Version 1.0.12:
- Merged Pull Request 11] which simplifies conversion of a
profit matrix to a cost matrix, in the default case.
- Merged Pull Request 7 which fixes a message.
- Added more tests.
Version 0.41.0
This release adds the following major features:
* Diagnostics showing the optimizations done by ParallelAccelerator
* Support for profiling Numba-compiled functions in Intel VTune
* Additional NumPy functions: partition, nancumsum, nancumprod, ediff1d, cov,
conj, conjugate, tri, tril, triu
* Initial support for Python 3 Unicode strings
General Enhancements:
* armv7 support
* invert mapping b/w binop operators and the operator module
* First attempt at parallel diagnostics
* Adding NUMBA_ENABLE_PROFILING envvar, enabling jit event
* Support for np.partition
* Support for np.nancumsum and np.nancumprod
* Add location information to exceptions.
* Support for np.ediff1d
* Support for np.cov
* Support user pipeline class in with lifting
* string support
* Improve error message for empty imprecise lists.
* Enable overload(operator.getitem)
* Support negative indexing in tuple.
* Refactor Const type
* Optimized usage of alloca out of the loop
* Updates for llvmlite 0.26
* Add support for `np.conj/np.conjugate`.
* np.tri, np.tril, np.triu - default optional args
* Permit dtype argument as sole kwarg in np.eye
CUDA Enhancements:
* Add max_registers Option to cuda.jit
Continuous Integration / Testing:
* CI with Azure Pipelines
* Workaround race condition with apt
* Fix issues with Azure Pipelines
* Fix `RuntimeWarning: 'numba.runtests' found in sys.modules`
* Disable openmp in wheel building
* Azure Pipelines templates
* Fix cuda tests and error reporting in test discovery
* Prevent faulthandler installation on armv7l
* Fix CUDA test that used negative indexing behaviour that's fixed.
* Start Flake8 checking of Numba source
Fixes:
* Fix dispatcher to only consider contiguous-ness.
* Fix 3119, raise for 0d arrays in reductions
* Reduce redundant module linking
* Fix AOT on windows.
* Fix memory management of __cuda_array_interface__ views.
* Fix typo in error name.
* Fix the default unboxing logic
* Allow non-global reference to objmode() context-manager
* Fix global reference in objmode for dynamically created function
* CUDA_ERROR_MISALIGNED_ADDRESS Using Multiple Const Arrays
* Correctly handle very old versions of colorama
* Add 32bit package guard for non-32bit installs
* Fix with-objmode warning
* Fix label offset in call inline after parfor pass
* Fixes raising of user defined exceptions for exec(<string>).
* Fix error due to function naming in CI in py2.7
* Fixed TBB's single thread execution and test added for
* Allow matching non-array objects in find_callname()
* Change getiter and iternext to not be pure.
* Make ir.UndefinedType singleton class.
* Fix np.random.shuffle sideeffect
* Raise unsupported for kwargs given to `print()`
* Remove dead script.
* Fix stencil support for boolean as return type
* Fix handling make_function literals
* Add missing unicode != unicode
* Fix complex math sqrt implementation for large -ve values
* This adds arg an check for the pattern supplied to Parfors.
* Sets list dtor linkage to `linkonce_odr` to fix visibility in AOT
Documentation Updates:
* Update 0.40 changelog with additional PRs
* Tweak spacing to avoid search box wrapping onto second line
* Add note about memory leaks with exceptions to docs.
* Add FAQ on CUDA + fork issue.
* Update docs for argsort, kind kwarg partially supported.
* Added mention of njit in 5minguide.rst
* Fix parallel reduction example in docs.
* Fix broken link and mark up problem.
* Size Numba logo in docs in em units.
* just two typos
* Document string support
* Documentation for parallel diagnostics.
