 'took', time.time() - start, 'seconds'
h2o.check_sandbox_for_errors()
inspect = h2o_cmd.runInspect(None, parseKey['destination_key'])
print "\n" + csvPathname, \
" num_rows:", "{:,}".format(inspect['num_rows']), \
" num_cols:", "{:,}".format(inspect['num_cols'])
kwargs = {
#appears not to take 'cols'?
'cols': None,
'epsilon': 1e-6,
'k': 3
}
start = time.time()
kmeans = h2o_cmd.runKMeansOnly(parseKey=parseKey, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
centers = h2o_kmeans.bigCheckResults(self, kmeans, csvPathname, parseKey, 'd', **kwargs)
print "Expected centers: [-2.276318, -0.965151], with 59 rows."
print " [0.0388763, 1.63886039], with 71 rows."
print " [2.740469, -1.237816], with 48 rows."
model_key = kmeans['destination_key']
kmeansScoreResult = h2o.nodes[0].kmeans_score(
key = parseKey['destination_key'], model_key = model_key)
score = kmeansScoreResult['score']
def test_KMeans_params_rand2(self):
if localhost:
csvFilenameList = [
# ('covtype.data', 60),
('covtype20x.data', 800),
]
else:
csvFilenameList = [
('covtype20x.data', 800),
]
importFolderPath = '/home/0xdiag/datasets/standard'
h2i.setupImportFolder(None, importFolderPath)
for csvFilename, timeoutSecs in csvFilenameList:
# creates csvFilename.hex from file in importFolder dir
parseKey = h2i.parseImportFolderFile(None,
csvFilename,
importFolderPath,
timeoutSecs=2000,
pollTimeoutSecs=60)
inspect = h2o_cmd.runInspect(None, parseKey['destination_key'])
csvPathname = importFolderPath + "/" + csvFilename
print "\n" + csvPathname, \
" num_rows:", "{:,}".format(inspect['num_rows']), \
" num_cols:", "{:,}".format(inspect['num_cols'])
paramDict = define_params(SEED)
for trial in range(3):
# default
params = {
'k': 1,
'destination_key': csvFilename + "_" + str(trial) + '.hex'
}
h2o_kmeans.pickRandKMeansParams(paramDict, params)
kwargs = params.copy()
start = time.time()
kmeans = h2o_cmd.runKMeansOnly(parseKey=parseKey, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
### print h2o.dump_json(kmeans)
inspect = h2o_cmd.runInspect(None,
key=kmeans['destination_key'])
print h2o.dump_json(inspect)
print "Trial #", trial, "completed\n"
def test_KMeansGrid_basic(self):
if localhost:
csvFilenameList = [
# ('covtype.data', 60),
('covtype.data', 800),
]
else:
csvFilenameList = [
('covtype.data', 800),
]
for csvFilename, timeoutSecs in csvFilenameList:
# creates csvFilename.hex from file in importFolder dir
parseResult = h2i.import_parse(bucket='home-0xdiag-datasets', path='standard/covtype.data', schema='local',
timeoutSecs=2000, pollTimeoutSecs=60)
inspect = h2o_cmd.runInspect(None, parseResult['destination_key'])
print "python_source:", parseResult['python_source']
csvPathname = parseResult['python_source']
print "\n" + csvPathname, \
" num_rows:", "{:,}".format(inspect['num_rows']), \
" num_cols:", "{:,}".format(inspect['num_cols'])
destination_key = 'd.hex'
params = {
'k': 2,
# 'initialization': 'Furthest',
'initialization': None,
'seed': 3923021996079663354,
'normalize': 0,
'max_iter': '2',
'destination_key': destination_key
}
for trial in range(3):
kwargs = params.copy()
h2o.beta_features = True
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans (with grid) end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
# This doesn't work (inspecting the model)
# inspect = h2o_cmd.runInspect(None,key=destination_key)
# print h2o.dump_json(inspect)
print "Trial #", trial, "completed\n"
def test_KMeans_covtype20x_fvec(self):
h2o.beta_features = True
if localhost:
csvFilenameList = [
# 68 secs on my laptop?
('covtype20x.data', 1200, 'cA'),
]
else:
# None is okay for hex_key
csvFilenameList = [
('covtype20x.data', 1200,'cA'),
# ('covtype200x.data', 1000,'cE'),
]
importFolderPath = "standard"
for csvFilename, timeoutSecs, hex_key in csvFilenameList:
csvPathname = importFolderPath + "/" + csvFilename
# creates csvFilename.hex from file in importFolder dir
start = time.time()
parseResult = h2i.import_parse(bucket='home-0xdiag-datasets', path=csvPathname,
timeoutSecs=2000, hex_key=hex_key) # noise=('JStack', None)
print "parse end on ", csvPathname, 'took', time.time() - start, 'seconds'
h2o.check_sandbox_for_errors()
inspect = h2o_cmd.runInspect(None, parseResult['destination_key'])
print "\n" + csvPathname, \
" numRows:", "{:,}".format(inspect['numRows']), \
" numCols:", "{:,}".format(inspect['numCols'])
k = 2
kwargs = {
'max_iter': 25,
'initialization': 'Furthest',
'k': k,
# reuse the same seed, to get deterministic results (otherwise sometimes fails
'seed': 265211114317615310,
}
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
(centers, tupleResultList) = h2o_kmeans.bigCheckResults(self, kmeans, csvPathname, parseResult, 'd', **kwargs)
gs = h2o.nodes[0].gap_statistic(source=hex_key, k_max=8)
print "gap_statistic:", h2o.dump_json(gs)
def test_GLM_covtype20x(self):
if localhost:
csvFilenameList = [
# 68 secs on my laptop?
('covtype20x.data', 480, 'cA'),
]
else:
# None is okay for key2
csvFilenameList = [
('covtype20x.data', 480,'cA'),
# ('covtype200x.data', 1000,'cE'),
]
# a browser window too, just because we can
h2b.browseTheCloud()
importFolderPath = '/home/0xdiag/datasets'
h2i.setupImportFolder(None, importFolderPath)
for csvFilename, timeoutSecs, key2 in csvFilenameList:
csvPathname = importFolderPath + "/" + csvFilename
# creates csvFilename.hex from file in importFolder dir
start = time.time()
parseKey = h2i.parseImportFolderFile(None, csvFilename, importFolderPath,
timeoutSecs=2000, key2=key2, noise=('JStack', None))
print "parse end on ", csvPathname, 'took', time.time() - start, 'seconds'
h2o.check_sandbox_for_errors()
inspect = h2o_cmd.runInspect(None, parseKey['destination_key'])
print "\n" + csvPathname, \
" num_rows:", "{:,}".format(inspect['num_rows']), \
" num_cols:", "{:,}".format(inspect['num_cols'])
kwargs = {
'cols': None,
'epsilon': 1e-4,
'k': 2
}
start = time.time()
kmeans = h2o_cmd.runKMeansOnly(parseKey=parseKey, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
### print h2o.dump_json(kmeans)
inspect = h2o_cmd.runInspect(None,key=kmeans['destination_key'])
print h2o.dump_json(inspect)
def test_C_kmeans_prostate(self):
csvFilename = "prostate.csv"
print "\nStarting", csvFilename
csvPathname = h2o.find_file('smalldata/logreg' + '/' + csvFilename)
parseKey = h2o_cmd.parseFile(csvPathname=csvPathname, key2=csvFilename + ".hex")
# loop, to see if we get same centers
for i in range(2):
kwargs = {'k': 3, 'epsilon': 1e-6, 'cols': None, 'destination_key': 'prostate_k.hex'}
kmeans = h2o_cmd.runKMeansOnly(parseKey=parseKey, timeoutSecs=5, **kwargs)
model_key = kmeans['destination_key']
kmeansResult = h2o_cmd.runInspect(key=model_key)
centers = kmeansResult['KMeansModel']['clusters']
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
show_results(csvPathname, parseKey, model_key, centers, 'd')
def test_kmeans_benign(self):
importFolderPath = "logreg"
csvFilename = "benign.csv"
hex_key = "benign.hex"
csvPathname = importFolderPath + "/" + csvFilename
parseResult = h2i.import_parse(bucket='smalldata', path=csvPathname, hex_key=hex_key, checkHeader=1,
timeoutSecs=180, doSummary=False)
numRows, numCols, parse_key = h2o_cmd.infoFromParse(parseResult)
inspectResult = h2o_cmd.runInspect(key=parse_key)
missingList, labelList, numRows, numCols = h2o_cmd.infoFromInspect(inspectResult)
expected = [
([8.86, 2.43, 35.53, 0.31, 13.22, 1.47, 1.33, 20.06, 13.08, 0.53, 2.12, 128.61, 35.33, 1.57], 49, None),
([33.47, 2.29, 50.92, 0.34, 12.82, 1.33, 1.36, 21.43, 13.30, 0.37, 2.52, 125.40, 43.91, 1.79], 87, None),
([27.64, 2.87, 48.11, 0.09, 11.80, 0.98, 1.51, 21.02, 12.53, 0.58, 2.89, 171.27, 42.73, 1.53], 55, None),
([26.00, 2.67, 46.67, 0.00, 13.00, 1.33, 1.67, 21.56, 11.44, 0.22, 2.89, 234.56, 39.22, 1.56], 9, None),
]
# all are multipliers of expected tuple value
allowedDelta = (0.01, 0.01, 0.01, 0.01)
# loop, to see if we get same centers
# no cols ignored
labelListUsed = list(labelList)
numColsUsed = numCols
for trial in range(5):
kmeansSeed = random.randint(0, sys.maxint)
# kmeansSeed = 6655548259421773879
parameters = {
'validation_frame': parse_key,
'ignored_columns': None,
'score_each_iteration': False,
'K': 4,
'max_iters': 50,
'normalize': False,
'seed': kmeansSeed,
'init': 'PlusPlus',
}
model_key = 'benign_k.hex'
kmeansResult = h2o.n0.build_model(
algo='kmeans',
destination_key=model_key,
training_frame=parse_key,
parameters=parameters,
timeoutSecs=10)
modelResult = h2o.n0.models(key=model_key)
# this prints too
tuplesSorted, iters, mse, names = \
h2o_kmeans.simpleCheckKMeans(self, modelResult, parameters, numRows, numColsUsed, labelListUsed)
h2o_cmd.runStoreView()
# zip with * is it's own inverse here. It's sorted by centers for easy comparisons
ids, mses, rows, clusters = zip(*tuplesSorted)
def test_KMeans_params_rand2_fvec(self):
h2o.beta_features = True
if localhost:
csvFilenameList = [
# ('covtype.data', 60),
('covtype.data', 800),
]
else:
csvFilenameList = [
('covtype.data', 800),
]
importFolderPath = "standard"
for csvFilename, timeoutSecs in csvFilenameList:
# creates csvFilename.hex from file in importFolder dir
csvPathname = importFolderPath + "/" + csvFilename
parseResult = h2i.import_parse(bucket='home-0xdiag-datasets',
path=csvPathname,
timeoutSecs=2000,
pollTimeoutSecs=60)
inspect = h2o_cmd.runInspect(None, parseResult['destination_key'])
print "\n" + csvPathname, \
" numRows:", "{:,}".format(inspect['numRows']), \
" numCols:", "{:,}".format(inspect['numCols'])
paramDict = define_params(SEED)
for trial in range(3):
# default
params = {
'max_iter': 20,
'k': 1,
'destination_key': csvFilename + "_" + str(trial) + '.hex'
}
h2o_kmeans.pickRandKMeansParams(paramDict, params)
kwargs = params.copy()
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
### print h2o.dump_json(kmeans)
print "Trial #", trial, "completed\n"
def test_kmeans_sphere3(self):
SYNDATASETS_DIR = h2o.make_syn_dir()
csvFilename = 'syn_spheres3_' + str(SEED) + '.csv'
csvPathname = SYNDATASETS_DIR + '/' + csvFilename
write_syn_dataset(csvPathname, 1000000, SEED)
print "\nStarting", csvFilename
parseKey = h2o_cmd.parseFile(csvPathname=csvPathname,
key2=csvFilename + ".hex")
kwargs = {
'k': 3,
'epsilon': 1e-6,
'cols': None,
'destination_key': 'spheres3.hex'
}
timeoutSecs = 30
start = time.time()
kmeans = h2o_cmd.runKMeansOnly(parseKey=parseKey,
timeoutSecs=timeoutSecs,
**kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', "%d pct. of timeout" % (
(elapsed / timeoutSecs) * 100)
kmeansResult = h2o_cmd.runInspect(key='spheres3.hex')
### print h2o.dump_json(kmeans)
print h2o.dump_json(kmeansResult)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
clusters = kmeansResult['KMeansModel']['clusters']
# cluster centers can return in any order
clustersSorted = sorted(clusters, key=itemgetter(0))
self.assertAlmostEqual(clustersSorted[0][0], 100, delta=.2)
self.assertAlmostEqual(clustersSorted[1][0], 200, delta=.2)
self.assertAlmostEqual(clustersSorted[2][0], 300, delta=.2)
self.assertAlmostEqual(clustersSorted[0][1], 100, delta=.2)
self.assertAlmostEqual(clustersSorted[1][1], 200, delta=.2)
self.assertAlmostEqual(clustersSorted[2][1], 300, delta=.2)
self.assertAlmostEqual(clustersSorted[0][2], 100, delta=.2)
self.assertAlmostEqual(clustersSorted[1][2], 200, delta=.2)
self.assertAlmostEqual(clustersSorted[2][2], 300, delta=.2)
def test_KMeans_winesPCA(self):
csvPathname = h2o.find_file('smalldata/winesPCA.csv')
start = time.time()
parseKey = h2o_cmd.parseFile(csvPathname=csvPathname, timeoutSecs=10)
print "parse end on ", csvPathname, 'took', time.time() - start, 'seconds'
h2o.check_sandbox_for_errors()
inspect = h2o_cmd.runInspect(None, parseKey['destination_key'])
print "\n" + csvPathname, \
" num_rows:", "{:,}".format(inspect['num_rows']), \
" num_cols:", "{:,}".format(inspect['num_cols'])
kwargs = {
#appears not to take 'cols'?
