Skip to content

Commit affb5d9

Browse files
mroeschkejreback
mroeschke
authored and
jreback
committed
CI: Run ASV on Travis for failed benchmarks (pandas-dev#19236)
1 parent 31973f5 commit affb5d9

22 files changed

+199
-69
lines changed

‎.travis.yml

+8
Original file line numberDiff line numberDiff line change
@@ -73,6 +73,10 @@ matrix:
7373
env:
7474
- JOB="3.6_NUMPY_DEV" TEST_ARGS="--skip-slow --skip-network" PANDAS_TESTING_MODE="deprecate"
7575
# In allow_failures
76+
- dist: trusty
77+
env:
78+
- JOB="3.6_ASV" ASV=true
79+
# In allow_failures
7680
- dist: trusty
7781
env:
7882
- JOB="3.6_DOC" DOC=true
@@ -93,6 +97,9 @@ matrix:
9397
- dist: trusty
9498
env:
9599
- JOB="3.6_NUMPY_DEV" TEST_ARGS="--skip-slow --skip-network" PANDAS_TESTING_MODE="deprecate"
100+
- dist: trusty
101+
env:
102+
- JOB="3.6_ASV" ASV=true
96103
- dist: trusty
97104
env:
98105
- JOB="3.6_DOC" DOC=true
@@ -128,6 +135,7 @@ script:
128135
- ci/script_single.sh
129136
- ci/script_multi.sh
130137
- ci/lint.sh
138+
- ci/asv.sh
131139
- echo "checking imports"
132140
- source activate pandas && python ci/check_imports.py
133141
- echo "script done"

‎asv_bench/benchmarks/algorithms.py

+3-1
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,4 @@
1+
import warnings
12
from importlib import import_module
23

34
import numpy as np
@@ -83,7 +84,8 @@ def setup(self):
8384
self.all = self.uniques.repeat(10)
8485

8586
def time_match_string(self):
86-
pd.match(self.all, self.uniques)
87+
with warnings.catch_warnings(record=True):
88+
pd.match(self.all, self.uniques)
8789

8890

8991
class Hashing(object):

‎asv_bench/benchmarks/categoricals.py

+8-2
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,5 @@
1+
import warnings
2+
13
import numpy as np
24
import pandas as pd
35
import pandas.util.testing as tm
@@ -119,11 +121,15 @@ def setup(self):
119121

120122
self.s_str = pd.Series(tm.makeCategoricalIndex(N, ncats)).astype(str)
121123
self.s_str_cat = self.s_str.astype('category')
122-
self.s_str_cat_ordered = self.s_str.astype('category', ordered=True)
124+
with warnings.catch_warnings(record=True):
125+
self.s_str_cat_ordered = self.s_str.astype('category',
126+
ordered=True)
123127

124128
self.s_int = pd.Series(np.random.randint(0, ncats, size=N))
125129
self.s_int_cat = self.s_int.astype('category')
126-
self.s_int_cat_ordered = self.s_int.astype('category', ordered=True)
130+
with warnings.catch_warnings(record=True):
131+
self.s_int_cat_ordered = self.s_int.astype('category',
132+
ordered=True)
127133

128134
def time_rank_string(self):
129135
self.s_str.rank()

‎asv_bench/benchmarks/frame_methods.py

+6-3
Original file line numberDiff line numberDiff line change
@@ -1,4 +1,6 @@
11
import string
2+
import warnings
3+
24
import numpy as np
35
import pandas.util.testing as tm
46
from pandas import (DataFrame, Series, MultiIndex, date_range, period_range,
@@ -15,7 +17,8 @@ def setup(self):
1517
self.df = DataFrame(np.random.randn(10000, 25))
1618
self.df['foo'] = 'bar'
1719
self.df['bar'] = 'baz'
18-
self.df = self.df.consolidate()
20+
with warnings.catch_warnings(record=True):
21+
self.df = self.df.consolidate()
1922

2023
def time_frame_get_numeric_data(self):
2124
self.df._get_numeric_data()
@@ -141,8 +144,8 @@ class Repr(object):
141144
def setup(self):
142145
nrows = 10000
143146
data = np.random.randn(nrows, 10)
144-
idx = MultiIndex.from_arrays(np.tile(np.random.randn(3, nrows / 100),
145-
100))
147+
arrays = np.tile(np.random.randn(3, int(nrows / 100)), 100)
148+
idx = MultiIndex.from_arrays(arrays)
146149
self.df3 = DataFrame(data, index=idx)
147150
self.df4 = DataFrame(data, index=np.random.randn(nrows))
148151
self.df_tall = DataFrame(np.random.randn(nrows, 10))

