Version 2.43
Installing the Python interface through PyPI is supported.
Version 2.42
For dual CD solvers (logistic/l2 losses but not l1 loss), if a maximal number of iterations is reached, LIBLINEAR directly switches to run a primal Newton solver.
KISS FFT - A mixed-radix Fast Fourier Transform
There are many great fft libraries already around. Kiss FFT is not trying
to be better than any of them. It only attempts to be a reasonably efficient,
moderately useful FFT that can use fixed or floating data types and can be
incorporated into someone's C program in a few minutes with trivial licensing.
This introduces a 'generic' variant for blas_opt_info() and
lapack_opt_info() that looks into BLAS_LIBS and LAPACK_LIBS. The
others stay alive, and you can choose via NPY_BLAS_ORDER and
NPY_LAPACK_ORDER.
Will post this upstream … hoping for the best.
Other projects seem just to abuse the [openblas] entry in
site.cfg, or whatever hacks to trick this into building with
a generic BLAS. One hurdle is that using openblas seems to
assume that CBLAS is present in there, which it is not in pkgsrc.
We could think about introducing some alternatives stuff that uses
stubs for all the related libs anyway, but that is not my application,
which builds things from source. For binary packages, it would be
interesting, though. See Debian and Genoot for prior art on that.
Many many changes including
Oneway ANOVA-type analysis
~~~~~~~~~~~~~~~~~~~~~~~~~~
Several statistical methods for ANOVA-type analysis of k independent samples
have been added in module :mod:`~statsmodels.stats.oneway`. This includes
standard Anova, Anova for unequal variances (Welch, Brown-Forsythe for mean),
Anova based on trimmed samples (Yuen anova) and equivalence testing using
the method of Wellek.
Anova for equality of variances or dispersion are available for several
transformations. This includes Levene test and Browne-Forsythe test for equal
variances as special cases. It uses the `anova_oneway` function, so unequal
variance and trimming options are also available for tests on variances.
Several functions for effect size measures have been added, that can be used
for reporting or for power and sample size computation.
Multivariate statistics
~~~~~~~~~~~~~~~~~~~~~~~
The new module :mod:`~statsmodels.stats.multivariate` includes one and
two sample tests for multivariate means, Hotelling's t-tests',
:func:`~statsmodels.stats.multivariate.test_mvmean`,
:func:`~statsmodels.stats.multivariate.test_mvmean_2indep` and confidence
intervals for one-sample multivariate mean
:func:`~statsmodels.stats.multivariate.confint_mvmean`
Additionally, hypothesis tests for covariance patterns, and for oneway equality
of covariances are now available in several ``test_cov`` functions.
New exponential smoothing model: ETS (Error, Trend, Seasonal)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
- Class implementing ETS models :class:`~statsmodels.tsa.exponential_smoothing.ets.ETSModel`.
- Includes linear and non-linear exponential smoothing models
- Supports parameter fitting, in-sample prediction and out-of-sample
forecasting, prediction intervals, simulation, and more.
- Based on the innovations state space approach.
Forecasting Methods
~~~~~~~~~~~~~~~~~~~
Two popular methods for forecasting time series, forecasting after
STL decomposition (:class:`~statsmodels.tsa.forecasting.stl.STLForecast`)
and the Theta model
(:class:`~statsmodels.tsa.forecasting.theta.ThetaModel`) have been added.
See 0.12.0-0.12.2 at https://www.statsmodels.org/stable/release/
for the full story, including deprecations.
2.7.3 (2021-02-25)
------------------
- Add pytest plugin options to skip and xfail individual tests
and xfail the unsupported ndarray-1.0.0 example.
- Fix bug resulting in invalid strides values for views over
FITS arrays.
2.7.2 (2021-01-15)
------------------
- Fix bug causing test collection failures in some environments.
- Fix bug when decompressing arrays with numpy 1.20
Highlights of this release
scipy.ndimage improvements: Fixes and ehancements to boundary extension
modes for interpolation functions. Support for complex-valued inputs in many
filtering and interpolation functions. New grid_mode option for
scipy.ndimage.zoom to enable results consistent with scikit-image's
rescale.
scipy.optimize.linprog has fast, new methods for large, sparse problems
from the HiGHS library.
scipy.stats improvements including new distributions, a new test, and
enhancements to existing distributions and tests
Deprecated features
scipy.spatial changes
Calling KDTree.query with k=None to find all neighbours is deprecated.
Use KDTree.query_ball_point instead.
distance.wminkowski was deprecated; use distance.minkowski and supply
weights with the w keyword instead.
