3
5
Fork 0
mirror of git://git.savannah.gnu.org/guix.git synced 2023-12-14 03:33:07 +01:00
guix/gnu/packages/patches/python-statsmodels-fix-tests.patch
Ricardo Wurmus 8b52167e77
gnu: python-statsmodels: Fix tests.
* gnu/packages/statistics.scm (python-statsmodels): Add patch to fix tests.
(python2-statsmodels)[propagated-inputs]: Add python2-pytz.
* gnu/packages/patches/python-statsmodels-fix-tests.patch: New file.
* gnu/local.mk (dist_patch_DATA): Add it.
2016-08-04 13:40:13 +02:00

196 lines
7.9 KiB
Diff

This patch fixes a couple of test failures introduced by changes to the pandas
package. It was extracted from this pull request:
https://github.com/statsmodels/statsmodels/pull/2675
From c9ef60a7bc4407766ab9e9f12c8a6b89013046ee Mon Sep 17 00:00:00 2001
From: Ralf Gommers <ralf.gommers@gmail.com>
Date: Tue, 20 Oct 2015 07:34:11 +0200
Subject: [PATCH 1/4] MAINT: fix use of old_behavior kw for numpy.correlate.
Was removed in 1.10.0
Numpy PR that removed it: https://github.com/numpy/numpy/pull/5991
Closes gh-2667.
---
statsmodels/tsa/ar_model.py | 6 ++----
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/statsmodels/tsa/ar_model.py b/statsmodels/tsa/ar_model.py
index 087a9e0..02984bd 100644
--- a/statsmodels/tsa/ar_model.py
+++ b/statsmodels/tsa/ar_model.py
@@ -261,10 +261,8 @@ def _presample_varcov(self, params):
Vpinv = np.zeros((p, p), dtype=params.dtype)
for i in range(1, p1):
- Vpinv[i-1, i-1:] = np.correlate(params0, params0[:i],
- old_behavior=False)[:-1]
- Vpinv[i-1, i-1:] -= np.correlate(params0[-i:], params0,
- old_behavior=False)[:-1]
+ Vpinv[i-1, i-1:] = np.correlate(params0, params0[:i],)[:-1]
+ Vpinv[i-1, i-1:] -= np.correlate(params0[-i:], params0,)[:-1]
Vpinv = Vpinv + Vpinv.T - np.diag(Vpinv.diagonal())
return Vpinv
From f1dc8979b09bc1736149993f895943b3158ee2db Mon Sep 17 00:00:00 2001
From: Ralf Gommers <ralf.gommers@gmail.com>
Date: Wed, 21 Oct 2015 22:05:52 +0200
Subject: [PATCH 2/4] MAINT: fix graphics module for changes in recent pandas
versions.
---
statsmodels/graphics/tests/test_mosaicplot.py | 2 +-
statsmodels/graphics/tests/test_tsaplots.py | 6 +++---
statsmodels/graphics/tsaplots.py | 2 +-
3 files changed, 5 insertions(+), 5 deletions(-)
diff --git a/statsmodels/graphics/tests/test_mosaicplot.py b/statsmodels/graphics/tests/test_mosaicplot.py
index cb9bbbe..e41020e 100644
--- a/statsmodels/graphics/tests/test_mosaicplot.py
+++ b/statsmodels/graphics/tests/test_mosaicplot.py
@@ -113,7 +113,7 @@ def test_mosaic():
# sort by the marriage quality and give meaningful name
# [rate_marriage, age, yrs_married, children,
# religious, educ, occupation, occupation_husb]
- datas = datas.sort(['rate_marriage', 'religious'])
+ datas = datas.sort_values(by=['rate_marriage', 'religious'])
num_to_desc = {1: 'awful', 2: 'bad', 3: 'intermediate',
4: 'good', 5: 'wonderful'}
datas['rate_marriage'] = datas['rate_marriage'].map(num_to_desc)
diff --git a/statsmodels/graphics/tests/test_tsaplots.py b/statsmodels/graphics/tests/test_tsaplots.py
index 511f18f..365be82 100644
--- a/statsmodels/graphics/tests/test_tsaplots.py
+++ b/statsmodels/graphics/tests/test_tsaplots.py
@@ -1,4 +1,4 @@
-from statsmodels.compat.python import lmap, lzip, map
+from statsmodels.compat.python import lmap, map
import numpy as np
import pandas as pd
from numpy.testing import dec
@@ -51,8 +51,8 @@ def test_plot_month():
dta = sm.datasets.elnino.load_pandas().data
dta['YEAR'] = dta.YEAR.astype(int).apply(str)
dta = dta.set_index('YEAR').T.unstack()
- dates = lmap(lambda x : pd.datetools.parse('1 '+' '.join(x)),
- dta.index.values)
+ dates = lmap(lambda x : pd.datetools.parse_time_string('1 '+' '.join(x))[0],
+ dta.index.values)
# test dates argument
fig = month_plot(dta.values, dates=dates, ylabel='el nino')
diff --git a/statsmodels/graphics/tsaplots.py b/statsmodels/graphics/tsaplots.py
index 3d04692..94626c9 100644
