Fix indexing errors in KaplanMeier

Author: Rebecca N. Palmer <rebecca_palmer@zoho.com>
Bug-Debian: https://bugs.debian.org/860449
Forwarded: not-needed (affected code no longer in upstream)

Gbp-Pq: Name kaplan_meier_860449.patch

Stop using no-longer-existing pd.ols

Fixes
module 'pandas' has no attribute 'ols'
e.g. in the vector_ar documentation examples.

Origin: upstream https://github.com/statsmodels/statsmodels/commit/4d76116c9c3f47c629503837b3db91bec4d22a5a
Author: Kevin "bashtage" Sheppard
Forwarded: not-needed

Gbp-Pq: Name vector_ar_window_ols.patch

Fix errors in examples

Be compatible with current pandas
https://github.com/statsmodels/statsmodels/commit/66a4f98d0ac86ea957fa6f8168cc2df178c64630
https://github.com/statsmodels/statsmodels/commit/67415ad8d4c04dc1dd9ea9ebb1e059af92646c7b
https://github.com/statsmodels/statsmodels/commit/572d507dfa28044166cf4f183177a046f90934ae

and with one dataset layout change within statsmodels, based on
https://github.com/statsmodels/statsmodels/commit/31ec838d1eab2ec559ab69d42151085bdf9f18e1

Look in the correct path for star_diagram.png
(using this fix instead of
https://github.com/statsmodels/statsmodels/commit/3363ce1dc41b1717db5b6779a2a9cbf8f72a0294
to be minimally risky during freeze)

Author: Kevin "bashtage" Sheppard, Rebecca N. Palmer <rebecca_palmer@zoho.com>
Origin: upstream (mostly)
Forwarded: not-needed

Gbp-Pq: Name examples_fixes.patch

Don't use np.genfromtxt on non-ASCII data

Fixes UnicodeDecodeError on file e6.dat

Author: Peter "thequackdaddy" Quackenbush, Rebecca N. Palmer <rebecca_palmer@zoho.com>
Origin: upstream
Forwarded: not-needed

Gbp-Pq: Name lutkepohl_data_load.patch

Create a normal array, not an array of 1x1 arrays

Fixes "object arrays are not supported".

Author: Kevin "bashtage" Sheppard
Origin: upstream https://github.com/statsmodels/statsmodels/pull/4911/commits/18dcceefbfed61b0d11ae5884a9f4cea82c2edb5
Forwarded: not-needed

Gbp-Pq: Name unconstrain_stationary_multivariate.patch

Make tests compatible with current numpy/scipy/pandas

These issues are all within test code, i.e. should not affect users.

https://github.com/statsmodels/statsmodels/commit/bb8fba9d4d9883be2991b43b29bfa3a9c3fb325a
https://github.com/statsmodels/statsmodels/commit/310f91da9f07d9483965d624d76dbb7abebbd67f
https://github.com/statsmodels/statsmodels/pull/4911/commits/18dcceefbfed61b0d11ae5884a9f4cea82c2edb5
https://github.com/statsmodels/statsmodels/commit/f1d5eddd44d61099a9f8f90c40c35716a8346cf7

and one new fix only needed on 32 bit Python 2 (integer data loads as
int64, which gets converted to long producing factor names like
C(agecat)[T.4L], which does not match C(agecat)[T.4])

Author: Kevin "bashtage" Sheppard, Peter "thequackdaddy" Quackenbush, Matthew Brett, Rebecca N. Palmer <rebecca_palmer@zoho.com>
Origin: upstream (mostly)
Forwarded: no

Gbp-Pq: Name test_fixes.patch

Allow loading the CO2 dataset in pandas >= 0.23

Fixes __new__() got an unexpected keyword argument 'format'

Author: Matthew Brett
Origin: upstream https://github.com/statsmodels/statsmodels/commit/bd2985d781484c2a2ce47a3db107690bbe688acd
Forwarded: not-needed

Gbp-Pq: Name co2_dataset_load.patch

Don't use gammaln on complex input

In scipy >= 1.0, it rejects such input with
TypeError: ufunc 'gammaln' not supported for the input types, and the
inputs could not be safely coerced to any supported types according
to the casting rule ''safe''

Author: Kevin "bashtage" Sheppard, Rebecca N. Palmer <rebecca_palmer@zoho.com>
Origin: upstream https://github.com/statsmodels/statsmodels/pull/3964/commits/c11c6242ff81a0f83d3e7fd83a8105397d3cbc1d
Forwarded: not-needed