Version 0.9.12 Release Notes
New features:
- SkewedVoigtModel was added as built-in model
- Parameter uncertainties and correlations are reported for least_squares
- Plotting of complex-valued models is now handled in ModelResult class
- A model's independent variable is allowed to be an object
- Added usersyms to Parameters() initialization to make it easier to add custom functions and symbols
- the numdifftools package can be used to calculate parameter uncertainties and correlations for all solvers that do not natively support this
- emcee can now be used as method keyword-argument to Minimizer.minimize and minimize function, which allows for using emcee in the Model class
(Bug)fixes:
- asteval errors are now flushed after raising
- max_time and evaluation time for ExpressionModel increased to 1 hour
- loading a saved ModelResult now restores all attributes
- development versions of scipy and emcee are now supported
- ModelResult.eval() do no longer overwrite the userkws dictionary
- running the test suite requires pytest only
- improved FWHM calculation for VoigtModel
Bump to version 2.84
Construct a separate case for lexing '0' from other octal numbers
This avoids needing code to identify this case within the octal number
token generation code.
Ignore '_' within numbers so it can be used as a separator
Adopted from languages like Rust where this can be used to group
digits in thousands for base-10 or power of two for base 2,8,16 to
make them easier to read.
123_456_789 = 123456789
0xafff_1212 = 0xafff1212
The patch makes both the lexer and string conversions support this.
Add array initalizer tests
Ignore SIGTTOU/SIGTTIN. Adjust TSTP processing
When run as other than the master of a process group, nickle may not
get all of its work done before the controlling shell takes over the
tty and resets that control group. This means that any processing of
terminal parameters may generate SIGTTOU/SIGTTIN signals, which end up
re-entering the stop function and causing general chaos. Only SIG_IGN
appears to appease the kernel into dealing with this case correctly.
Make assignments to stdin/stdout/stderr affect C code
Use values of these variables in the C code.
Allow repeat initializers in implicit sized arrays for small dimensions
When initializing an implicitly sized multi-dimensional array, permit
repeat initializers ("...") to be used to fill out smaller dimensions.
This is done by requiring that the maximum number of elements in each
dimension be taken from an initalizer without a repeat element.
This also requires allowing repeat initializers to be useful for
resizable arrays, which required a change to the execution machinery
as well.
Remove incorrect tree walking in CompileSizeDimensions
CompileSizeDimensions was incorrectly interpreting the tree structure
for bare value nodes -- it is either passed an ARRAY, or a single
expression value, which can either be a comprehension, an empty
initializer or a simple expression. The simple expression case was
incorrectly looking for a '...' element to the left, which cannot
happen. It further recursed if it hadn't reached the end of the
dimensions, which is pointless as that case will generate an
error when the initializers are compiled later.
Debian-bug: #911926
Bump to version 2.83
debian: Run CI tests from installed location (/usr/share/nickle/test)
Install tests so they can be run later
debian: Bump standards-version to 4.2.1
Sign tarfiles. Check signature in debian/watch file.
debian: Add autopkgtest CI script
debian: Depend on debhelper version 11
debian: Add debian/watch file
Check to make sure debian package is up-to-date with upstream version.
debian: Add hardening flag to debian/rules
debian: Make debian/copyright conform to standard copyright-format
debian: Add Homepage entry in debian/control
debian: Bump compat to 11
debian: Remove trailing whitespace from debian/changelog
test: Remove spurious printf in check_exp test
Spelling fixes in nickle.1.in
expresion → expression
formating → formatting
Add AC_SYS_LARGEFILE to configure.ac
This ensures that the correct macros to get large file support enabled
are included in the config.h file.
Include debian/source/format in tarball
The tarball is used to build the debian package and without this file
the resulting bits have a lintian warning.
Update to version 2.82
Add .gitignore and .dir-locals.el files
Add tests for exp/log which cover a range of precisions
These check log and exp with mathematical identities to see if they
produce reasonable results at a range of precisions.
Add Math::e_value. Fix exp and log precision error.
e_value computes e to any desired precision.
exp was incorrectly truncating the value just before the last
computation step, which caused it to lose significant precision.
log was range-reducing to 0-1 instead of 2-3, which put it outside the
range of rapid convergence. It also needed another newtons method
iteration step.
Use mktime instead of timelocal
timelocal is a glibc alias for mktime and isn't provided by the musl C
library.
2.2:
This release is the result of 8 months of work with over 149 commits by
58 contributors. Highlights include:
- Add support for Python 3.7. This is the last release to support Python 2.
- Uniform random number generator (RNG) handling which defaults to global
RNGs but allows specification of a single RNG for all random numbers in NX.
- Improved GraphViews to ease subclassing and remove cyclic references
which caused trouble with deepcopy and pickle.