'cols': None,
'initialization': 'Furthest',
'k': 3,
# reuse the same seed, to get deterministic results (otherwise sometimes fails
'seed': 265211114317615310,
}
timeoutSecs = 480
# try the same thing 5 times
for trial in range (10):
start = time.time()
kmeans = h2o_cmd.runKMeansOnly(parseKey=parseKey, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans #", trial, "end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
(centers, tupleResultList) = \
h2o_kmeans.bigCheckResults(self, kmeans, csvPathname, parseKey, 'd', **kwargs)
# tupleResultList has tuples = center, rows_per_cluster, sqr_error_per_cluster
# now compare expected vs actual. By sorting on center, we should be able to compare
# since the centers should be separated enough to have the order be consistent
expected = [
([-2.25977535371875, -0.8631572635625001], 64, 83.77800617624794) ,
([0.16232721958461543, 1.7626161107230771], 65, 111.64440134649745) ,
([2.7362112930204074, -1.2107751495102044], 49, 62.6290553489474) ,
]
# multipliers on the expected values for allowed
allowedDelta = (0.01, 0.01, 0.01)
h2o_kmeans.compareResultsToExpected(self, tupleResultList, expected, allowedDelta, trial)
def test_KMeans_params_rand2(self):
if localhost:
csvFilenameList = [
# ('covtype.data', 60),
('covtype.data', 800),
]
else:
csvFilenameList = [
('covtype.data', 800),
]
importFolderPath = '/home/0xdiag/datasets/standard'
h2i.setupImportFolder(None, importFolderPath)
for csvFilename, timeoutSecs in csvFilenameList:
# creates csvFilename.hex from file in importFolder dir
parseKey = h2i.parseImportFolderFile(None, csvFilename, importFolderPath,
timeoutSecs=2000, pollTimeoutSecs=60)
inspect = h2o_cmd.runInspect(None, parseKey['destination_key'])
csvPathname = importFolderPath + "/" + csvFilename
print "\n" + csvPathname, \
" num_rows:", "{:,}".format(inspect['num_rows']), \
" num_cols:", "{:,}".format(inspect['num_cols'])
paramDict = define_params(SEED)
for trial in range(3):
# default
params = {'k': 1 }
# 'destination_key': csvFilename + "_" + str(trial) + '.hex'}
h2o_kmeans.pickRandKMeansParams(paramDict, params)
kwargs = params.copy()
start = time.time()
kmeans = h2o_cmd.runKMeansGridOnly(parseKey=parseKey, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans grid end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
### print h2o.dump_json(kmeans)
inspect = h2o_cmd.runInspect(None,key=kmeans['destination_key'])
print h2o.dump_json(inspect)
print "Trial #", trial, "completed\n"
def test_KMeans_params_rand2_fvec(self):
h2o.beta_features = True
if localhost:
csvFilenameList = [
# ('covtype.data', 60),
('covtype.data', 800),
]
else:
csvFilenameList = [
('covtype.data', 800),
]
importFolderPath = "standard"
for csvFilename, timeoutSecs in csvFilenameList:
# creates csvFilename.hex from file in importFolder dir
csvPathname = importFolderPath + "/" + csvFilename
parseResult = h2i.import_parse(bucket='home-0xdiag-datasets', path=csvPathname,
timeoutSecs=2000, pollTimeoutSecs=60)
inspect = h2o_cmd.runInspect(None, parseResult['destination_key'])
print "\n" + csvPathname, \
" numRows:", "{:,}".format(inspect['numRows']), \
" numCols:", "{:,}".format(inspect['numCols'])
paramDict = define_params(SEED)
for trial in range(3):
# default
params = {
'max_iter': 20,
'k': 1,
'destination_key': csvFilename + "_" + str(trial) + '.hex'
}
h2o_kmeans.pickRandKMeansParams(paramDict, params)
kwargs = params.copy()
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
### print h2o.dump_json(kmeans)
print "Trial #", trial, "completed\n"
def test_KMeans_covtype20x(self):
if localhost:
csvFilenameList = [
# 68 secs on my laptop?
('covtype20x.data', 480, 'cA'),
]
else:
# None is okay for key2
csvFilenameList = [
('covtype20x.data', 480,'cA'),
# ('covtype200x.data', 1000,'cE'),
]
importFolderPath = '/home/0xdiag/datasets/standard'
h2i.setupImportFolder(None, importFolderPath)
for csvFilename, timeoutSecs, key2 in csvFilenameList:
csvPathname = importFolderPath + "/" + csvFilename
# creates csvFilename.hex from file in importFolder dir
start = time.time()
parseKey = h2i.parseImportFolderFile(None, csvFilename, importFolderPath,
timeoutSecs=2000, key2=key2) # noise=('JStack', None)
print "parse end on ", csvPathname, 'took', time.time() - start, 'seconds'
h2o.check_sandbox_for_errors()
inspect = h2o_cmd.runInspect(None, parseKey['destination_key'])
print "\n" + csvPathname, \
" num_rows:", "{:,}".format(inspect['num_rows']), \
" num_cols:", "{:,}".format(inspect['num_cols'])
kwargs = {
'cols': None,
'epsilon': 1e-4,
'k': 2,
# reuse the same seed, to get deterministic results (otherwise sometimes fails
'seed': 265211114317615310,
}
start = time.time()
kmeans = h2o_cmd.runKMeansOnly(parseKey=parseKey, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
(centers, tupleResultList) = h2o_kmeans.bigCheckResults(self, kmeans, csvPathname, parseKey, 'd', **kwargs)
def test_many_cols_with_syn(self):
SYNDATASETS_DIR = h2o.make_syn_dir()
tryList = [
(100, 11, 'cA', 5),
(100, 10, 'cB', 5),
(100, 9, 'cC', 5),
(100, 8, 'cD', 5),
(100, 7, 'cE', 5),
(100, 6, 'cF', 5),
(100, 5, 'cG', 5),
]
### h2b.browseTheCloud()
lenNodes = len(h2o.nodes)
cnum = 0
for (rowCount, colCount, key2, timeoutSecs) in tryList:
cnum += 1
csvFilename = 'syn_' + str(SEED) + "_" + str(rowCount) + 'x' + str(colCount) + '.csv'
csvPathname = SYNDATASETS_DIR + '/' + csvFilename
print "Creating random", csvPathname
write_syn_dataset(csvPathname, rowCount, colCount, SEED)
parseKey = h2o_cmd.parseFile(csvPathname=csvPathname, key2=csvFilename + ".hex")
print "Parse result['destination_key']:", parseKey['destination_key']
kwargs = {'k': 2, 'epsilon': 1e-6, 'cols': None, 'destination_key': 'benign_k.hex'}
kmeans = h2o_cmd.runKMeansOnly(parseKey=parseKey, timeoutSecs=5, **kwargs)
model_key = kmeans['destination_key']
kmeansResult = h2o_cmd.runInspect(key=model_key)
## h2o.nodes[0].kmeans_apply(data_key=parseKey['destination_key'], model_key=model_key, destination_key='a')
# this is failing for some reason
## h2o.nodes[0].kmeans_score(key=parseKey['destination_key'], model_key=model_key)
clusters = kmeansResult['KMeansModel']['clusters']
for i,c in enumerate(clusters):
print "clusters["+str(i)+"]: ", clusters[i]
## print h2o.dump_json(kmeans)
## print h2o.dump_json(kmeansResult)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
def test_C_kmeans_prostate(self):
csvFilename = "prostate.csv"
print "\nStarting", csvFilename
csvPathname = h2o.find_file('smalldata/logreg' + '/' + csvFilename)
parseKey = h2o_cmd.parseFile(csvPathname=csvPathname,
key2=csvFilename + ".hex")
kwargs = {
'k': 1,
'epsilon': 1e-6,
'cols': None,
'destination_key': 'prostate_k.hex'
}
kmeans = h2o_cmd.runKMeansOnly(parseKey=parseKey,
timeoutSecs=5,
**kwargs)
kmeansResult = h2o_cmd.runInspect(key='prostate_k.hex')
print h2o.dump_json(kmeans)
print h2o.dump_json(kmeansResult)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
def test_KMeans_covtype_fvec(self):
h2o.beta_features = True
csvFilenameList = [
('covtype.data', 800),
]
importFolderPath = "standard"
for csvFilename, timeoutSecs in csvFilenameList:
# creates csvFilename.hex from file in importFolder dir
csvPathname = importFolderPath + "/" + csvFilename
parseResult = h2i.import_parse(bucket='home-0xdiag-datasets',
path=csvPathname,
timeoutSecs=2000,
pollTimeoutSecs=60)
inspect = h2o_cmd.runInspect(None, parseResult['destination_key'])
print "\n" + csvPathname, \
" numRows:", "{:,}".format(inspect['numRows']), \
" numCols:", "{:,}".format(inspect['numCols'])
for trial in range(3):
kwargs = {
'source': u'covtype.hex',
'destination_key': 'covtype.data_2.hex',
'initialization': 'Furthest',
# 'max_iter': 20,
'max_iter': 50,
'k': 2,
}
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
### print h2o.dump_json(kmeans)
print "Trial #", trial, "completed\n"
def test_kmeans_sphere3(self):
SYNDATASETS_DIR = h2o.make_syn_dir()
csvFilename = 'syn_spheres3_' + str(SEED) + '.csv'
csvPathname = SYNDATASETS_DIR + '/' + csvFilename
write_syn_dataset(csvPathname, 1000000, SEED)
print "\nStarting", csvFilename
parseKey = h2o_cmd.parseFile(csvPathname=csvPathname, key2=csvFilename + ".hex")
kwargs = {'k': 3, 'epsilon': 1e-6, 'cols': None, 'destination_key': 'spheres3.hex'}
timeoutSecs = 30
start = time.time()
kmeans = h2o_cmd.runKMeansOnly(parseKey=parseKey, timeoutSecs=timeoutSecs, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', "%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
kmeansResult = h2o_cmd.runInspect(key='spheres3.hex')
### print h2o.dump_json(kmeans)
print h2o.dump_json(kmeansResult)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
clusters = kmeansResult['KMeansModel']['clusters']
# cluster centers can return in any order
clustersSorted = sorted(clusters, key=itemgetter(0))
self.assertAlmostEqual(clustersSorted[0][0],100,delta=.2)
self.assertAlmostEqual(clustersSorted[1][0],200,delta=.2)
