docs/source/contingency_tables.rst | 4 2 + 2 - 0 !
docs/source/datasets/index.rst | 2 1 + 1 - 0 !
docs/source/duration.rst | 4 2 + 2 - 0 !
docs/source/example_formulas.rst | 2 1 + 1 - 0 !
docs/source/gee.rst | 2 1 + 1 - 0 !
docs/source/gettingstarted.rst | 2 1 + 1 - 0 !
docs/source/index.rst | 2 1 + 1 - 0 !
docs/source/mixed_linear.rst | 2 1 + 1 - 0 !
docs/source/release/version0.6.rst | 2 1 + 1 - 0 !
examples/notebooks/markov_regression.ipynb | 9 1 + 8 - 0 !
examples/notebooks/mixed_lm_example.ipynb | 4 2 + 2 - 0 !
examples/notebooks/regression_diagnostics.ipynb | 4 2 + 2 - 0 !
statsmodels/stats/tests/test_dist_dependant_measures.py | 4 2 + 2 - 0 !
13 files changed, 18 insertions(+), 25 deletions(-)

 use cached datasets in tests and examples

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

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

The cached data is extracted from R packages by debian/datasets/*.

docs/source/conf.py | 8 4 + 4 - 0 !
1 file changed, 4 insertions(+), 4 deletions(-)

 use intersphinx inventory files from other debian packages.

examples/notebooks/generic_mle.ipynb | 68 43 + 25 - 0 !
1 file changed, 43 insertions(+), 25 deletions(-)

 use example data from an r package we have

statsmodels/discrete/tests/test_truncated_model.py | 12 11 + 1 - 0 !
statsmodels/iolib/tests/test_summary2.py | 19 19 + 0 - 0 !
statsmodels/regression/tests/test_lme.py | 6 5 + 1 - 0 !
statsmodels/tsa/holtwinters/tests/test_holtwinters.py | 17 16 + 1 - 0 !
statsmodels/tsa/statespace/tests/test_exact_diffuse_filtering.py | 7 7 + 0 - 0 !
statsmodels/tsa/statespace/tests/test_multivariate_switch_univariate.py | 8 6 + 2 - 0 !
statsmodels/tsa/statespace/tests/test_smoothing.py | 10 6 + 4 - 0 !
7 files changed, 70 insertions(+), 9 deletions(-)

 loosen tolerances on some tests on i386

and xfail a test that fails for not producing ConvergenceWarning.

Failure logs:
test_smoothing
https://buildd.debian.org/status/fetch.php?pkg=statsmodels&arch=i386&ver=0.9.0-3&stamp=1567157609&raw=0
test_multivariate_switch_univariate
https://buildd.debian.org/status/fetch.php?pkg=statsmodels&arch=i386&ver=0.12.0-1&stamp=1599693472&raw=0

archive/docs/fix_longtable.py | 2 1 + 1 - 0 !
examples/python/plots_boxplots.py | 2 1 + 1 - 0 !
examples/python/robust_models_1.py | 2 1 + 1 - 0 !
statsmodels/regression/quantile_regression.py | 2 1 + 1 - 0 !
statsmodels/sandbox/examples/example_pca.py | 2 1 + 1 - 0 !
statsmodels/tools/print_version.py | 2 1 + 1 - 0 !
tools/export_notebooks_to_python.py | 2 1 + 1 - 0 !
tools/generate_formula_api.py | 2 1 + 1 - 0 !
8 files changed, 8 insertions(+), 8 deletions(-)

 use python 3 in script shebangs

(Skipping the two .pyx files with #!python - they need to be compiled
with Cython, not directly run with any version of plain Python)

statsmodels/datasets/tests/test_utils.py | 18 8 + 10 - 0 !
1 file changed, 8 insertions(+), 10 deletions(-)

 don't try to write to the source directory

Not allowed in autopkgtest

docs/source/conf.py | 5 2 + 3 - 0 !
1 file changed, 2 insertions(+), 3 deletions(-)

 don't try to use a theme we don't have

Upstream switched to sphinx_material in
1d7aafad5c3ad634c9ae2122881c1809c2ccbcbe
and later to sphinx_immaterial; Debian doesn't have those themes.

(I haven't investigated packaging it, which might be a better long-term option.)

docs/Makefile | 2 2 + 0 - 0 !
docs/source/conf.py | 1 1 + 0 - 0 !
2 files changed, 3 insertions(+)

 don't fail documentation build on examples exceptions

If example code outputs an error when run,
ipython (ipython:: blocks) now fails the whole build by default,
and tools/nbgenerate.py (.ipynb files) now leaves out that file
(leaving a broken link in the examples index).