‎asv_bench/benchmarks/gil.py

+20-23
Original file line numberDiff line numberDiff line change
@@ -1,9 +1,13 @@
11
import numpy as np
22
import pandas.util.testing as tm
3-
from pandas import (DataFrame, Series, rolling_median, rolling_mean,
4-
rolling_min, rolling_max, rolling_var, rolling_skew,
5-
rolling_kurt, rolling_std, read_csv, factorize, date_range)
3+
from pandas import DataFrame, Series, read_csv, factorize, date_range
64
from pandas.core.algorithms import take_1d
5+
try:
6+
from pandas import (rolling_median, rolling_mean, rolling_min, rolling_max,
7+
rolling_var, rolling_skew, rolling_kurt, rolling_std)
8+
have_rolling_methods = True
9+
except ImportError:
10+
have_rolling_methods = False
711
try:
812
from pandas._libs import algos
913
except ImportError:
@@ -171,8 +175,7 @@ def run(period):
171175
class ParallelRolling(object):
172176

173177
goal_time = 0.2
174-
params = ['rolling_median', 'rolling_mean', 'rolling_min', 'rolling_max',
175-
'rolling_var', 'rolling_skew', 'rolling_kurt', 'rolling_std']
178+
params = ['median', 'mean', 'min', 'max', 'var', 'skew', 'kurt', 'std']
176179
param_names = ['method']
177180

178181
def setup(self, method):
@@ -181,34 +184,28 @@ def setup(self, method):
181184
win = 100
182185
arr = np.random.rand(100000)
183186
if hasattr(DataFrame, 'rolling'):
184-
rolling = {'rolling_median': 'median',
185-
'rolling_mean': 'mean',
186-
'rolling_min': 'min',
187-
'rolling_max': 'max',
188-
'rolling_var': 'var',
189-
'rolling_skew': 'skew',
190-
'rolling_kurt': 'kurt',
191-
'rolling_std': 'std'}
192187
df = DataFrame(arr).rolling(win)
193188

194189
@test_parallel(num_threads=2)
195190
def parallel_rolling():
196-
getattr(df, rolling[method])()
191+
getattr(df, method)()
197192
self.parallel_rolling = parallel_rolling
198-
else:
199-
rolling = {'rolling_median': rolling_median,
200-
'rolling_mean': rolling_mean,
201-
'rolling_min': rolling_min,
202-
'rolling_max': rolling_max,
203-
'rolling_var': rolling_var,
204-
'rolling_skew': rolling_skew,
205-
'rolling_kurt': rolling_kurt,
206-
'rolling_std': rolling_std}
193+
elif have_rolling_methods:
194+
rolling = {'median': rolling_median,
195+
'mean': rolling_mean,
196+
'min': rolling_min,
197+
'max': rolling_max,
198+
'var': rolling_var,
199+
'skew': rolling_skew,
200+
'kurt': rolling_kurt,
201+
'std': rolling_std}
207202

208203
@test_parallel(num_threads=2)
209204
def parallel_rolling():
210205
rolling[method](arr, win)
211206
self.parallel_rolling = parallel_rolling
207+
else:
208+
raise NotImplementedError
212209

213210
def time_rolling(self, method):
214211
self.parallel_rolling()

‎asv_bench/benchmarks/groupby.py

+4-2
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,4 @@
1+
import warnings
12
from string import ascii_letters
23
from itertools import product
34
from functools import partial
@@ -340,7 +341,8 @@ def time_dt_size(self):
340341
self.df.groupby(['dates']).size()
341342

342343
def time_dt_timegrouper_size(self):
343-
self.df.groupby(TimeGrouper(key='dates', freq='M')).size()
344+
with warnings.catch_warnings(record=True):
345+
self.df.groupby(TimeGrouper(key='dates', freq='M')).size()
344346

345347
def time_category_size(self):
346348
self.draws.groupby(self.cats).size()
@@ -467,7 +469,7 @@ class SumMultiLevel(object):
467469