self.assertAlmostEqual(clustersSorted[2][0],300,delta=.2)
self.assertAlmostEqual(clustersSorted[0][1],100,delta=.2)
self.assertAlmostEqual(clustersSorted[1][1],200,delta=.2)
self.assertAlmostEqual(clustersSorted[2][1],300,delta=.2)
self.assertAlmostEqual(clustersSorted[0][2],100,delta=.2)
self.assertAlmostEqual(clustersSorted[1][2],200,delta=.2)
self.assertAlmostEqual(clustersSorted[2][2],300,delta=.2)
def test_KMeans_allstate_s3n_thru_hdfs(self):
bucket = 'home-0xdiag-datasets'
importFolderPath = 'allstate'
csvFilename = "train_set.csv"
csvPathname = importFolderPath + "/" + csvFilename
timeoutSecs = 600
trialMax = 3
for trial in range(trialMax):
trialStart = time.time()
hex_key = csvFilename + "_" + str(trial) + ".hex"
start = time.time()
parseResult = h2i.import_parse(bucket=bucket,
path=csvPathname,
schema='s3n',
hex_key=hex_key,
timeoutSecs=timeoutSecs,
retryDelaySecs=10,
pollTimeoutSecs=60)
elapsed = time.time() - start
print 'h2o reported parse time:', parseResult['response']['time']
print "parse end on ", csvPathname, 'took', elapsed, 'seconds',\
"%d pct. of timeout" % ((elapsed*100)/timeoutSecs)
print "parse result:", parseResult['destination_key']
kwargs = {'cols': None, 'initialization': 'Furthest', 'k': 12}
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=120, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvFilename, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
inspect = h2o_cmd.runInspect(None, key=kmeans['destination_key'])
print h2o.dump_json(inspect)
print "Trial #", trial, "completed in", time.time() - trialStart, "seconds.", \
def test_KMeans_covtype_fvec(self):
h2o.beta_features = True
csvFilenameList = [
('covtype.data', 800),
]
importFolderPath = "standard"
for csvFilename, timeoutSecs in csvFilenameList:
# creates csvFilename.hex from file in importFolder dir
csvPathname = importFolderPath + "/" + csvFilename
parseResult = h2i.import_parse(bucket='home-0xdiag-datasets', path=csvPathname,
timeoutSecs=2000, pollTimeoutSecs=60)
inspect = h2o_cmd.runInspect(None, parseResult['destination_key'])
print "\n" + csvPathname, \
" numRows:", "{:,}".format(inspect['numRows']), \
" numCols:", "{:,}".format(inspect['numCols'])
for trial in range(3):
kwargs = {
'k': 3,
'initialization': 'Furthest',
'ignored_cols': range(11, inspect['numCols']),
'max_iter': 10,
# 'normalize': 0,
# reuse the same seed, to get deterministic results
'seed': 265211114317615310
}
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
### print h2o.dump_json(kmeans)
print "Trial #", trial, "completed\n"
def test_KMeans_allstate_s3n_thru_hdfs(self):
bucket = 'home-0xdiag-datasets'
importFolderPath = 'allstate'
csvFilename = "train_set.csv"
csvPathname = importFolderPath + "/" + csvFilename
timeoutSecs = 600
trialMax = 3
for trial in range(trialMax):
trialStart = time.time()
hex_key = csvFilename + "_" + str(trial) + ".hex"
start = time.time()
parseResult = h2i.import_parse(bucket=bucket, path=csvPathname, schema='s3n', hex_key=hex_key,
timeoutSecs=timeoutSecs, retryDelaySecs=10, pollTimeoutSecs=60)
elapsed = time.time() - start
print 'h2o reported parse time:', parseResult['response']['time']
print "parse end on ", csvPathname, 'took', elapsed, 'seconds',\
"%d pct. of timeout" % ((elapsed*100)/timeoutSecs)
print "parse result:", parseResult['destination_key']
kwargs = {
'cols': None,
'initialization': 'Furthest',
'k': 12
}
start = time.time()
kmeans = h2o_cmd.runKMeansOnly(parseResult=parseResult, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=120, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvFilename, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
inspect = h2o_cmd.runInspect(None,key=kmeans['destination_key'])
print h2o.dump_json(inspect)
print "Trial #", trial, "completed in", time.time() - trialStart, "seconds.", \
def test_KMeans_covtype(self):
csvFilenameList = [
('covtype.data', 800),
]
importFolderPath = "standard"
for csvFilename, timeoutSecs in csvFilenameList:
# creates csvFilename.hex from file in importFolder dir
csvPathname = importFolderPath + "/" + csvFilename
parseResult = h2i.import_parse(bucket='home-0xdiag-datasets', path=csvPathname,
timeoutSecs=2000, pollTimeoutSecs=60)
inspect = h2o_cmd.runInspect(None, parseResult['destination_key'])
print "\n" + csvPathname, \
" num_rows:", "{:,}".format(inspect['num_rows']), \
" num_cols:", "{:,}".format(inspect['num_cols'])
for trial in range(3):
kwargs = {
'source_key': u'covtype.hex',
'destination_key': 'covtype.data_2.hex',
'initialization': 'Furthest',
# 'max_iter': 20,
'max_iter': 50,
'k': 2,
}
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
### print h2o.dump_json(kmeans)
print "Trial #", trial, "completed\n"
def test_parse_bounds_libsvm(self):
# just do the import folder once
importFolderPath = "/home/0xdiag/datasets/libsvm"
# make the timeout variable per dataset. it can be 10 secs for covtype 20x (col key creation)
# so probably 10x that for covtype200
csvFilenameList = [
("mnist_train.svm", "cM", 30, 1),
# FIX! fails KMeansScore
("tmc2007_train.svm", "cJ", 30, 1),
("covtype.binary.svm", "cC", 30, 1),
("colon-cancer.svm", "cA", 30, 1),
("connect4.svm", "cB", 30, 1),
("duke.svm", "cD", 30, 1),
# too many features? 150K inspect timeout?
# ("E2006.train.svm", "cE", 30, 1),
("gisette_scale.svm", "cF", 30, 1),
("mushrooms.svm", "cG", 30, 1),
("news20.svm", "cH", 30, 1),
("syn_6_1000_10.svm", "cK", 30, 1),
("syn_0_100_1000.svm", "cL", 30, 1),
# normal csv
]
### csvFilenameList = random.sample(csvFilenameAll,1)
# h2b.browseTheCloud()
lenNodes = len(h2o.nodes)
firstDone = False
for (csvFilename, key2, timeoutSecs, resultMult) in csvFilenameList:
# have to import each time, because h2o deletes source after parse
h2i.setupImportFolder(None, importFolderPath)
csvPathname = importFolderPath + "/" + csvFilename
# PARSE******************************************
# creates csvFilename.hex from file in importFolder dir
parseKey = h2i.parseImportFolderFile(None, csvFilename, importFolderPath,
key2=key2, timeoutSecs=2000)
print csvPathname, 'parse time:', parseKey['response']['time']
print "Parse result['destination_key']:", parseKey['destination_key']
# INSPECT******************************************
start = time.time()
inspect = h2o_cmd.runInspect(None, parseKey['destination_key'], timeoutSecs=360)
print "Inspect:", parseKey['destination_key'], "took", time.time() - start, "seconds"
h2o_cmd.infoFromInspect(inspect, csvFilename)
# KMEANS******************************************
for trial in range(2):
kwargs = {
'k': 3,
'epsilon': 1e-6,
# 'cols': 2,
# 'max_iter': 10,
# 'normalize': 0,
# reuse the same seed, to get deterministic results (otherwise sometimes fails
'seed': 265211114317615310
}
# fails if I put this in kwargs..i.e. source = dest
# 'destination_key': parseKey['destination_key'],
timeoutSecs = 600
start = time.time()
kmeans = h2o_cmd.runKMeansOnly(parseKey=parseKey, timeoutSecs=timeoutSecs, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
# this does an inspect of the model and prints the clusters
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
(centers, tupleResultList) = h2o_kmeans.bigCheckResults(self, kmeans, csvPathname, parseKey, 'd', **kwargs)
def test_KMeans_create_frame_fvec(self):
for trial in range(20):
cfParamDict = define_create_frame_params(SEED)
# default
params = {
'rows': 5,
'cols': 10
}
h2o_util.pickRandParams(cfParamDict, params)
i = params.get('integer_fraction', None)
c = params.get('categorical_fraction', None)
r = params.get('randomize', None)
v = params.get('value', None)
# h2o does some strict checking on the combinations of these things
# fractions have to add up to <= 1 and only be used if randomize
# h2o default randomize=1?
if r:
if not i:
i = 0
if not c:
c = 0
if (i and c) and (i + c) >= 1.0:
c = 1.0 - i
params['integer_fraction'] = i
params['categorical_fraction'] = c
params['value'] = None
else:
params['randomize'] = 0
params['integer_fraction'] = 0
params['categorical_fraction'] = 0
kwargs = params.copy()
timeoutSecs = 300
hex_key = 'temp_%s.hex' % trial
cfResult = h2o.nodes[0].create_frame(key=hex_key, timeoutSecs=timeoutSecs, **kwargs)
inspect = h2o_cmd.runInspect(None, hex_key)
print "\n%s" % hex_key, \
" numRows:", "{:,}".format(inspect['numRows']), \
" numCols:", "{:,}".format(inspect['numCols'])
kmeansParamDict = define_KMeans_params(SEED)
# default
params = {
'max_iter': 20,
'k': 1,
'destination_key': "KM_" + str(trial) + '.hex'
}
h2o_kmeans.pickRandKMeansParams(kmeansParamDict, params)
kwargs = params.copy()
start = time.time()
parseResult = {'destination_key': hex_key }
kmeans = h2o_cmd.runKMeans(parseResult=parseResult, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans trial %s end on ", trial, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
### print h2o.dump_json(kmeans)
print "Trial #", trial, "completed\n"
def test_KMeans_libsvm_fvec(self):