--- a/statsmodels/graphics/tsaplots.py
+++ b/statsmodels/graphics/tsaplots.py
@@ -200,7 +200,7 @@ def seasonal_plot(grouped_x, xticklabels, ylabel=None, ax=None):
ticks = []
for season, df in grouped_x:
df = df.copy() # or sort balks for series. may be better way
- df.sort()
+ df.sort_values(inplace=True)
nobs = len(df)
x_plot = np.arange(start, start + nobs)
ticks.append(x_plot.mean())
From 4cfbef6af137629c6953f1f025d9cfc781874256 Mon Sep 17 00:00:00 2001
From: Ralf Gommers <ralf.gommers@gmail.com>
Date: Wed, 21 Oct 2015 22:15:25 +0200
Subject: [PATCH 3/4] MAINT: work around pandas breaking backwards compat for
pandas.version
---
setup.py | 5 ++++-
statsmodels/tools/testing.py | 6 ++----
2 files changed, 6 insertions(+), 5 deletions(-)
diff --git a/setup.py b/setup.py
index 0002840..74aefb8 100644
--- a/setup.py
+++ b/setup.py
@@ -134,7 +134,10 @@ def check_dependency_versions(min_versions):
(spversion, min_versions['scipy']))
try:
- from pandas.version import short_version as pversion
+ import pandas
+ #FIXME: this will break for pandas 1.0.0. Needs elaborate parsing now,
+ # due to pandas removing version.short_version
+ pversion = pandas.__version__[:6]
except ImportError:
install_requires.append('pandas')
else:
diff --git a/statsmodels/tools/testing.py b/statsmodels/tools/testing.py
index e207e44..643f79f 100644
--- a/statsmodels/tools/testing.py
+++ b/statsmodels/tools/testing.py
@@ -16,10 +16,8 @@ def strip_rc(version):
def is_pandas_min_version(min_version):
- '''check whether pandas is at least min_version
- '''
- from pandas.version import short_version as pversion
- return StrictVersion(strip_rc(pversion)) >= min_version
+ '''check whether pandas is at least min_version '''
+ return StrictVersion((pandas.__version__[:6])) >= min_version
# local copies, all unchanged
From c894c3f4882d570efb517950069d83afa9794db8 Mon Sep 17 00:00:00 2001
From: Ralf Gommers <ralf.gommers@gmail.com>
Date: Mon, 26 Oct 2015 20:47:51 +0100
Subject: [PATCH 4/4] BUG: fix use of Series.sort_values for older pandas.
Some failing tests in the previous commits because older ``pandas`` versions
don't have ``Series.sort_values``. That method was only added in pandas 0.17,
in https://github.com/pydata/pandas/pull/10726
---
statsmodels/graphics/tests/test_mosaicplot.py | 6 +++++-
statsmodels/graphics/tsaplots.py | 6 +++++-
2 files changed, 10 insertions(+), 2 deletions(-)
diff --git a/statsmodels/graphics/tests/test_mosaicplot.py b/statsmodels/graphics/tests/test_mosaicplot.py
index e41020e..2a873e7 100644
--- a/statsmodels/graphics/tests/test_mosaicplot.py
+++ b/statsmodels/graphics/tests/test_mosaicplot.py
@@ -113,7 +113,11 @@ def test_mosaic():
# sort by the marriage quality and give meaningful name
# [rate_marriage, age, yrs_married, children,
# religious, educ, occupation, occupation_husb]
- datas = datas.sort_values(by=['rate_marriage', 'religious'])
+ if pandas.__version__ < '0.17.0':
+ datas = datas.sort(['rate_marriage', 'religious'])
+ else:
+ datas = datas.sort_values(by=['rate_marriage', 'religious'])
+
num_to_desc = {1: 'awful', 2: 'bad', 3: 'intermediate',
4: 'good', 5: 'wonderful'}
datas['rate_marriage'] = datas['rate_marriage'].map(num_to_desc)
diff --git a/statsmodels/graphics/tsaplots.py b/statsmodels/graphics/tsaplots.py
index 94626c9..217724f 100644
--- a/statsmodels/graphics/tsaplots.py
+++ b/statsmodels/graphics/tsaplots.py
@@ -2,6 +2,7 @@
import numpy as np
+import pandas
from statsmodels.graphics import utils
from statsmodels.tsa.stattools import acf, pacf
@@ -200,7 +201,10 @@ def seasonal_plot(grouped_x, xticklabels, ylabel=None, ax=None):
ticks = []
for season, df in grouped_x:
df = df.copy() # or sort balks for series. may be better way
- df.sort_values(inplace=True)
+ if pandas.__version__ < '0.17.0':
+ df.sort()
+ else:
+ df.sort_values(inplace=True)
nobs = len(df)
x_plot = np.arange(start, start + nobs)
ticks.append(x_plot.mean())