Backwards incompatible changes
Using scipy.fft as a function aliasing numpy.fft.fft was removed after
being deprecated in SciPy 1.4.0. As a result, the scipy.fft submodule
must be explicitly imported now, in line with other SciPy subpackages.
scipy.signal changes
The output of decimate, lfilter_zi, lfiltic, sos2tf, and
sosfilt_zi have been changed to match numpy.result_type of their inputs.
The window function slepian was removed.
The frechet_l and frechet_r distributions were removed.
Packaging changes:
- Some patches got upstreamed
- Upstream now defaults to name=sc-im so it no longer needs to be set
Changes, (reformatted) from the official changelog:
New features:
- New motions in edit mode: df cf F dF cF d0 d$ c0 c$ t T dt dT ct cT ^
g_ d^ dg_ c^ cg_
- Simple mouse support. Can handle selection of single cells and scroll.
- Added markdown export.
- Add custom colors with RGB definitions.
- Add support for italic attribute to be appled to cells.
- Added digraphs.
- Added EXECUTE command.
- Added --export_csv --export_tab --export_txt parameters.
- New input_bar_bottom config variable.
- New underline_grid config variable.
- Added filename in status bar (filename_with_mode config variable).
- Added option to truncate cells.
- Added xlsx sheet parameter.
- New @fact function.
Other:
- modified config and history file locations
- renamed winch to sig_winch.
- current fg and bg colors are kept if they are not specified in
:cellcolor.
- OpenBSD support (mostly wordexp() workaround)
- Correct lua api function names in doc
- always generate backup file when opening file
- renamed scim to sc-im in Makefile
- Document trigger return value convention
Issues fixed:
- fix bug when deleting an ent that for instance has a @sum()... it
used to remove its vertex and THE ENTS LINKED to THEM!
- fix error messages when using "scim --version" and color definitions
are intented to be applied from .scimrc.
- fix in waiting for valid command to complete after pressing ESC key.
- fix in parse_str so a word with white space can be stored in
dictionary value
- fix when setting default_paste_from_clipboard_cmd
- Don't clobber startup message w/ err opening file
- Allow for lua scripts in local .sc directory #259
- Send informational messages to stderr rather than the output when
used in non-interactive mode #263
- fixed cell_negative color #271
- fixed locked cells when saving #261
- fixed DEL key in insert mode #272
- fix when resizing column #266
- "Autobackup - case insensitive QER options when backup exists"
- fixed segfault when fcopy'ing with no selection while on first column
- Call write triggers on value clear
- Use sc_info to report trigger exit code
- fixed#277
- Grow table up to MAXROWS exactly
- Remove redundant if clause (God only knows where that was).
- fix in getVertex that prevented rebuild_graph to work properly.
- fix annoying bug because of not resetting inputline_pos to 0 when
confirming a command in COMMAND_MODE
- fix in let and slet. existing vertexs should not be removed.
- added "eval_visited" in vertex struct for not to collide with
current "visited", since:
EvalAll uses EvalBottomUp
EvalBottomUp uses EvalJustOneVertex
EvalJustOneVertex uses eval
eval uses GraphAddEdge
GraphAddEdge uses GraphIsReachable
GraphIsReachable uses visited
and EvalBottomUp also uses uses visited!
- Also changed markAllVerticesNotVisited and
All_vertexs_of_edges_visited functions.
- Some other issues fixed: #228, #234, #239, #240, #244, #246, #260,
#295, #308
- .. and many other fixes and improvements!
Pending:
The most significative issue is regarding circular references.
This have to be deeply analyzed and with the collaboration of the
community
it would be hopefully fixed by v0.9.
The downstream configure code freaked out when info['define_macros'] was not set.
It is now defined and empty.
I hope it's fine to just push this without further notice.
Build was confirmed by oster on NetBSD-9.1/amd64 and tested by me
on Linux/amd64.
Fix openblas build issues, mainly by updating to 0.3.10.
This pulls in these commits from WIP (newest first):
commit 3c6284cba90280bc367cf4d1d8252ae4d6e92e76
Author: Jason Bacon <bacon@NetBSD.org>
Date: Thu Feb 25 11:56:13 2021 -0600
openblas: Update ONLY_FOR_PLATFORMS documentation
commit 8071bf28f3ffc95af046ff3eaaac6983f4f70035
Author: Jason Bacon <bacon@NetBSD.org>
Date: Thu Feb 25 11:51:32 2021 -0600
openblas*: Successful build on NetBSD
commit 056e3d5c972a4b286e8755dbee323a9951855165
Author: Dr. Thomas Orgis <thomas.orgis@uni-hamburg.de>
Date: Wed Feb 24 18:40:17 2021 +0100
openblas: flags from environment again to un-break PICy build
The bug that prompted us to force the compiler flags in the
make arguments is fixed. Now we got a different one: The logic
that decides to add -fPIC where needed is overridden when doing
this, resulting in relocation errors (strangely, not with every
toolchain).