h2o.beta_features = True
# just do the import folder once
# make the timeout variable per dataset. it can be 10 secs for covtype 20x (col key creation)
# so probably 10x that for covtype200
csvFilenameList = [
# FIX! fails KMeansScore
("colon-cancer.svm", "cA", 30, 1),
("connect4.svm", "cB", 30, 1),
("covtype.binary.svm", "cC", 30, 1),
# multi-label class
# ("tmc2007_train.svm", "cJ", 30, 1),
("mnist_train.svm", "cM", 30, 1),
("duke.svm", "cD", 30, 1),
# too many features? 150K inspect timeout?
# ("E2006.train.svm", "cE", 30, 1),
("gisette_scale.svm", "cF", 120, 1
), #Summary2 is slow with 5001 columns
("mushrooms.svm", "cG", 30, 1),
# ("news20.svm", "cH", 120, 1), #Summary2 is very slow - disable for now
("syn_6_1000_10.svm", "cK", 30, 1),
("syn_0_100_1000.svm", "cL", 30, 1),
]
### csvFilenameList = random.sample(csvFilenameAll,1)
# h2b.browseTheCloud()
lenNodes = len(h2o.nodes)
firstDone = False
importFolderPath = "libsvm"
for (csvFilename, hex_key, timeoutSecs, resultMult) in csvFilenameList:
# have to import each time, because h2o deletes source after parse
csvPathname = importFolderPath + "/" + csvFilename
# PARSE******************************************
# creates csvFilename.hex from file in importFolder dir
parseResult = h2i.import_parse(bucket='home-0xdiag-datasets',
path=csvPathname,
hex_key=hex_key,
timeoutSecs=2000)
print "Parse result['destination_key']:", parseResult[
'destination_key']
# INSPECT******************************************
start = time.time()
inspect = h2o_cmd.runInspect(None,
parseResult['destination_key'],
timeoutSecs=360)
print "Inspect:", parseResult[
'destination_key'], "took", time.time() - start, "seconds"
h2o_cmd.infoFromInspect(inspect, csvFilename)
numRows = inspect['numRows']
numCols = inspect['numCols']
# KMEANS******************************************
for trial in range(1):
kwargs = {
'k': 3,
'initialization': 'Furthest',
'ignored_cols':
None, #range(11, numCols), # THIS BREAKS THE REST API
'max_iter': 10,
# 'normalize': 0,
# reuse the same seed, to get deterministic results (otherwise sometimes fails
'seed': 265211114317615310,
}
# fails if I put this in kwargs..i.e. source = dest
# 'destination_key': parseResult['destination_key'],
timeoutSecs = 600
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult,
timeoutSecs=timeoutSecs,
**kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
# this does an inspect of the model and prints the clusters
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
(centers, tupleResultList) = h2o_kmeans.bigCheckResults(
self, kmeans, csvPathname, parseResult, 'd', **kwargs)
parseResult = h2i.import_parse(bucket=bucket, path=csvPathname, schema='s3n', hex_key=hex_key,
timeoutSecs=timeoutSecs, retryDelaySecs=10, pollTimeoutSecs=60))
elapsed = time.time() - start
print 'h2o reported parse time:', parseResult['response']['time']
print "parse end on ", hex_key, 'took', elapsed, 'seconds',\
"%d pct. of timeout" % ((elapsed*100)/timeoutSecs)
print "parse result:", parseResult['destination_key']
kwargs = {
'cols': None,
'initialization': 'Furthest',
'k': 12
}
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=120, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvFilename, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
### print h2o.dump_json(kmeans)
inspect = h2o_cmd.runInspect(None,key=kmeans['destination_key'])
print h2o.dump_json(inspect)
print "Trial #", trial, "completed in", time.time() - trialStart, "seconds.", \
if __name__ == '__main__':
h2o.unit_main()
def test_KMeans_allstate_s3n_thru_hdfs(self):
# csvFilename = "covtype20x.data"
# csvPathname = csvFilename
csvFilename = "CAT*"
csvPathname = "cats/" + csvFilename
# https://s3.amazonaws.com/home-0xdiag-datasets/allstate/train_set.csv
URI = "s3n://home-0xdiag-datasets/"
s3nKey = URI + csvPathname
trialMax = 1
for trial in range(trialMax):
trialStart = time.time()
# since we delete the key, we have to re-import every iteration
# s3n URI thru HDFS is not typical.
importHDFSResult = h2o.nodes[0].import_hdfs(URI)
s3nFullList = importHDFSResult['succeeded']
### print "s3nFullList:", h2o.dump_json(s3nFullList)
self.assertGreater(len(s3nFullList), 8,
"Didn't see more than 8 files in s3n?")
storeView = h2o.nodes[0].store_view()
### print "storeView:", h2o.dump_json(storeView)
for s in storeView['keys']:
print "\nkey:", s['key']
if 'rows' in s:
print "rows:", s['rows'], "value_size_bytes:", s[
'value_size_bytes']
key2 = csvFilename + "_" + str(trial) + ".hex"
print "Loading s3n key: ", s3nKey, 'thru HDFS'
# ec2 is about 400 secs on four m2.4xlarge nodes
# should be less on more nodes?
timeoutSecs = 600
start = time.time()
parseKey = h2o.nodes[0].parse(s3nKey,
key2,
timeoutSecs=timeoutSecs,
retryDelaySecs=10,
pollTimeoutSecs=60,
noise=('JStack', None))
elapsed = time.time() - start
print s3nKey, 'h2o reported parse time:', parseKey['response'][
'time']
print "parse end on ", s3nKey, 'took', elapsed, 'seconds',\
"%d pct. of timeout" % ((elapsed*100)/timeoutSecs)
print "parse result:", parseKey['destination_key']
kwargs = {'cols': None, 'epsilon': 1e-6, 'k': 12}
start = time.time()
kmeans = h2o_cmd.runKMeansOnly(parseKey=parseKey, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=120, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
### print h2o.dump_json(kmeans)
inspect = h2o_cmd.runInspect(None, key=kmeans['destination_key'])
print h2o.dump_json(inspect)
print "Deleting key in H2O so we get it from S3 (if ec2) or nfs again.", \
"Otherwise it would just parse the cached key."
storeView = h2o.nodes[0].store_view()
# pattern matching problem
# h2o removes key afte parse now
### print "Removing", s3nKey
### removeKeyResult = h2o.nodes[0].remove_key(key=s3nKey)
print "Trial #", trial, "completed in", time.time() - trialStart, "seconds.", \
def test_kmeans_sphere100(self):
SYNDATASETS_DIR = h2o.make_syn_dir()
csvFilename = 'syn_spheres100.csv'
csvPathname = SYNDATASETS_DIR + '/' + csvFilename
centersList = write_spheres_dataset(csvPathname, CLUSTERS, SPHERE_PTS)
if SHUFFLE_SPHERES:
# since we create spheres in order
csvFilename2 = 'syn_spheres100_shuffled.csv'
csvPathname2 = SYNDATASETS_DIR + '/' + csvFilename2
h2o_util.file_shuffle(csvPathname, csvPathname2)
else:
csvFilename2 = csvFilename
csvPathname2 = csvPathname
print "\nStarting", csvFilename
parseResult = h2i.import_parse(path=csvPathname2, schema='put', hex_key=csvFilename2 + ".hex")
### h2b.browseTheCloud()
# try 5 times, to see if all inits by h2o are good
# does it break if cols is not specified?
cols = ",".join(map(str,range(DIMENSIONS)))
for trial in range(10):
kwargs = {
'k': CLUSTERS,
'initialization': 'Furthest',
'cols': cols,
'destination_key': 'syn_spheres100.hex'
}
timeoutSecs = 100
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult, timeoutSecs=timeoutSecs, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.',\
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
kmeansResult = h2o_cmd.runInspect(key='syn_spheres100.hex')
# print h2o.dump_json(kmeansResult)
### print h2o.dump_json(kmeans)
### print h2o.dump_json(kmeansResult)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
# cluster centers can return in any order
clusters = kmeansResult['KMeansModel']['clusters']
# the way we create the centers above, if we sort on the sum of xyz
# we should get the order the same as when they were created.
# to be safe, we'll sort the centers that were generated too, the same way
clustersSorted = sorted(clusters, key=sum)
centersSorted = sorted(centersList, key=sum)
### print clustersSorted
print "\ntrial #", trial, "h2o result, centers (sorted by key=sum)"
cf = '{0:6.2f}'
for c in clustersSorted:
print ' '.join(map(cf.format,c))
print "\ngenerated centers (sorted by key=sum)"
for c in centersSorted:
print ' '.join(map(cf.format,c))
for i,center in enumerate(centersSorted):
# Doing the compare of gen'ed/actual centers is kind of a hamming distance problem.
# Assuming that the difference between adjacent sums of all center values,
# is greater than 2x the sum of all max allowed variance on each value,
# Then the sums will be unique and non-overlapping with allowed variance.
# So a sort of the centers, keyed on sum of all values for a center.
# will create an ordering that can be compared.
# sort gen'ed and actual separately.
# Adjacent center hamming distance check is done during gen above.
a = center
b = clustersSorted[i]
print "\nexpected:", a
print "h2o:", b # h2o result
aStr = ",".join(map(str,a))
bStr = ",".join(map(str,b))
iStr = str(i)
for i, v in enumerate(a):
emsg = aStr+" != "+bStr+". Sorted cluster center "+iStr+" axis "+str(i)+" not correct."
self.assertAlmostEqual(a[i], b[i], delta=ALLOWED_CENTER_DELTA, msg=emsg)
print "Trial #", trial, "completed"
def test_kmeans2_sphere100(self):
SYNDATASETS_DIR = h2o.make_syn_dir()
csvFilename = 'syn_spheres100.csv'
csvPathname = SYNDATASETS_DIR + '/' + csvFilename
centersList = write_spheres_dataset(csvPathname, CLUSTERS, SPHERE_PTS)
if SHUFFLE_SPHERES:
# since we create spheres in order
csvFilename2 = 'syn_spheres100_shuffled.csv'
csvPathname2 = SYNDATASETS_DIR + '/' + csvFilename2
h2o_util.file_shuffle(csvPathname, csvPathname2)
else:
csvFilename2 = csvFilename
csvPathname2 = csvPathname
print "\nStarting", csvFilename
parseResult = h2i.import_parse(path=csvPathname2, schema='put', hex_key=csvFilename2 + ".hex")
### h2b.browseTheCloud()
# try 5 times, to see if all inits by h2o are good
# does it break if cols is not specified?
cols = ",".join(map(str,range(DIMENSIONS)))
for trial in range(10):
kwargs = {
'k': CLUSTERS,
'initialization': 'Furthest',
'destination_key': 'syn_spheres100.hex',
'max_iter': 15,
}
timeoutSecs = 100
start = time.time()
kmeansResult = h2o_cmd.runKMeans(parseResult=parseResult, timeoutSecs=timeoutSecs, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.',\
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
# can't inspect a kmeans2 model
# kmeansResult = h2o_cmd.runInspect(key='syn_spheres100.hex')
# print h2o.dump_json(kmeansResult)
### print h2o.dump_json(kmeans)
### print h2o.dump_json(kmeansResult)
h2o_kmeans.simpleCheckKMeans(self, kmeansResult, **kwargs)
# cluster centers can return in any order
model = kmeansResult['model']
clusters = model["centers"]
cluster_variances = model["within_cluster_variances"]
error = model["total_within_SS"]
iterations = model["iterations"]
normalized = model["normalized"]
max_iter = model["max_iter"]
# the way we create the centers above, if we sort on the sum of xyz
# we should get the order the same as when they were created.
# to be safe, we'll sort the centers that were generated too, the same way
clustersSorted = sorted(clusters, key=sum)
centersSorted = sorted(centersList, key=sum)
### print clustersSorted
print "\ntrial #", trial, "h2o result, centers (sorted by key=sum)"
cf = '{0:6.2f}'
for c in clustersSorted:
print ' '.join(map(cf.format,c))
print "\ngenerated centers (sorted by key=sum)"
for c in centersSorted:
print ' '.join(map(cf.format,c))
for i,center in enumerate(centersSorted):
# Doing the compare of gen'ed/actual centers is kind of a hamming distance problem.
# Assuming that the difference between adjacent sums of all center values,
# is greater than 2x the sum of all max allowed variance on each value,
# Then the sums will be unique and non-overlapping with allowed variance.
# So a sort of the centers, keyed on sum of all values for a center.
# will create an ordering that can be compared.
# sort gen'ed and actual separately.
# Adjacent center hamming distance check is done during gen above.
a = center
b = clustersSorted[i]
print "\nexpected:", a
print "h2o:", b # h2o result
aStr = ",".join(map(str,a))
bStr = ",".join(map(str,b))
iStr = str(i)
for i, v in enumerate(a):
emsg = aStr+" != "+bStr+". Sorted cluster center "+iStr+" axis "+str(i)+" not correct."