As some examples use downloaded data and are hence expected to be
unrunnable on a Debian buildd, this instead keeps them
(::ipython with error output, .ipynb without output).

docs/source/conf.py | 2 2 + 0 - 0 !
1 file changed, 2 insertions(+)

 use local requirejs/mathjax

statsmodels/graphics/tests/test_functional.py | 38 38 + 0 - 0 !
1 file changed, 38 insertions(+)

 allow tests to fail if multiprocessing is not available

This is currently the case on hurd-i386

statsmodels/examples/ex_generic_mle_tdist.py | 3 3 + 0 - 0 !
statsmodels/graphics/regressionplots.py | 5 3 + 2 - 0 !
statsmodels/tools/sequences.py | 29 14 + 15 - 0 !
3 files changed, 20 insertions(+), 17 deletions(-)

 remove code from stack overflow and wasteful loop

Stack Overflow content is CC-BY-SA licensed,
which this package is not supposed to be.  This snippet may be
too small to be copyrightable, but removing it to be safe.

statsmodels/stats/tests/test_corrpsd.py | 8 4 + 4 - 0 !
1 file changed, 4 insertions(+), 4 deletions(-)

 xfail / increase tolerance on tests

TestMICE and test_mixedlm vary enough with the random state, and
corrpsd is close enough to 0, that this is plausibly a rounding non-bug

statsmodels/tsa/exponential_smoothing/ets.py | 9 9 + 0 - 0 !
statsmodels/tsa/tests/test_exponential_smoothing.py | 3 2 + 1 - 0 !
2 files changed, 11 insertions(+), 1 deletion(-)

 test_seasonal_order fails on i386

Warn on use and xfail

statsmodels/sandbox/regression/tests/test_gmm.py | 3 2 + 1 - 0 !
1 file changed, 2 insertions(+), 1 deletion(-)

 avoid test failure on mipsel

statsmodels/tsa/statespace/tests/test_dynamic_factor.py | 2 1 + 1 - 0 !
1 file changed, 1 insertion(+), 1 deletion(-)

 loosen tolerance on a possibly unstable test

statsmodels/base/tests/test_generic_methods.py | 4 3 + 1 - 0 !
1 file changed, 3 insertions(+), 1 deletion(-)

 ignore missing hessianinversionwarning on some mips*

statsmodels/distributions/tests/test_discrete.py | 4 3 + 1 - 0 !
1 file changed, 3 insertions(+), 1 deletion(-)

 avoid cast exception in np.bincount

statsmodels/stats/tests/test_corrpsd.py | 5 5 + 0 - 0 !
1 file changed, 5 insertions(+)

 ignore a failure to converge on mipsel and armel

(Seen to fail in 0.13.5+dfsg-5, twice on mipsel, and in 0.14.0 on armel.
Not a plain mark.xfail to not accept "converged" to the wrong answer.)

statsmodels/tsa/statespace/tests/test_dynamic_factor_mq_frbny_nowcast.py | 3 2 + 1 - 0 !
1 file changed, 2 insertions(+), 1 deletion(-)

 allow a memory-heavy test to run out of memory

Failed in 0.14.0 on armel,armhf,mipsel (but not i386)

statsmodels/stats/tests/test_rates_poisson.py | 8 7 + 1 - 0 !
1 file changed, 7 insertions(+), 1 deletion(-)

 don't require a warning armel numpy doesn't have

statsmodels/discrete/tests/test_discrete.py | 4 3 + 1 - 0 !
statsmodels/regression/tests/test_regression.py | 4 3 + 1 - 0 !
statsmodels/tsa/holtwinters/tests/test_holtwinters.py | 11 7 + 4 - 0 !
3 files changed, 13 insertions(+), 6 deletions(-)

 ignore extra warnings that may occur on mips*

Some mips* hardware seems to emit
RuntimeWarning: invalid value encountered in multiply
on multiplication by explicit np.nan, which breaks these tests.

statsmodels/tsa/exponential_smoothing/_ets_smooth.pyx | 2 1 + 1 - 0 !
statsmodels/tsa/holtwinters/_exponential_smoothers.pyx | 2 1 + 1 - 0 !
statsmodels/tsa/statespace/_filters/_inversions.pyx.in | 1 1 + 0 - 0 !
statsmodels/tsa/statespace/_filters/_univariate.pyx.in | 3 2 + 1 - 0 !
statsmodels/tsa/statespace/_filters/_univariate_diffuse.pyx.in | 3 2 + 1 - 0 !
5 files changed, 7 insertions(+), 4 deletions(-)

 allow building with cython 3.x

(= Cython _version_ 3.x, not to be confused with the cython3 binary)

docs/source/emplike.rst | 1 1 + 0 - 0 !
examples/notebooks/copula.ipynb | 3 2 + 1 - 0 !
examples/notebooks/ets.ipynb | 3 2 + 1 - 0 !
examples/notebooks/exponential_smoothing.ipynb | 1 1 + 0 - 0 !
examples/notebooks/gee_score_test_simulation.ipynb | 3 2 + 1 - 0 !
examples/notebooks/plots_boxplots.ipynb | 3 2 + 1 - 0 !
examples/notebooks/predict.ipynb | 3 2 + 1 - 0 !
examples/notebooks/robust_models_0.ipynb | 3 2 + 1 - 0 !
8 files changed, 14 insertions(+), 6 deletions(-)

 fix random seeds for reproducibility

statsmodels/tsa/statespace/tests/test_dynamic_factor_mq_frbny_nowcast.py | 2 2 + 0 - 0 !
1 file changed, 2 insertions(+)

 ignore test exception on big-endian systems

failed in 0.14.1+dfsg-1

statsmodels/discrete/tests/test_count_model.py | 3 2 + 1 - 0 !
1 file changed, 2 insertions(+), 1 deletion(-)

 ignore a non-converging test

(it currently does warn about possible non-convergence and
it would be better to check that it does so)