468470
def setup(self):
469471
N = 50
470-
self.df = DataFrame({'A': range(N) * 2,
472+
self.df = DataFrame({'A': list(range(N)) * 2,
471473
'B': range(N * 2),
472474
'C': 1}).set_index(['A', 'B'])
473475

‎asv_bench/benchmarks/indexing.py

+13-6
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,5 @@
1+
import warnings
2+
13
import numpy as np
24
import pandas.util.testing as tm
35
from pandas import (Series, DataFrame, MultiIndex, Int64Index, Float64Index,
@@ -91,7 +93,8 @@ def time_getitem_pos_slice(self, index):
9193
self.s[:80000]
9294

9395
def time_get_value(self, index):
94-
self.s.get_value(self.lbl)
96+
with warnings.catch_warnings(record=True):
97+
self.s.get_value(self.lbl)
9598

9699
def time_getitem_scalar(self, index):
97100
self.s[self.lbl]
@@ -112,7 +115,8 @@ def setup(self):
112115
self.bool_obj_indexer = self.bool_indexer.astype(object)
113116

114117
def time_get_value(self):
115-
self.df.get_value(self.idx_scalar, self.col_scalar)
118+
with warnings.catch_warnings(record=True):
119+
self.df.get_value(self.idx_scalar, self.col_scalar)
116120

117121
def time_ix(self):
118122
self.df.ix[self.idx_scalar, self.col_scalar]
@@ -231,11 +235,13 @@ class PanelIndexing(object):
231235
goal_time = 0.2
232236

233237
def setup(self):
234-
self.p = Panel(np.random.randn(100, 100, 100))
235-
self.inds = range(0, 100, 10)
238+
with warnings.catch_warnings(record=True):
239+
self.p = Panel(np.random.randn(100, 100, 100))
240+
self.inds = range(0, 100, 10)
236241

237242
def time_subset(self):
238-
self.p.ix[(self.inds, self.inds, self.inds)]
243+
with warnings.catch_warnings(record=True):
244+
self.p.ix[(self.inds, self.inds, self.inds)]
239245

240246

241247
class MethodLookup(object):
@@ -295,7 +301,8 @@ def setup(self):
295301
def time_insert(self):
296302
np.random.seed(1234)
297303
for i in range(100):
298-
self.df.insert(0, i, np.random.randn(self.N))
304+
self.df.insert(0, i, np.random.randn(self.N),
305+
allow_duplicates=True)
299306

300307
def time_assign_with_setitem(self):
301308
np.random.seed(1234)

‎asv_bench/benchmarks/io/hdf.py

+13-8
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,5 @@
1+
import warnings
2+
13
import numpy as np
24
from pandas import DataFrame, Panel, date_range, HDFStore, read_hdf
35
import pandas.util.testing as tm
@@ -105,22 +107,25 @@ class HDFStorePanel(BaseIO):
105107

106108
def setup(self):
107109
self.fname = '__test__.h5'
108-
self.p = Panel(np.random.randn(20, 1000, 25),
109-
items=['Item%03d' % i for i in range(20)],
110-
major_axis=date_range('1/1/2000', periods=1000),
111-
minor_axis=['E%03d' % i for i in range(25)])
112-
self.store = HDFStore(self.fname)
113-
self.store.append('p1', self.p)
110+
with warnings.catch_warnings(record=True):
111+
self.p = Panel(np.random.randn(20, 1000, 25),
112+
items=['Item%03d' % i for i in range(20)],
113+
major_axis=date_range('1/1/2000', periods=1000),
114+
minor_axis=['E%03d' % i for i in range(25)])
115+
self.store = HDFStore(self.fname)
116+
self.store.append('p1', self.p)
114117

115118
def teardown(self):
116119
self.store.close()
117120
self.remove(self.fname)
118121

119122
def time_read_store_table_panel(self):
120-
self.store.select('p1')
123+
with warnings.catch_warnings(record=True):
124+
self.store.select('p1')
121125

122126
def time_write_store_table_panel(self):
123-
self.store.append('p2', self.p)
127+
with warnings.catch_warnings(record=True):
128+
self.store.append('p2', self.p)
124129

125130

126131
class HDF(BaseIO):

‎asv_bench/benchmarks/join_merge.py

+16-7
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,4 @@
1+
import warnings
12
import string
23