self.assertAlmostEqual(a[i], b[i], delta=ALLOWED_CENTER_DELTA, msg=emsg)
print "Trial #", trial, "completed"
def test_KMeans_libsvm_fvec(self):
# hack this into a function so we can call it before and after kmeans
# kmeans is changing the last col to enum?? (and changing the data)
def do_summary_and_inspect():
# SUMMARY******************************************
summaryResult = h2o_cmd.runSummary(key=hex_key)
coltypeList = h2o_cmd.infoFromSummary(summaryResult)
# INSPECT******************************************
inspect = h2o_cmd.runInspect(None, parseResult['destination_key'], timeoutSecs=360)
h2o_cmd.infoFromInspect(inspect, csvFilename)
numRows = inspect['numRows']
numCols = inspect['numCols']
# Now check both inspect and summary
if csvFilename=='covtype.binary.svm':
for k in range(55):
naCnt = inspect['cols'][k]['naCnt']
self.assertEqual(0, naCnt, msg='col %s naCnt %d should be %s' % (k, naCnt, 0))
stype = inspect['cols'][k]['type']
print k, stype
self.assertEqual('Int', stype, msg='col %s type %s should be %s' % (k, stype, 'Int'))
# summary may report type differently than inspect..check it too!
# we could check na here too
for i,c in enumerate(coltypeList):
print "column index: %s column type: %s" % (i, c)
# inspect says 'int?"
assert c=='Numeric', "All cols in covtype.binary.svm should be parsed as Numeric! %s %s" % (i,c)
# just do the import folder once
# make the timeout variable per dataset. it can be 10 secs for covtype 20x (col key creation)
# so probably 10x that for covtype200
csvFilenameList = [
# FIX! fails KMeansScore
("colon-cancer.svm", "cA", 30, 1),
("connect4.svm", "cB", 30, 1),
("covtype.binary.svm", "cC", 30, 1),
# multi-label class
# ("tmc2007_train.svm", "cJ", 30, 1),
("mnist_train.svm", "cM", 30, 1),
("duke.svm", "cD", 30, 1),
# too many features? 150K inspect timeout?
# ("E2006.train.svm", "cE", 30, 1),
("gisette_scale.svm", "cF", 120, 1), #Summary2 is slow with 5001 columns
("mushrooms.svm", "cG", 30, 1),
# ("news20.svm", "cH", 120, 1), #Summary2 is very slow - disable for now
("syn_6_1000_10.svm", "cK", 30, 1),
("syn_0_100_1000.svm", "cL", 30, 1),
]
csvFilenameList = [
("covtype.binary.svm", "cC", 30, 1),
]
### csvFilenameList = random.sample(csvFilenameAll,1)
h2b.browseTheCloud()
lenNodes = len(h2o.nodes)
firstDone = False
importFolderPath = "libsvm"
for (csvFilename, hex_key, timeoutSecs, resultMult) in csvFilenameList:
# have to import each time, because h2o deletes source after parse
csvPathname = importFolderPath + "/" + csvFilename
# PARSE******************************************
# creates csvFilename.hex from file in importFolder dir
parseResult = h2i.import_parse(bucket='home-0xdiag-datasets', path=csvPathname,
hex_key=hex_key, timeoutSecs=2000, doSummary=False)
do_summary_and_inspect()
# KMEANS******************************************
for trial in range(1):
kwargs = {
'k': 3,
'initialization': 'Furthest',
'ignored_cols': None, #range(11, numCols), # THIS BREAKS THE REST API
'max_iter': 10,
# 'normalize': 0,
# reuse the same seed, to get deterministic results (otherwise sometimes fails
'seed': 265211114317615310,
}
# fails if I put this in kwargs..i.e. source = dest
# 'destination_key': parseResult['destination_key'],
timeoutSecs = 600
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult, timeoutSecs=timeoutSecs, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
do_summary_and_inspect()
# this does an inspect of the model and prints the clusters
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
print "hello"
(centers, tupleResultList) = h2o_kmeans.bigCheckResults(self, kmeans, csvPathname, parseResult, 'd', **kwargs)
do_summary_and_inspect()
def test_KMeans_covtype_fvec(self):
csvFilenameList = [
('covtype.data', 800),
]
importFolderPath = "standard"
for csvFilename, timeoutSecs in csvFilenameList:
# creates csvFilename.hex from file in importFolder dir
csvPathname = importFolderPath + "/" + csvFilename
parseResult = h2i.import_parse(bucket='home-0xdiag-datasets',
path=csvPathname,
timeoutSecs=2000,
pollTimeoutSecs=60)
inspect = h2o_cmd.runInspect(None, parseResult['destination_key'])
print "\n" + csvPathname, \
" numRows:", "{:,}".format(inspect['numRows']), \
" numCols:", "{:,}".format(inspect['numCols'])
for trial in range(2):
kwargs = {
'k': 6,
'initialization': 'Furthest',
# 'initialization': '',
# 'ignored_cols': range(11, inspect['numCols']),
# ignore the response
'ignored_cols_by_name': 'C55',
'max_iter': 100,
# 'normalize': 0,
# reuse the same seed, to get deterministic results
'seed': 265211114317615310
}
start = time.time()
kmeansResult = h2o_cmd.runKMeans(parseResult=parseResult, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeansResult, **kwargs)
expected = [
([
2781.64184460309, 162.69950733599902,
16.545275983574268, 243.73547234768156,
50.48239522121315, 942.4480922085701,
208.3915356763203, 218.7135425941215,
140.10956243018794, 1040.6795741397266,
0.22024185323685105, 0.0845245225799837,
0.4957505706376572, 0.19948305354550802,
0.01635558145683929, 0.033196811983660604,
0.026025394050259283, 0.04566180477986607,
0.008617572941792261, 0.03547936261257615, 0.0, 0.0,
0.006189327591882107, 0.13606268110663236,
0.037222303163733886, 0.024007252359445064,
0.040891651692487006, 0.003232264365769295,
1.6188302332734367e-05, 0.004667627172605076,
0.00910861811255187, 9.173371321882807e-05,
0.0025415634662392956, 0.008946735089224526,
0.0023095311328034363, 0.04957397784361021,
0.09252154393235448, 0.03887890610245037, 0.0, 0.0,
0.0010792201555156243, 0.004867282901375466,
0.08281935473426902, 0.045640220376755754,
0.04933654940939677, 0.08426550974265995,
0.07772003949945769, 0.001327440791284218,
0.0014191745045030462, 0.0, 0.0, 0.009513325670870229,
0.010970272880816322, 0.009443176360761713
], 185319, 116283720155.37769),
([
2892.8730376693256, 119.94759695676377,
11.22516236778623, 189.0301354611245,
24.621525329374652, 2631.9842642419744,
219.94967526442753, 223.3794395991835,
135.71226572647987, 5409.1797365002785,
0.883243644460939, 0.11675635553906105, 0.0, 0.0, 0.0,
0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0015587307478196325,
0.0, 0.0, 0.0, 0.23410651326776769, 0.0, 0.0, 0.0,
0.026498422712933754, 0.0, 0.04152904063833735,
0.005158656522545927, 0.0695490814622379, 0.0,
0.0634997216552236, 0.05418444980515866,
0.010391538318797551, 0.0002969010948227871, 0.0, 0.0,
0.0, 0.3677862312117276, 0.07596956763778066, 0.0,
0.01109667841900167, 0.005641120801632956, 0.0,
0.0018185192057895714, 0.0, 0.0, 0.0021154203006123586,
0.018444980515865652, 0.010354425681944703
], 26945, 46932273891.61873),
([
3022.020861415003, 137.8546989122598, 13.3449108178427,
282.99227296949937, 45.23691263596753,
1606.0215197015768, 216.64941537882825,
222.64791856054669, 137.40339644525253,
2529.4366555907336, 0.4113429046111407,
0.08617284724616782, 0.5024842481426914, 0.0, 0.0,
0.0052506191028494405, 0.0, 0.014176671577693489, 0.0,
0.0, 0.0, 0.0, 0.0, 0.018949249239835743,
0.029850161436945546, 0.05403435628977148,
0.020892761982382997, 0.0, 0.0, 0.0018494718033917432,
0.011731607159650168, 0.005979436381304661,
0.0047098837027052445, 0.013714303626845553,
0.0007601642581737249, 0.047788470580859534,
0.10631328171530674, 0.04641704021817498,
0.0036519231372057308, 0.011872668568383437, 0.0,
0.00034481677690354536, 0.17267483777937995,
0.044473527475627724, 0.05637754302372967,
0.1292435973793925, 0.11970627880003762,
0.0013871038525438075, 0.004858781856368139, 0.0, 0.0,
0.03151155136202627, 0.028988119494686687,
0.012491771417823892
], 127604, 95229063588.02844),
([
3051.365089986695, 168.1268450579292,
14.114846831985933, 287.6101588092033,
50.702549817536706, 2835.266162979793,
209.89460702308608, 226.92302305495684,
148.84282479633362, 1461.8985753079312,
0.3284728328107128, 0.0006069141527711857,
0.670920253036516, 0.0, 0.0, 0.0054700083256172235,
0.0, 0.01653452018767653, 0.0, 0.0, 0.0, 0.0, 0.0,
0.03886584862938554, 0.013250959002170886,
0.04277966681969203, 0.05480901656564399, 0.0, 0.0,
0.0010426473906581905, 0.0018440853103432178, 0.0,
0.0035014278044491476, 0.011671426014830491,
0.002435437561761296, 0.044405885511091744,
0.10662236712081483, 0.042756323967662366, 0.0,
0.007384122192049426, 0.006263665294625696, 0.0,
0.14390868276285998, 0.022152366576148275,
0.07071327974851968, 0.14799368186805065,
0.1011367968938445, 0.009111493242244337,
0.006427065258833325, 0.0009259331305098857,
0.002318723301612991, 0.03055579330682623,
0.041044514818820564, 0.024074261393257027
], 128519, 106432862495.53804),
([
3052.088693852026, 149.15056174929376,
11.549996765359152, 328.4748452763461,
44.2420589567205, 4786.68757682272, 215.8348392383499,
226.91413106764713, 143.9780260065124,
4192.589071226791, 0.8949819938326181, 0.0,
0.10501800616738188, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0,
0.0022642485929312314, 0.002415198499126647, 0.0,
0.00012938563388178466, 0.0, 0.1351648588618377, 0.0,
0.0, 0.0, 0.014836219351777974, 0.0, 0.0,
0.010674314795247235, 0.03553792077286352, 0.0,
0.039290104155435275, 0.09289888512712138,
0.03864317598602636, 0.0, 0.0, 0.0, 0.0,
0.4371509283419232, 0.08636491061609126,
0.0003665926293317232, 0.002717098311517478,
0.017100467944709204, 0.0, 0.0028249196730856323, 0.0,
0.0, 0.03226015138119164, 0.017316110667845514,
0.03204450865805533
], 46373, 77991941653.19676),
([
3119.4885286481917, 165.13178470083923,
11.672206122079334, 271.2690333876713,
39.407851838435064, 4959.81440560285,
212.5861709835175, 227.95909557447322,
148.6725381875264, 1613.4457676749382,
0.9052556903942522, 0.0, 0.09474430960574776, 0.0, 0.0,
0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.00037734709895550323,
0.0, 0.0, 0.0, 0.008346917828895732,
0.0021584254060254783, 0.0, 0.0, 0.0031395278633097865,
0.0, 0.0, 0.02815009358208054, 0.012512829801364487,
0.0, 0.13355068526233171, 0.11424560767976816,