- New Graph method `G.update(H)`
v0.5.1
* The Python 3.6.7 and 3.7.1 point releases changed the standard
tokenizer module in a way that broke patsy. Updated patsy to work
with these point releases.
This release contains the following changes from 0.14.0:
Enhancements
* "Reflect" mode in transforms now works fine when an image dimension has
size 1
* img_as_float now allows single-precision (32-bit) float arrays to pass
through unmodified, rather than being up-converted to 64-bit
* Speed up rgb2gray computation
* The scikit-image viewer now works with different PyQt versions
* The cycle_spin function for enhanced denoising works single-threaded when
dask is not installed now
* scikit-image's io module will no longer inadvertently set the matplotlib
backend when imported
* Fix deprecated get keyword from dask in favor of scheduler
* Add missing cval parameter to threshold_local
API changes
* Remove deprecated dynamic_range in measure.compare_psnr
0.14.0
New Features
* Lookfor function to search across the library: skimage.lookfor.
* nD support for skimage.transform.rescale, skimage.transform.resize,
and skimage.transform.pyramid_* transforms.
* Chan-Vese segmentation algorithm.
* Manual segmentation with matplotlib for fast data annotation:
skimage.future.manual_polygon_segmentation,
skimage.future.manual_lasso_segmentation.
* Hysteresis thresholding: skimage.filters.apply_hysteresis_threshold.
* Segmentation with morphological snakes:
skimage.segmentation.morphological_chan_vese (2D),
skimage.segmentation.morphological_geodesic_active_contour (2D and 3D).
* nD support for image moments: skimage.measure.moments_central,
skimage.measure.moments_central, skimage.measure.moments_normalized,
skimage.measure.moments_hu. This change leads to 3D/nD compatibility
for many regionprops.
* Image moments from coordinate input: skimage.measure.moments_coords,
skimage.measure.moments_coords_central.
* Added 3D support to blob_dog and blob_log.
* Inertia tensor and its eigenvalues can now be computed outside of
regionprops; available in skimage.measure.inertia_tensor.
* Cycle-spinning function for approximating shift-invariance by averaging
results from a series of spatial shifts: skimage.restoration.cycle_spin.
* Haar-like feature: skimage.feature.haar_like_feature,
skimage.feature.haar_like_feature_coord,
skimage.feature.draw_haar_like_feature.
* Data generation with random_shapes function: skimage.draw.random_shapes.
* Subset of LFW (Labeled Faces in the Wild) database:
skimage.data.cbcl_face_database.
* Fully reworked montage function (now with a better padding behavior):
skimage.util.montage.
* YDbDr colorspace conversion routines: skimage.color.rgb2ydbdr,
skimage.color.ydbdr2rgb.
Improvements
* VisuShrink method for skimage.restoration.denoise_wavelet.
* New max_ratio parameter for skimage.feature.match_descriptors.
* skimage.transform.resize and skimage.transform.rescale have a new
anti_aliasing option to avoid aliasing artifacts when down-sampling images.
* Support for multichannel images for skimage.feature.hog.
* Non-local means denoising (skimage.restoration.denoise_nl_means) has a new
optional parameter, sigma, that can be used to specify the noise standard
deviation. This enables noise-robust patch distance estimation.
* New alignment parameter in skimage.feature.plot_matches.
* New seed parameter in skimage.transform.probabilistic_hough_line.
* Various performance improvements.
API Changes
* skimage.util.montage. namespace has been removed, and
skimage.util.montage.montage2d function is now available as
skimage.util.montage2d.
* skimage.morphology.binary_erosion now uses True as border value, and is
now consistent with skimage.morphology.erosion.
Deprecations
* freeimage plugin has been removed from skimage.io.
* skimage.util.montage2d is deprecated and will be removed in 0.15.
Use skimage.util.montage function instead.
* skimage.novice is deprecated and will be removed in 0.16.
* skimage.transform.resize and skimage.transform.rescale have a new
anti_aliasing option that avoids aliasing artifacts when down-sampling
images. This option will be enabled by default in 0.15.
* regionprops will use row-column coordinates in 0.16.
* skimage.morphology.remove_small_holes min_size argument is deprecated
and will be removed in 0.16. Use area_threshold instead.
Upstream changes:
1.13
30 Oct 2018
- Add function pl_translate(), by request.