34
import numpy as np
@@ -26,7 +27,8 @@ def setup(self):
2627
self.mdf1['obj2'] = 'bar'
2728
self.mdf1['int1'] = 5
2829
try:
29-
self.mdf1.consolidate(inplace=True)
30+
with warnings.catch_warnings(record=True):
31+
self.mdf1.consolidate(inplace=True)
3032
except:
3133
pass
3234
self.mdf2 = self.mdf1.copy()
@@ -75,16 +77,23 @@ class ConcatPanels(object):
7577
param_names = ['axis', 'ignore_index']
7678

7779
def setup(self, axis, ignore_index):
78-
panel_c = Panel(np.zeros((10000, 200, 2), dtype=np.float32, order='C'))
79-
self.panels_c = [panel_c] * 20
80-
panel_f = Panel(np.zeros((10000, 200, 2), dtype=np.float32, order='F'))
81-
self.panels_f = [panel_f] * 20
80+
with warnings.catch_warnings(record=True):
81+
panel_c = Panel(np.zeros((10000, 200, 2),
82+
dtype=np.float32,
83+
order='C'))
84+
self.panels_c = [panel_c] * 20
85+
panel_f = Panel(np.zeros((10000, 200, 2),
86+
dtype=np.float32,
87+
order='F'))
88+
self.panels_f = [panel_f] * 20
8289

8390
def time_c_ordered(self, axis, ignore_index):
84-
concat(self.panels_c, axis=axis, ignore_index=ignore_index)
91+
with warnings.catch_warnings(record=True):
92+
concat(self.panels_c, axis=axis, ignore_index=ignore_index)
8593

8694
def time_f_ordered(self, axis, ignore_index):
87-
concat(self.panels_f, axis=axis, ignore_index=ignore_index)
95+
with warnings.catch_warnings(record=True):
96+
concat(self.panels_f, axis=axis, ignore_index=ignore_index)
8897

8998

9099
class ConcatDataFrames(object):

‎asv_bench/benchmarks/offset.py

+5-2
Original file line numberDiff line numberDiff line change
@@ -1,4 +1,5 @@
11
# -*- coding: utf-8 -*-
2+
import warnings
23
from datetime import datetime
34

45
import numpy as np
@@ -76,7 +77,8 @@ def setup(self, offset):
7677
self.data = pd.Series(rng)
7778

7879
def time_add_offset(self, offset):
79-
self.data + offset
80+
with warnings.catch_warnings(record=True):
81+
self.data + offset
8082

8183

8284
class OffsetDatetimeIndexArithmetic(object):
@@ -90,7 +92,8 @@ def setup(self, offset):
9092
self.data = pd.date_range(start='1/1/2000', periods=N, freq='T')
9193

9294
def time_add_offset(self, offset):
93-
self.data + offset
95+
with warnings.catch_warnings(record=True):
96+
self.data + offset
9497

9598

9699
class OffestDatetimeArithmetic(object):

‎asv_bench/benchmarks/panel_ctor.py

+7-3
Original file line numberDiff line numberDiff line change
@@ -1,3 +1,4 @@
1+
import warnings
12
from datetime import datetime, timedelta
23

34
from pandas import DataFrame, DatetimeIndex, date_range
@@ -19,7 +20,8 @@ def setup(self):
1920
self.data_frames[x] = df
2021

2122
def time_from_dict(self):
22-
Panel.from_dict(self.data_frames)
23+
with warnings.catch_warnings(record=True):
24+
Panel.from_dict(self.data_frames)
2325

2426

2527
class SameIndexes(object):
@@ -34,7 +36,8 @@ def setup(self):
3436
self.data_frames = dict(enumerate([df] * 100))
3537

3638
def time_from_dict(self):
37-
Panel.from_dict(self.data_frames)
39+
with warnings.catch_warnings(record=True):
40+
Panel.from_dict(self.data_frames)
3841

3942

4043
class TwoIndexes(object):
@@ -53,4 +56,5 @@ def setup(self):
5356
self.data_frames = dict(enumerate(dfs))
5457

5558
def time_from_dict(self):
56-
Panel.from_dict(self.data_frames)
59+
with warnings.catch_warnings(record=True):
60+
Panel.from_dict(self.data_frames)

0 commit comments

Comments
 (0)