Roadmap
-
-
Bug-fix release of 0.7.0:
- #739
- scipy.weave.converters.blitz fails with g++4.3
- #746
- Scipy 0.6.0 release is missing setup,py files for stsci
- #786
- scipy.stats problems on 64 bits python 2.4 (cython)
- #823
- betainc function in scipy.special
- #859
- spatial.distance.squareform - arg 'force' behaves inconsistently
- #871
- bicgstab fails on Win32
- #880
- scipy.signal.lfilter segfaults if zi is an empty array
- #883
- Broken: scipy.io.mmread with scipy.sparse.lil_matrix
- #885
- io.loadmat and io.savemat fail with 0-length fields
- #887
- stats.binom.ppf crashes
- #891
- Memory leaking
- #912
- scipy.signal.iirdesign segfaults in scipy 0.6.0
- #927
- ellipj segfaults for m=nan
- #944
- [Ubuntu Jaunty]python-scipy install SyntaxWarning
- #947
- wrong min max in stats.describe
- #953
- scipy.special test failures on mac os x
- #958
- 2.6 binaries for windows
- #959
- Scipy mac os x binary is not universal
- #960
- scipy mac os x binary should not depend on gfortran
- #1030
- curve_fit not available in latest build (0.7.1)
-
Release schedule
The SciPy? 0.8.x development has now started. Here is the preliminary release schedule:
5/25 0.8.0b1 released 7/13 0.8.x branch frozen 7/15 0.8.0rc1 tagged 7/16 0.8.0rc1 announced, binaries posted to sourceforge 7/24 0.8.0 tagged 7/27 0.8.0 announced, binaries posted to sourceforge
SciPy 0.8.0 Release Notes
SciPy 0.8.0 is the culmination of 6 months of hard work. It contains many new features, numerous bug-fixes, improved test coverage and better documentation. There have been a number of deprecations and API changes in this release, which are documented below. All users are encouraged to upgrade to this release, as there are a large number of bug-fixes and optimizations. Moreover, our development attention will now shift to bug-fix releases on the 0.8.x branch, and on adding new features on the development trunk. This release requires Python 2.4 - 2.6 and NumPy 1.3 or greater.
Please note that SciPy is still considered to have "Beta" status, as we work toward a SciPy 1.0.0 release. The 1.0.0 release will mark a major milestone in the development of SciPy, after which changing the package structure or API will be much more difficult. Whilst these pre-1.0 releases are considered to have "Beta" status, we are committed to making them as bug-free as possible. For example, in addition to fixing numerous bugs in this release, we have also doubled the number of unit tests since the last release.
However, until the 1.0 release, we are aggressively reviewing and refining the functionality, organization, and interface. This is being done in an effort to make the package as coherent, intuitive, and useful as possible. To achieve this, we need help from the community of users. Specifically, we need feedback regarding all aspects of the project - everything - from which algorithms we implement, to details about our function's call signatures.
Python 3.0
Python 3.0 is not supported at all; it requires NumPy to be ported to Python 3.0. This requires immense effort, since a lot of C code has to be ported. The transition to 3.0 is still under consideration; currently, we don't have any timeline or roadmap for this transition.
Swapping inputs for correlation functions (scipy.signal)
Concern correlate, correlate2d, convolve and convolve2d. If the second input is larger than the first input, the inputs are swapped before calling the underlying computation routine. This behavior is deprecated, and will be removed in scipy 0.9.0.
DCT support (scipy.fftpack)
New realtransforms have been added, namely dct and idct for Discrete Cosine Transform; type I, II and III are available, for both single and double precision.
Single precision support for fft functions (scipy.fftpack)
fft functions can now handle single precision inputs as well: fft(x) will return a single precision array if x is single precision.
correlation functions now implements the usual definition (scipy.signal)
The outputs should now correspond to their matlab and R counterparts, and do what most people expect if the old_behavior=False argument is passed:
- correlate, convolve and their 2d counterparts do not swap their inputs depending on their relative shape anymore.
- correlation functions now conjugate their second argument while computing the slided sum-products, which correspond to the usual definition of correlation
ARPACK-based sparse SVD
A naive implementation of SVD for sparse matrices is available in scipy.sparse.linalg.eigen.arpack. It is based on using an symmetric solver on <A, A>, and as such may not be very precise.