- Turn README into README.md. With MetaCPAN replacing CPAN,
there is no reason to have a plain-text README anymore
(MetaCPAN doesn't treat the file as special).
- Add a CONTRIBUTING.md file, which is becoming a thing.
- New test file for pl_translate(), and upgrade test files
that for some reason were still using Test::Simple to Test::More.
- Added an :all tag. Why not.
NumPy 1.15.4:
This is a bugfix release for bugs and regressions reported following the 1.15.3
release. The Python versions supported by this release are 2.7, 3.4-3.7. The
wheels are linked with OpenBLAS v0.3.0, which should fix some of the linalg
problems reported for NumPy 1.14.
NumPy 1.15.3:
This is a bugfix release for bugs and regressions reported following the 1.15.2
release. The Python versions supported by this release are 2.7, 3.4-3.7. The
wheels are linked with OpenBLAS v0.3.0, which should fix some of the linalg
problems reported for NumPy 1.14.
Upstream changes:
1.999815 2018-10-19
* Move bitwise operators signed and, signed or, and signed xor from
lib/Math/BigInt/CalcEmu.pm into lib/Math/BigInt/Lib.pm. The file
lib/Math/BigInt/CalcEmu.pm is no longer needed and thus removed.
1.999814 2018-10-01
* Add to_base() and from_base() to Math::BigInt and corresponding library
methods _to_base() and _from_base() to Math::BigInt::Lib. This was inspired
by CPAN RT #122681.
* Fix Makefile.PL to reflect that Test::More is only needed for testing, not
for building.
* In the documentation for each of the to_(bin|hex|oct|bytes) methods, add a
reference to the corresponding from_(bin|hex|oct|bytes) method.
Highlights:
Missing values in features, represented by NaNs, are now accepted in
column-wise preprocessing such as scalers. Each feature is fitted
disregarding NaNs, and data containing NaNs can be transformed. The
new impute module provides estimators for learning despite missing
data.
ColumnTransformer handles the case where different features or columns
of a pandas.DataFrame need different preprocessing. String or pandas
Categorical columns can now be encoded with OneHotEncoder or
OrdinalEncoder.
TransformedTargetRegressor helps when the regression target needs to
be transformed to be modeled. PowerTransformer and KBinsDiscretizer
join QuantileTransformer as non-linear transformations.
Added sample_weight support to several estimators (including KMeans,
BayesianRidge and KernelDensity) and improved stopping criteria in
others (including MLPRegressor, GradientBoostingRegressor and
SGDRegressor).
This release is also the first to be accompanied by a Glossary of
Common Terms and API Elements.
What is new in gsl-2.5:
** doc bug fix in binomial distribution figure
** added Wishart distribution
** added new module for digital filtering (gsl_filter); current filters include:
Gaussian filter
median filter
recursive median filter
impulse detection filter
** added new module for moving window statistics (gsl_movstat)
** added statistics functions:
gsl_stats_median()
gsl_stats_select()
gsl_stats_mad()
gsl_stats_mad0()
gsl_stats_Sn_from_sorted_data()
gsl_stats_Qn_from_sorted_data()
gsl_stats_gastwirth_from_sorted_data()
gsl_stats_trmean_from_sorted_data()
** added Romberg integration (gsl_integration_romberg)
** bug fix in deprecated functions gsl_multifit_wlinear_svd and
gsl_multifit_wlinear_usvd
** documention corrected to state that gsl_sf_legendre functions do
not include Condon-Shortley phase by default
** bug fix in exponential fitting example when using larger number
of points
** changed internal workspace inside gsl_spmatrix to a union to
avoid casting
** bug fixes in ode-initval2 for very rare solver crashing cases
** add histogram2d figure to manual (was missing in 2.4)
** bug fix in gsl_spmatrix_add for duplicate input arguments
** add support for negative arguments nu in gsl_sf_bessel_Jnu and
gsl_sf_bessel_Ynu
** better texinfo documentation for gsl_sf_hyperg functions
** fix vector and matrix fread/fwrite testing on windows systems
when tmpfile() fails
** fix for rstat/test.c on PPC64
NumPy 1.15.2:
This is a bugfix release for bugs and regressions reported following the 1.15.1
release.
* The matrix PendingDeprecationWarning is now suppressed in pytest 3.8.