0.008799734347642335, 0.0, 0.0018867354947775161,
0.0012226046006158305, 0.0, 0.44056028497252914,
0.10774014369377528, 0.0033810300066413087,
0.014580691903640641, 0.02313892410795146,
0.0002565960272897422, 3.018776791644026e-05, 0.0, 0.0,
0.06503954597597053, 0.022625732053371973,
0.008256354525146411
], 66252, 74666940350.2879),
]
### print h2o.dump_json(kmeans)
predictKey = 'd'
(centers, tupleResultList) = h2o_kmeans.bigCheckResults(
self, kmeansResult, csvPathname, parseResult, predictKey,
**kwargs)
# all are multipliers of expected tuple value
allowedDelta = (0.01, 0.01, 0.01)
# these clusters were sorted compared to the cluster order in training
h2o_kmeans.showClusterDistribution(self,
tupleResultList,
expected,
trial=trial)
# why is the expected # of rows not right in KMeans2. That means predictions are wrong
h2o_kmeans.compareResultsToExpected(self,
tupleResultList,
expected,
allowedDelta,
allowError=False,
allowRowError=True,
trial=trial)
print "Trial #", trial, "completed\n"
def test_KMeans_libsvm_fvec(self):
# hack this into a function so we can call it before and after kmeans
# kmeans is changing the last col to enum?? (and changing the data)
def do_summary_and_inspect():
# SUMMARY******************************************
summaryResult = h2o_cmd.runSummary(key=hex_key)
coltypeList = h2o_cmd.infoFromSummary(summaryResult)
# INSPECT******************************************
inspect = h2o_cmd.runInspect(None,
parseResult['destination_key'],
timeoutSecs=360)
h2o_cmd.infoFromInspect(inspect, csvFilename)
numRows = inspect['numRows']
numCols = inspect['numCols']
# Now check both inspect and summary
if csvFilename == 'covtype.binary.svm':
for k in range(55):
naCnt = inspect['cols'][k]['naCnt']
self.assertEqual(0,
naCnt,
msg='col %s naCnt %d should be %s' %
(k, naCnt, 0))
stype = inspect['cols'][k]['type']
print k, stype
self.assertEqual('Int',
stype,
msg='col %s type %s should be %s' %
(k, stype, 'Int'))
# summary may report type differently than inspect..check it too!
# we could check na here too
for i, c in enumerate(coltypeList):
print "column index: %s column type: %s" % (i, c)
# inspect says 'int?"
assert c == 'Numeric', "All cols in covtype.binary.svm should be parsed as Numeric! %s %s" % (
i, c)
# just do the import folder once
# make the timeout variable per dataset. it can be 10 secs for covtype 20x (col key creation)
# so probably 10x that for covtype200
csvFilenameList = [
# FIX! fails KMeansScore
("colon-cancer.svm", "cA", 30, 1),
("connect4.svm", "cB", 30, 1),
("covtype.binary.svm", "cC", 30, 1),
# multi-label class
# ("tmc2007_train.svm", "cJ", 30, 1),
("mnist_train.svm", "cM", 30, 1),
("duke.svm", "cD", 30, 1),
# too many features? 150K inspect timeout?
# ("E2006.train.svm", "cE", 30, 1),
("gisette_scale.svm", "cF", 120, 1
), #Summary2 is slow with 5001 columns
("mushrooms.svm", "cG", 30, 1),
# ("news20.svm", "cH", 120, 1), #Summary2 is very slow - disable for now
("syn_6_1000_10.svm", "cK", 30, 1),
("syn_0_100_1000.svm", "cL", 30, 1),
]
csvFilenameList = [
("covtype.binary.svm", "cC", 30, 1),
]
### csvFilenameList = random.sample(csvFilenameAll,1)
h2b.browseTheCloud()
lenNodes = len(h2o.nodes)
firstDone = False
importFolderPath = "libsvm"
for (csvFilename, hex_key, timeoutSecs, resultMult) in csvFilenameList:
# have to import each time, because h2o deletes source after parse
csvPathname = importFolderPath + "/" + csvFilename
# PARSE******************************************
# creates csvFilename.hex from file in importFolder dir
parseResult = h2i.import_parse(bucket='home-0xdiag-datasets',
path=csvPathname,
hex_key=hex_key,
timeoutSecs=2000,
doSummary=False)
do_summary_and_inspect()
# KMEANS******************************************
for trial in range(1):
kwargs = {
'k': 3,
'initialization': 'Furthest',
'ignored_cols':
None, #range(11, numCols), # THIS BREAKS THE REST API
'max_iter': 10,
# 'normalize': 0,
# reuse the same seed, to get deterministic results (otherwise sometimes fails
'seed': 265211114317615310,
}
# fails if I put this in kwargs..i.e. source = dest
# 'destination_key': parseResult['destination_key'],
timeoutSecs = 600
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult,
timeoutSecs=timeoutSecs,
**kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
do_summary_and_inspect()
# this does an inspect of the model and prints the clusters
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
print "hello"
(centers, tupleResultList) = h2o_kmeans.bigCheckResults(
self, kmeans, csvPathname, parseResult, 'd', **kwargs)
do_summary_and_inspect()
def test_kmeans_sphere5(self):
SYNDATASETS_DIR = h2o.make_syn_dir()
CLUSTERS = 5
SPHERE_PTS = 10000
csvFilename = 'syn_spheres100.csv'
csvPathname = SYNDATASETS_DIR + '/' + csvFilename
centersList = write_spheres_dataset(csvPathname, CLUSTERS, SPHERE_PTS)
print "\nStarting", csvFilename
parseResult = h2i.import_parse(path=csvPathname,
schema='put',
hex_key=csvFilename + ".hex")
# try 5 times, to see if all inits by h2o are good
for trial in range(5):
# pass SEED so it's repeatable
kwargs = {
'k': CLUSTERS,
'max_iter': 10,
'initialization': 'Furthest',
'cols': None,
'destination_key': 'syn_spheres100.hex',
'seed': SEED
}
timeoutSecs = 30
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult,
timeoutSecs=timeoutSecs,
**kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.',\
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
kmeansResult = h2o_cmd.runInspect(key='syn_spheres100.hex')
### print h2o.dump_json(kmeans)
### print h2o.dump_json(kmeansResult)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
# cluster centers can return in any order
clusters = kmeansResult['KMeansModel']['clusters']
clustersSorted = sorted(clusters, key=itemgetter(0))
### print clustersSorted
print "\nh2o result, centers sorted"
print clustersSorted
print "\ngenerated centers"
print centersList
for i, center in enumerate(centersList):
a = center
b = clustersSorted[i]
print "\nexpected:", a
print "h2o:", b # h2o result
aStr = ",".join(map(str, a))
bStr = ",".join(map(str, b))
iStr = str(i)
self.assertAlmostEqual(a[0],
b[0],
delta=1,
msg=aStr + "!=" + bStr +
". Sorted cluster center " + iStr +
" x not correct.")
self.assertAlmostEqual(a[1],
b[1],
delta=1,
msg=aStr + "!=" + bStr +
". Sorted cluster center " + iStr +
" y not correct.")
self.assertAlmostEqual(a[2],
b[2],
delta=1,
msg=aStr + "!=" + bStr +
". Sorted cluster center " + iStr +
" z not correct.")
print "Trial #", trial, "completed"
start = time.time()
parseResult = h2i.import_parse(bucket=bucket, path=csvPathname, schema='s3n', hex_key=hex_key,
timeoutSecs=timeoutSecs, retryDelaySecs=10, pollTimeoutSecs=60))
elapsed = time.time() - start
print "parse end on ", hex_key, 'took', elapsed, 'seconds',\
"%d pct. of timeout" % ((elapsed*100)/timeoutSecs)
print "parse result:", parseResult['destination_key']
kwargs = {
'cols': None,
'initialization': 'Furthest',
'k': 12
}
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=120, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvFilename, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
### print h2o.dump_json(kmeans)
inspect = h2o_cmd.runInspect(None,key=kmeans['destination_key'])
print h2o.dump_json(inspect)
print "Trial #", trial, "completed in", time.time() - trialStart, "seconds.", \
if __name__ == '__main__':
h2o.unit_main()
def test_KMeans2_sphere5_inits(self):
SYNDATASETS_DIR = h2o.make_syn_dir()
CLUSTERS = 5
SPHERE_PTS = 10000
csvFilename = 'syn_spheres100.csv'
csvPathname = SYNDATASETS_DIR + '/' + csvFilename
expectedCenters = write_spheres_dataset(csvPathname, CLUSTERS,
SPHERE_PTS)
print "\nStarting", csvFilename
parseResult = h2i.import_parse(path=csvPathname,
schema='put',
hex_key=csvFilename + ".hex")
# try 5 times, to see if all inits by h2o are good
savedResults = []
Result = collections.namedtuple(
'Result',
'trial clusters size cluster_variances error iterations normalized max_iter clustersSorted'
)
# save the best for comparison. Print messages when we update best
sameAsBest = 0
# big number? to init
bestResult = Result(None, None, None, None, None, None, None, None,
None)
for trial in range(TRIALS):
# pass SEED so it's repeatable
kwargs = {
'normalize': 0,
'k': CLUSTERS,
'max_iter': MAX_ITER,
'initialization': INIT,
# 'initialization': 'PlusPlus',
'destination_key': 'syn_spheres100.hex',
'seed': SEED
}
timeoutSecs = 30
start = time.time()
kmeansResult = h2o_cmd.runKMeans(parseResult=parseResult,
timeoutSecs=timeoutSecs,
**kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.',\
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
# see if we took the full limit to get an answer
# inspect of model doesn't work
# kmeansResult = h2o_cmd.runInspect(key='syn_spheres100.hex')
### print h2o.dump_json(kmeans)
### print h2o.dump_json(kmeansResult)
h2o_kmeans.simpleCheckKMeans(self, kmeansResult, **kwargs)
model = kmeansResult['model']
clusters = model["centers"]
size = model["size"]
cluster_variances = model["within_cluster_variances"]
# round to int to avoid fp error when saying "same"
error = int(model["total_within_SS"])
iterations = model["iterations"]
normalized = model["normalized"]
max_iter = model["max_iter"]
# clustersSorted = sorted(clusters, key=itemgetter(2))
clustersSorted = sorted(clusters)
r = Result(
trial,
clusters,
size,
cluster_variances,
error,
iterations,
normalized,
max_iter,
clustersSorted,
)
savedResults.append(r)
if iterations >= (
max_iter -
1): # h2o hits the limit at max_iter-1..shouldn't hit it
raise Exception(
"KMeans unexpectedly took %s iterations..which was the full amount allowed by max_iter %s",
(iterations, max_iter))
print "iterations", iterations
### print clustersSorted
# For now, just analyze the one with the lowest error
# we could analyze how many are not best, and how many are best (maybe just look at error
print "savedResults, error"
print r.error
if bestResult.error and r.error <= bestResult.error:
sameAsBest += 1
# we can check that if it has the same error, the sizes should be the same (integer) and reflects centers?
# should
if r.size != bestResult.size:
raise Exception(
"Would expect that if two trials got the same error (rounded to int), the cluster sizes would likely be the same? %s %s"
% (r.size, bestResult.size))
if not bestResult.error: # init case
bestResult = r
elif r.error < bestResult.error:
print "Trial", r.trial, "has a lower error", r.error, "than current lowest error", bestResult.error
print "Using it for best now"
bestResult = r
print "Trial #", trial, "completed"
print "\nApparently, %s out of %s trials, got the same best error: %s (lowest) " % (
sameAsBest, TRIALS, bestResult.error)
print "\nh2o best result was from trial %s, centers sorted:" % bestResult.trial
print bestResult.clustersSorted
print "\ngenerated centers for comparison"
print expectedCenters
for i, center in enumerate(expectedCenters):
a = center
bb = bestResult.clustersSorted
print "bb:", bb
b = bb[i]
print "\nexpected:", a
print "h2o:", b # h2o result
aStr = ",".join(map(str, a))
bStr = ",".join(map(str, b))
iStr = str(i)
self.assertAlmostEqual(a[0],
b[0],
delta=1,
msg=aStr + "!=" + bStr +
". Sorted cluster center " + iStr +
"; x not correct.")
self.assertAlmostEqual(a[1],
b[1],
delta=1,
msg=aStr + "!=" + bStr +
". Sorted cluster center " + iStr +
"; y not correct.")
self.assertAlmostEqual(a[2],
b[2],
delta=1,
msg=aStr + "!=" + bStr +
". Sorted cluster center " + iStr +
"; z not correct.")
def test_KMeans2_sphere5_bad_inits(self):
SYNDATASETS_DIR = h2o.make_syn_dir()
csvFilename = 'syn_spheres100.csv'
csvPathname = SYNDATASETS_DIR + '/' + csvFilename
expectedCenters = write_spheres_dataset(csvPathname, CLUSTERS, SPHERE_PTS)
print "\nStarting", csvFilename
parseResult = h2i.import_parse(path=csvPathname, schema='put', hex_key=csvFilename + ".hex")
# try 5 times, to see if all inits by h2o are good
savedResults = []
Result = collections.namedtuple('Result',
'trial clusters size cluster_variances error iterations normalized max_iter clustersSorted')
# save the best for comparison. Print messages when we update best
sameAsBest = 1
# big number? to init
bestResult = Result(None, None, None, None, None, None, None, None, None)
for trial in range(TRIALS):
# pass SEED so it's repeatable
kwargs = {
'normalize': 0,
'k': CLUSTERS,
'max_iter': MAX_ITER,
'initialization': INIT,
# 'initialization': 'PlusPlus',
'destination_key': 'syn_spheres100.hex',
'seed': SEED
}
timeoutSecs = 30
start = time.time()
kmeansResult = h2o_cmd.runKMeans(parseResult=parseResult, timeoutSecs=timeoutSecs, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.',\
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
# see if we took the full limit to get an answer
# inspect of model doesn't work
# kmeansResult = h2o_cmd.runInspect(key='syn_spheres100.hex')
### print h2o.dump_json(kmeans)
### print h2o.dump_json(kmeansResult)
h2o_kmeans.simpleCheckKMeans(self, kmeansResult, **kwargs)
model = kmeansResult['model']
clusters = model["centers"]
size = model["size"]
cluster_variances = model["within_cluster_variances"]
# round to int to avoid fp error when saying "same"
error = int(model["total_within_SS"])
iterations = model["iterations"]
normalized = model["normalized"]
max_iter = model["max_iter"]
# clustersSorted = sorted(clusters, key=itemgetter(0))
clustersSorted = sorted(clusters)
r = Result (
trial,
clusters,
size,
cluster_variances,
error,
iterations,
normalized,
max_iter,
clustersSorted,
)
savedResults.append(r)
if iterations >= (max_iter-1): # h2o hits the limit at max_iter-1..shouldn't hit it
raise Exception("KMeans unexpectedly took %s iterations..which was the full amount allowed by max_iter %s",
(iterations, max_iter))
print "iterations", iterations
### print clustersSorted
# For now, just analyze the one with the lowest error
# we could analyze how many are not best, and how many are best (maybe just look at error
print "savedResults, error"
print r.error
if bestResult.error and r.error <= bestResult.error:
sameAsBest += 1
# we can check that if it has the same error, the sizes should be the same (integer) and reflects centers?
# should
if sorted(r.size)!=sorted(bestResult.size):
raise Exception("Would expect that if two trials got the same error (rounded to int), the cluster sizes would likely be the same? %s %s" %
(r.size, bestResult.size))
if not bestResult.error: # init case
bestResult = r
elif r.error < bestResult.error:
print "Trial", r.trial, "has a lower error", r.error, "than current lowest error", bestResult.error
print "Using it for best now"
bestResult = r
print "Trial #", trial, "completed"
print "\nApparently, %s out of %s trials, got the same best error: %s (lowest) " % (sameAsBest, TRIALS, bestResult.error)
print "\nh2o best result was from trial %s, centers sorted:" % bestResult.trial
print bestResult.clustersSorted
print "\ngenerated centers for comparison"
print expectedCenters
for i,center in enumerate(expectedCenters):
a = center
bb = bestResult.clustersSorted
print "bb:", bb
b = bb[i]
print "\nexpected:", a
print "h2o:", b # h2o result
aStr = ",".join(map(str,a))
bStr = ",".join(map(str,b))
iStr = str(i)
self.assertAlmostEqual(a[0], b[0], delta=2, msg=aStr+"!="+bStr+". Sorted cluster center "+iStr+"; x not correct.")
self.assertAlmostEqual(a[1], b[1], delta=2, msg=aStr+"!="+bStr+". Sorted cluster center "+iStr+"; y not correct.")
self.assertAlmostEqual(a[2], b[2], delta=2, msg=aStr+"!="+bStr+". Sorted cluster center "+iStr+"; z not correct.")
def test_KMeans_covtype_fvec(self):
csvFilenameList = [
('covtype.data', 800),
]
importFolderPath = "standard"
for csvFilename, timeoutSecs in csvFilenameList:
# creates csvFilename.hex from file in importFolder dir
csvPathname = importFolderPath + "/" + csvFilename
parseResult = h2i.import_parse(bucket='home-0xdiag-datasets', path=csvPathname,
timeoutSecs=2000, pollTimeoutSecs=60)
inspect = h2o_cmd.runInspect(None, parseResult['destination_key'])
print "\n" + csvPathname, \
" numRows:", "{:,}".format(inspect['numRows']), \
" numCols:", "{:,}".format(inspect['numCols'])
for trial in range(2):
kwargs = {
'k': 6,
'initialization': 'Furthest',
# 'initialization': '',
# 'ignored_cols': range(11, inspect['numCols']),
# ignore the response
'ignored_cols_by_name': 'C55',
'max_iter': 100,
# 'normalize': 0,
# reuse the same seed, to get deterministic results
'seed': 265211114317615310
}
start = time.time()
kmeansResult = h2o_cmd.runKMeans(parseResult=parseResult, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeansResult, **kwargs)
expected = [
([2781.64184460309, 162.69950733599902, 16.545275983574268, 243.73547234768156, 50.48239522121315, 942.4480922085701, 208.3915356763203, 218.7135425941215, 140.10956243018794, 1040.6795741397266, 0.22024185323685105, 0.0845245225799837, 0.4957505706376572, 0.19948305354550802, 0.01635558145683929, 0.033196811983660604, 0.026025394050259283, 0.04566180477986607, 0.008617572941792261, 0.03547936261257615, 0.0, 0.0, 0.006189327591882107, 0.13606268110663236, 0.037222303163733886, 0.024007252359445064, 0.040891651692487006, 0.003232264365769295, 1.6188302332734367e-05, 0.004667627172605076, 0.00910861811255187, 9.173371321882807e-05, 0.0025415634662392956, 0.008946735089224526, 0.0023095311328034363, 0.04957397784361021, 0.09252154393235448, 0.03887890610245037, 0.0, 0.0, 0.0010792201555156243, 0.004867282901375466, 0.08281935473426902, 0.045640220376755754, 0.04933654940939677, 0.08426550974265995, 0.07772003949945769, 0.001327440791284218, 0.0014191745045030462, 0.0, 0.0, 0.009513325670870229, 0.010970272880816322, 0.009443176360761713], 185319, 116283720155.37769) ,
([2892.8730376693256, 119.94759695676377, 11.22516236778623, 189.0301354611245, 24.621525329374652, 2631.9842642419744, 219.94967526442753, 223.3794395991835, 135.71226572647987, 5409.1797365002785, 0.883243644460939, 0.11675635553906105, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0015587307478196325, 0.0, 0.0, 0.0, 0.23410651326776769, 0.0, 0.0, 0.0, 0.026498422712933754, 0.0, 0.04152904063833735, 0.005158656522545927, 0.0695490814622379, 0.0, 0.0634997216552236, 0.05418444980515866, 0.010391538318797551, 0.0002969010948227871, 0.0, 0.0, 0.0, 0.3677862312117276, 0.07596956763778066, 0.0, 0.01109667841900167, 0.005641120801632956, 0.0, 0.0018185192057895714, 0.0, 0.0, 0.0021154203006123586, 0.018444980515865652, 0.010354425681944703], 26945, 46932273891.61873) ,
([3022.020861415003, 137.8546989122598, 13.3449108178427, 282.99227296949937, 45.23691263596753, 1606.0215197015768, 216.64941537882825, 222.64791856054669, 137.40339644525253, 2529.4366555907336, 0.4113429046111407, 0.08617284724616782, 0.5024842481426914, 0.0, 0.0, 0.0052506191028494405, 0.0, 0.014176671577693489, 0.0, 0.0, 0.0, 0.0, 0.0, 0.018949249239835743, 0.029850161436945546, 0.05403435628977148, 0.020892761982382997, 0.0, 0.0, 0.0018494718033917432, 0.011731607159650168, 0.005979436381304661, 0.0047098837027052445, 0.013714303626845553, 0.0007601642581737249, 0.047788470580859534, 0.10631328171530674, 0.04641704021817498, 0.0036519231372057308, 0.011872668568383437, 0.0, 0.00034481677690354536, 0.17267483777937995, 0.044473527475627724, 0.05637754302372967, 0.1292435973793925, 0.11970627880003762, 0.0013871038525438075, 0.004858781856368139, 0.0, 0.0, 0.03151155136202627, 0.028988119494686687, 0.012491771417823892], 127604, 95229063588.02844) ,