So let's remove that again and take FFLAGS and friends from the
environment again.
commit 86af17db8526e629c2c02c6af1f1ce7db6f6ba6d
Author: Dr. Thomas Orgis <thomas.orgis@uni-hamburg.de>
Date: Thu Nov 12 12:44:39 2020 +0100
openblas: version 0.3.10
This updated fixes the build with gcc 10 (segfault in cblat1 test).
I did not go all the way to the current 0.3.12, as that would
need some hacking of chosen compiler flags. 0.3.13 should be
the next one.
This commit also fixes the ARCH → ARCH_ sed to change all
occurences on a line.
This fulfills pkg/55999 and was approved by wiz during freeze.
The previous state included our BLAS choice but the build
did not honour the selected libs. This patches the config
script to not search for LAPACK and BLAS, using our settings
instead. Hopefull we can get that option upstreamed.
Approved during freeze by wiz.
The previous state included our BLAS choice but the build
did not honour the selected libs. This patches the config
script to not search for LAPACK and BLAS, using our settings
instead. Hopefull we can get that option upstreamed.
Approved during freeze by wiz.
This is a bug fix release and contains the following two bug fixes:
* Fix indexing error for labelling in large (>2GB) arrays
* Only use retry_if_failed with recent pooch
2021-01-31: 1.0.7 release:
* ga.sor: new module Irish by László Németh
* lg.sor: new module Luganda by Phillip Samuel
* mr.sor: new module Marathi by Ankur Heramb Joshi
* mt.sor: new module Maltase by László Németh
* fi.sor: fix ordinal 0, bug report by Tuomas Hietala
* hu_Hung.sor:
- fix transliteration of Vörösmarty, bug report by Zoltán Óvári
- replace ü and ű by Viktor Kovács
- fix transliteration of dates with affixation
* th.sor: fix big (>999,999) numbers by Vladislav Kiper
* ua.sor: fix euro cent, bug report by Kty0mka
Benchmarking:
- Speed tests now call cpucycles() before setting resource limits. This
is important on platforms where cpucycles() needs to read files.
Verification:
- Support for SignExt and several more peephole optimizations, working
towards support for simpler symbolic-execution backend. Various
updates to work with angr8 and python3.
## 3.3.0
This is a production release that changes one behavior and fixes documentation
bugs.
The changed behavior is the treatment of `-e` and `-f` when given through
`BC_ENV_ARGS` or `DC_ENV_ARGS`. Now `bc` and `dc` do not exit when those options
(or their equivalents) are given through those environment variables. However,
`bc` and `dc` still exit when they or their equivalents are given on the
command-line.
Note that "DC_ENV_ARGS" will lead to dc exiting; this was new in 3.1.4
and I have reported that upstream.
Upstream changes, less bugfixes (and hence first/second do not make
sense; see the full NEWS.md in the sources if that upsets you):
## 3.2.5
This is a production release that fixes several bugs and adds a couple small
things.
The two most important bugs were bugs that causes `dc` to access memory
out-of-bounds (crash in debug builds). This was found by upgrading to `afl++`
from `afl`. Both were caused by a failure to distinguish between the same two
cases.
Fourth, the pseudo-random number now attempts to seed itself with `/dev/random`
if `/dev/urandom` fails.
## 3.2.0
This is a production release that has one bug fix and a major addition.
The major addition is a way to build a version of `bc`'s math code as a library.
This is done with the `-a` option to `configure.sh`. The API for the library can
be read in `./manuals/bcl.3.md` or `man bcl` once the library is installed with
`make install`.
This library was requested by developers before I even finished version 1.0, but
I could not figure out how to do it until now.
If the library has API breaking changes, the major version of `bc` will be
incremented.
## 3.1.4
This is a production release that fixes one bug, changes two behaviors, and
removes one environment variable.
The behavior that was changed is that `bc` now exits when given `-e`, `-f`,
`--expression` or `--file`. However, if the last one of those is `-f-` (using
`stdin` as the file), `bc` does not exit. If `-f-` exists and is not the last of
the `-e` and `-f` options (and equivalents), `bc` gives a fatal error and exits.
Next, I removed the `BC_EXPR_EXIT` and `DC_EXPR_EXIT` environment variables
since their use is not needed with the behavior change.
Finally, I made it so `bc` does not print the header, though the `-q` and
`--quiet` options were kept for compatibility with GNU `bc`.