* The new cached allocations machinery has been fixed to be thread safe.
* The boolean indexing of subclasses now works correctly.
* A small memory leak in PyArray_AdaptFlexibleDType has been fixed.
The Python versions supported by this release are 2.7, 3.4-3.7. The wheels are
linked with OpenBLAS v0.3.0, which should fix some of the linalg problems
reported for NumPy 1.14.
Changes:
* Reintroduced support for Yahoo! price, dividends, and splits data
for stocks and currency pairs.
* A new data connector for data provided by Alpha Vantage was
introduced to obtain Foreign Exchange (FX) data
* A new data connector for data provided by Alpha Vantage was
introduced to obtain historical time series data.
* A new data connector for data provided by Alpha Vantage was
introduced to obtain sector performance data.
* A new data connector for data provided by Alpha Vantage was
introduced to obtain real-time Batch Stock Quotes.
* Immediate deprecation of Google finance and Morningstar for
historical price data, as these API endpoints are no longer
supported by their respective providers. Alternate methods are
welcome via pull requests, as PDR would like to restore these
features.
* Removal of EDGAR, which was deprecated in v0.6.0.
### 1.1.8 / 20.08.2018
Author: VitaliyAdamkov <adamkov@tex.ua>
Date: Mon Aug 20 09:48:31 2018 +0300
* Cancel :lazy usage
* Use lazy select to speed up a little
* Omit rails :try usage
* stub for :postread_worksheet method
* sometimes it selects empty array..
Author: 545ch4 <s@rprojekt.org>
Date: Wed Mar 28 15:33:04 2018 +0200
* [ruby-2.4] Fix weird first line of spreadsheet.gemspec
* Doesn't seem to be a valid .gemspec command/field.
An R interface to the NetCDF file format designed by Unidata for
efficient storage of array-oriented scientific data and descriptions.
The R interface is closely based on the C API of the NetCDF library,
and it includes calendar conversions from the Unidata UDUNITS library.
The current implementation supports all operations on NetCDF datasets
in classic and 64-bit offset file formats, and NetCDF4-classic format
is supported for reading and modification of existing files.
From Kai-Uwe Eckhardt, updated as the previous distfile wasn't available.
PR pkg/51607
0.79 - 29 Aug 2018, H.Merijn Brand
* Environmant control hardening
* ReadData () docs (opening from scalar ref: issue#21)
* Check for XLSX content on opening from plain scalar content
* Fix typo in ss-dup-tk.pl (PRC, Branislav Zahradnik)
* Allow source from command line option in xls2csv (PRC, Branislav Zahradnik)
* Allow --in-sep=tab in xlscat/xlsgrep
* Add --hash to xlscat/xlsgrep
* Help and usage for examples/xls2csv PR#23)
0.71 2018-08-28
[ADDED]
- forfactored { ... } a,b loop n=a..b setting $_=n, @_=factor(n)
- forsquarefree { ... } a,b as forfactored, but only square-free n
- forsemiprimes { ... } a,b as forcomposites, but only semiprimes
- random_factored_integer(n) random [1..n] w/ array ref of factors
- semiprime_count([lo],hi) counts semiprimes in range
[FIXES]
- Monolithic sieves beyond 30*2^32 (~ 1.2 * 10^11) overflowed.
- is_semiprime was wrong for five small values since 0.69. Fixed.
[FUNCTIONALITY AND PERFORMANCE]
- is_primitive_root much faster (doesn't need to calulate totient,
and faster rejection when n has no primitive root).
- znprimroot and znorder use Montgomery, 1.2x to 2x faster.
- slightly faster sieve_range for native size inputs (use factor_one).
- bin/primes.pl faster for palindromic primes and works for 10^17
[OTHER]
- Added ability to use -DBENCH_SEG for benchmarking sieves using
prime_count and ntheory::_segment_pi without table optimizations.
- Reorg of main factor loop. Should be identical from external view.
- Internal change to is_semiprime and is_catalan_pseudoprime.
Upstream changes:
Version 0.4-91
Fixed a bug in heatmap.circular
Version 0.4-8 2014/08/08
minusPiPlusPi now accepts missing values
Added the start.sep argument to plot.circular and points.circular. Also the internal function PointsCircularRad has changed the API.