Gbp-Pq: Name gammaln_complex.patch

Fix shape mismatch errors in MICE imputation

numpy/pandas no longer allows assigning an nx1 2D value at a 1D index

Author: Peter "thequackdaddy" Quackenbush, Josef "josef-pkt" Perktold
Origin: upstream https://github.com/statsmodels/statsmodels/commit/df92bde1a85c238db9561304a196dbc618cfeaa1 https://github.com/statsmodels/statsmodels/commit/0871eba510e7ebe8aa9323f4f3ee50226576ed29
Forwarded: not-needed

Gbp-Pq: Name imputation_broadcasting.patch

Stop using no-longer-existing chisqprob

Needed for compatibility with scipy >= 1.0

Author: Kevin "bashtage" Sheppard
Origin: upstream https://github.com/statsmodels/statsmodels/commit/575fc69257171e1dbc86461b510eb818a63c5fab
Forwarded: not-needed

Gbp-Pq: Name chisqprob.patch

Make the datasets cache work in Python 3

Fixes URLError messages appearing in the built documentation

Author: Rebecca N. Palmer <rebecca_palmer@zoho.com>
Forwarded: not-needed (upstream fixed this in _cache_it instead)

Gbp-Pq: Name python3_dataset_caching.patch

Don't fetch Twitter logo

Author: Rebecca N. Palmer <rebecca_palmer@zoho.com>
Forwarded: no

Gbp-Pq: Name privacy.patch

Fix spelling

Author: Rebecca N. Palmer <rebecca_palmer@zoho.com>
Origin: lintian (partly)
Forwarded: https://github.com/statsmodels/statsmodels/pull/5513

Gbp-Pq: Name spelling.patch

Use mathjax instead of no-longer-existing pngmath

Either imgmath or mathjax fixes the FTBFS;
mathjax also makes TeX display as equations, i.e. fixes #687301

Author: Rebecca N. Palmer <rebecca_palmer@zoho.com>
Forwarded: not-needed (mathjax is already the upstream default)

Gbp-Pq: Name no-sphinx-pngmath.patch

Do not try to access remote locations on buildtime test

Bug-Debian: https://bugs.debian.org/882641

Gbp-Pq: Name prevent_online_access_in_buildtime_tests.patch

Use intersphinx inventory files from other Debian packages.

Forwarded: not-needed

Gbp-Pq: Name use-system-inventory

Use Python3 to create documentation

Bug: https://github.com/statsmodels/statsmodels/pull/5348

Gbp-Pq: Name use-python3-for-doc-build

fix_numpydoc_issue

Author; Denis Laxalde <denis@laxalde.org>
Last-Update: Thu, 17 Aug 2017 22:06:46 +0200
Description:  Fix numpy doc input

Gbp-Pq: Name fix_numpydoc_issue.patch

Seems the tests are failing due to some problem in pandas

Last-Update: Thu, 17 Aug 2017 12:47:48 +0200
Bug-Debian: https://bugs.debian.org/868977

which we can not fix here.  Ignore the issue for the moment to be able
to drop severity of bug #868977 from serious to important

Gbp-Pq: Name skip_tests_failing_due_to_panda_issue.patch

deb_skip_test_ons390

Gbp-Pq: Name deb_skip_test_ons390

to prevent failed test on i386

Bug: https://github.com/statsmodels/statsmodels/issues/1831
Last-Update: 2014-07-16

Gbp-Pq: Name up_reduce_test_precision

Use Debian packaged IPhythonSphinx extensions

Last-Update: Wed, 13 Feb 2019 08:43:05 +0100

Also drop matplotlib.sphinxext.only_directives which fail with
python3-matplotlib and are not used

Gbp-Pq: Name deb_use_internal_ipythonsphinx_extensions

Use cached datasets for building documentation/examples

Also remove a download that isn't actually used in that example.

This allows at least some of the examples to be built in an
offline environment such as a Debian buildd.

See debian/README.source for more information about the cached datasets.

Author: Diane Trout <diane@ghic.org>, Rebecca N. Palmer <rebecca_palmer@zoho.com>

Gbp-Pq: Name use-cached-datasets

statsmodels (0.8.0-9) unstable; urgency=medium

  * Team upload.
  * Disable KalmanFilter on armhf, as it gives wrong results
    (and probably always has).

[dgit import unpatched statsmodels 0.8.0-9]

Import statsmodels_0.8.0-9.debian.tar.xz

[dgit import tarball statsmodels 0.8.0-9 statsmodels_0.8.0-9.debian.tar.xz]

Import statsmodels_0.8.0.orig.tar.gz

[dgit import orig statsmodels_0.8.0.orig.tar.gz]