([3051.365089986695, 168.1268450579292, 14.114846831985933, 287.6101588092033, 50.702549817536706, 2835.266162979793, 209.89460702308608, 226.92302305495684, 148.84282479633362, 1461.8985753079312, 0.3284728328107128, 0.0006069141527711857, 0.670920253036516, 0.0, 0.0, 0.0054700083256172235, 0.0, 0.01653452018767653, 0.0, 0.0, 0.0, 0.0, 0.0, 0.03886584862938554, 0.013250959002170886, 0.04277966681969203, 0.05480901656564399, 0.0, 0.0, 0.0010426473906581905, 0.0018440853103432178, 0.0, 0.0035014278044491476, 0.011671426014830491, 0.002435437561761296, 0.044405885511091744, 0.10662236712081483, 0.042756323967662366, 0.0, 0.007384122192049426, 0.006263665294625696, 0.0, 0.14390868276285998, 0.022152366576148275, 0.07071327974851968, 0.14799368186805065, 0.1011367968938445, 0.009111493242244337, 0.006427065258833325, 0.0009259331305098857, 0.002318723301612991, 0.03055579330682623, 0.041044514818820564, 0.024074261393257027], 128519, 106432862495.53804) ,
([3052.088693852026, 149.15056174929376, 11.549996765359152, 328.4748452763461, 44.2420589567205, 4786.68757682272, 215.8348392383499, 226.91413106764713, 143.9780260065124, 4192.589071226791, 0.8949819938326181, 0.0, 0.10501800616738188, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0022642485929312314, 0.002415198499126647, 0.0, 0.00012938563388178466, 0.0, 0.1351648588618377, 0.0, 0.0, 0.0, 0.014836219351777974, 0.0, 0.0, 0.010674314795247235, 0.03553792077286352, 0.0, 0.039290104155435275, 0.09289888512712138, 0.03864317598602636, 0.0, 0.0, 0.0, 0.0, 0.4371509283419232, 0.08636491061609126, 0.0003665926293317232, 0.002717098311517478, 0.017100467944709204, 0.0, 0.0028249196730856323, 0.0, 0.0, 0.03226015138119164, 0.017316110667845514, 0.03204450865805533], 46373, 77991941653.19676) ,
([3119.4885286481917, 165.13178470083923, 11.672206122079334, 271.2690333876713, 39.407851838435064, 4959.81440560285, 212.5861709835175, 227.95909557447322, 148.6725381875264, 1613.4457676749382, 0.9052556903942522, 0.0, 0.09474430960574776, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.00037734709895550323, 0.0, 0.0, 0.0, 0.008346917828895732, 0.0021584254060254783, 0.0, 0.0, 0.0031395278633097865, 0.0, 0.0, 0.02815009358208054, 0.012512829801364487, 0.0, 0.13355068526233171, 0.11424560767976816, 0.008799734347642335, 0.0, 0.0018867354947775161, 0.0012226046006158305, 0.0, 0.44056028497252914, 0.10774014369377528, 0.0033810300066413087, 0.014580691903640641, 0.02313892410795146, 0.0002565960272897422, 3.018776791644026e-05, 0.0, 0.0, 0.06503954597597053, 0.022625732053371973, 0.008256354525146411], 66252, 74666940350.2879) ,
]
### print h2o.dump_json(kmeans)
predictKey = 'd'
(centers, tupleResultList) = h2o_kmeans.bigCheckResults(self, kmeansResult, csvPathname, parseResult, predictKey, **kwargs)
# all are multipliers of expected tuple value
allowedDelta = (0.01, 0.01, 0.01)
# these clusters were sorted compared to the cluster order in training
h2o_kmeans.showClusterDistribution(self, tupleResultList, expected, trial=trial)
# why is the expected # of rows not right in KMeans2. That means predictions are wrong
h2o_kmeans.compareResultsToExpected(self, tupleResultList, expected, allowedDelta, allowError=False,
allowRowError=True, trial=trial)
print "Trial #", trial, "completed\n"
def test_parse_bounds_libsvm(self):
# just do the import folder once
# make the timeout variable per dataset. it can be 10 secs for covtype 20x (col key creation)
# so probably 10x that for covtype200
csvFilenameList = [
# FIX! fails KMeansScore
("colon-cancer.svm", "cA", 30, 1),
("connect4.svm", "cB", 30, 1),
("covtype.binary.svm", "cC", 30, 1),
# multi-label class
# ("tmc2007_train.svm", "cJ", 30, 1),
("mnist_train.svm", "cM", 30, 1),
("duke.svm", "cD", 30, 1),
# too many features? 150K inspect timeout?
# ("E2006.train.svm", "cE", 30, 1),
("gisette_scale.svm", "cF", 30, 1),
("mushrooms.svm", "cG", 30, 1),
("news20.svm", "cH", 30, 1),
("syn_6_1000_10.svm", "cK", 30, 1),
("syn_0_100_1000.svm", "cL", 30, 1),
# normal csv
]
### csvFilenameList = random.sample(csvFilenameAll,1)
# h2b.browseTheCloud()
lenNodes = len(h2o.nodes)
firstDone = False
importFolderPath = "libsvm"
for (csvFilename, hex_key, timeoutSecs, resultMult) in csvFilenameList:
# have to import each time, because h2o deletes source after parse
csvPathname = importFolderPath + "/" + csvFilename
# PARSE******************************************
# creates csvFilename.hex from file in importFolder dir
parseResult = h2i.import_parse(bucket='home-0xdiag-datasets', path=csvPathname,
hex_key=hex_key, timeoutSecs=2000)
print csvPathname, 'parse time:', parseResult['response']['time']
print "Parse result['destination_key']:", parseResult['destination_key']
# INSPECT******************************************
start = time.time()
inspect = h2o_cmd.runInspect(None, parseResult['destination_key'], timeoutSecs=360)
print "Inspect:", parseResult['destination_key'], "took", time.time() - start, "seconds"
h2o_cmd.infoFromInspect(inspect, csvFilename)
# KMEANS******************************************
for trial in range(1):
kwargs = {
'k': 3,
'initialization': 'Furthest',
'cols': range(10),
# 'max_iter': 10,
# 'normalize': 0,
# reuse the same seed, to get deterministic results (otherwise sometimes fails
'seed': 265211114317615310,
}
# fails if I put this in kwargs..i.e. source = dest
# 'destination_key': parseResult['destination_key'],
timeoutSecs = 600
start = time.time()
kmeans = h2o_cmd.runKMeans(parseResult=parseResult, timeoutSecs=timeoutSecs, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
# this does an inspect of the model and prints the clusters
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
(centers, tupleResultList) = h2o_kmeans.bigCheckResults(self, kmeans, csvPathname, parseResult, 'd', **kwargs)
def test_KMeans_params_rand2(self):
SEED = random.randint(0, sys.maxint)
# if you have to force to redo a test
# SEED =
random.seed(SEED)
print "\nUsing random seed:", SEED
if localhost:
csvFilenameList = [
# ('covtype.data', 60),
('covtype20x.data', 400),
]
else:
csvFilenameList = [
('covtype20x.data', 400),
('covtype200x.data', 2000),
]
importFolderPath = '/home/0xdiag/datasets'
h2i.setupImportFolder(None, importFolderPath)
for csvFilename, timeoutSecs in csvFilenameList:
# creates csvFilename.hex from file in importFolder dir
parseKey = h2i.parseImportFolderFile(None,
csvFilename,
importFolderPath,
timeoutSecs=2000,
pollTimeoutSecs=60)
inspect = h2o_cmd.runInspect(None, parseKey['destination_key'])
csvPathname = importFolderPath + "/" + csvFilename
print "\n" + csvPathname, \
" num_rows:", "{:,}".format(inspect['num_rows']), \
" num_cols:", "{:,}".format(inspect['num_cols'])
paramDict = define_params()
for trial in range(3):
randomV = paramDict['k']
k = random.choice(randomV)
randomV = paramDict['epsilon']
epsilon = random.choice(randomV)
randomV = paramDict['cols']
cols = random.choice(randomV)
kwargs = {
'k': k,
'epsilon': epsilon,
'cols': cols,
'destination_key': csvFilename + "_" + str(trial) + '.hex'
}
start = time.time()
kmeans = h2o_cmd.runKMeansOnly(parseKey=parseKey, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=60, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvPathname, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
### print h2o.dump_json(kmeans)
inspect = h2o_cmd.runInspect(None,
key=kmeans['destination_key'])
print h2o.dump_json(inspect)
print "Trial #", trial, "completed\n"
def test_KMeans_allstate_s3n_thru_hdfs(self):
csvFilename = "CAT*"
URI = "s3n://home-0xdiag-datasets/cats"
s3nKey = URI + "/" + csvFilename
trialMax = 1
for trial in range(trialMax):
trialStart = time.time()
# since we delete the key, we have to re-import every iteration
# s3n URI thru HDFS is not typical.
importHDFSResult = h2o.nodes[0].import_hdfs(URI)
s3nFullList = importHDFSResult['succeeded']
### print "s3nFullList:", h2o.dump_json(s3nFullList)
self.assertGreater(len(s3nFullList),1,"Didn't see more than 1 files in s3n?")
storeView = h2o.nodes[0].store_view()
### print "storeView:", h2o.dump_json(storeView)
for s in storeView['keys']:
print "\nkey:", s['key']
if 'rows' in s:
print "rows:", s['rows'], "value_size_bytes:", s['value_size_bytes']
key2 = csvFilename + "_" + str(trial) + ".hex"
print "Loading s3n key: ", s3nKey, 'thru HDFS'
# ec2 is about 400 secs on four m2.4xlarge nodes
# should be less on more nodes?
timeoutSecs = 600
start = time.time()
parseKey = h2o.nodes[0].parse(s3nKey, key2,
timeoutSecs=timeoutSecs, retryDelaySecs=10, pollTimeoutSecs=60, noise=('JStack', None))
elapsed = time.time() - start
print s3nKey, 'h2o reported parse time:', parseKey['response']['time']
print "parse end on ", s3nKey, 'took', elapsed, 'seconds',\
"%d pct. of timeout" % ((elapsed*100)/timeoutSecs)
print "parse result:", parseKey['destination_key']
kwargs = {
'cols': None,
'epsilon': 1e-6,
'k': 12
}
start = time.time()
kmeans = h2o_cmd.runKMeansOnly(parseKey=parseKey, \
timeoutSecs=timeoutSecs, retryDelaySecs=2, pollTimeoutSecs=120, **kwargs)
elapsed = time.time() - start
print "kmeans end on ", csvFilename, 'took', elapsed, 'seconds.', \
"%d pct. of timeout" % ((elapsed/timeoutSecs) * 100)
h2o_kmeans.simpleCheckKMeans(self, kmeans, **kwargs)
### print h2o.dump_json(kmeans)
inspect = h2o_cmd.runInspect(None,key=kmeans['destination_key'])
print h2o.dump_json(inspect)
print "Deleting key in H2O so we get it from S3 (if ec2) or nfs again.", \
"Otherwise it would just parse the cached key."
storeView = h2o.nodes[0].store_view()
# pattern matching problem
# h2o removes key afte parse now
### print "Removing", s3nKey
### removeKeyResult = h2o.nodes[0].remove_key(key=s3nKey)
print "Trial #", trial, "completed in", time.time() - trialStart, "seconds.", \