Removed features
scipy.stsci: the package was removed
Bug-fixes
- #490
- weave.test(5) failures
- #503
- bug in special function iv for large argumen
- #623
- scipy.special.jv incorrect for certain argument(s)
- #640
- special package numerical errors
- #679
- struve function has discontinuity
- #733
- Implement dct
- #759
- errors in special.hyp2f1 for some values as used in stats.rdist
- #794
- linalg.eig segmentation fault on unpickled arrays
- #803
- pbdv produces invalid output
- #805
- Faster implementation of argsreduce in scipy.stats.distributions
- #806
- More accurate upper tail quantiles for small survival probabilities of the normal distribution
- #807
- More accurate tail probability and quantile for exponential distribution
- #829
- Python2.6 issue in test_implicit (test_odr.TestODR)
- #830
- firwin docstring suggestion
- #836
- scons build produces buggy scipy
- #841
- ImportError: No module named convolve
- #844
- Orthogonal polynomials inheritance appears to be broken
- #852
- special.jn_zeros numerical errors for large s
- #853
- scipy.special.jv incorrect values for complex x & negative non-integer order
- #854
- special.jv, yv etc. return 0 when they should return nan
- #855
- Python2.6 test_add_function_ordered Key error
- #858
- faster sparse matrices
- #860
- Ridder's method undercounts function evaluations
- #861
- The progress table printed by sp.integrate.romberg shows the wrong step size
- #867
- scipy.stats.mstats.mquantiles - limit keyword produces unexpected results
- #870
- dok_matrix can't handle setting unset elements to zero (was: KDtree sparse distance matrix doesn't handle distance=0 case)
- #874
- incorrect slope standard error in stats.linregress
- #877
- Depreciation of samplevar in scipy.stats
- #890
- LOBPCG uses depreciated write_array
- #897
- Fix jacobian in test_minpack.TestFSolve and style fixes
- #899
- scipy.special.ndtr / ndtri broken under OSX, Python 2.6
- #900
- exponential integral Ei of complex values
- #906
- Depreciate erfc function defined in scipy/stats/stats.py
- #908
- Enhanced harmonic mean function
- #911
- lfilter leaks memory
- #917
- column slicing is broken for lilmatrix
- #918
- New regression test for reading compressed matlab streams.
- #924
- Visual Studio 9.0 compile error
- #925
- lfilter segfaults for object arrays
- #929
- lfilter should support extended precision
- #932
- Wiener filter docstring states about a median filter window
- #946
- cephes/isnan was broken by r5693
- #951
- Website: The svn link on the weave documentation page is broken
- #971
- docstring of freqz
- #972
- factorial2 docstring typo
- #975
- special.gammaincinv(x,y) fails when y=0.25
- #978
- nbinom._ppf incorrect for n<1 at integer values
- #985
- Problem with 1-D hilbert transform on N-D arrays
- #986
- binom_test(50, 100, 0.1) gives bogus results
- #992
- scipy.stats.norm generates an error when given an array_like for scale argument
- #1000
- stats.chi2.cdf and ppf return nan with 0<df<1
- #1006
- hilbert2 being slightly verbose
- #1017
- normalize function from filter_design.py broken
- #1019
- distance does not take Unicode strings as metric names
- #1023
- polyint documentation is confusing
- #1025
- stats.chisquare needs ddof
- #1032
- von Mises CDF broken outside -pi..pi
- #1033
- Add description of how to merge docstrings
- #1037
- SciPy.linalg.basic.norm raises ZeroDivisionError for ord=0 argument
- #1041
- chi2.pdf numerically unstable for large values of df
- #1044
- [Patch] Typo fixes in scipy.optimize.*
- #1045
- Build error in scipy/special/cephes/scipy_iv.c
- #1047
- SuperLU/SRC/scomplex.h
- #1049
- Implement Lambert W special function
- #1050
- skellam distribution
- #1051
- Unnecessary debug output in scipy.io.wavfile
- #1056
- Clarify SLSQP license status
- #1059
- io.loadmat: "Dtype must be a numpy data-type"
- #1060
- sici (sine/cosine integrals) function returns wrong values for +-inf input
- #1066
- precision loss in cdf in stats.hypergeom for large samples
- #1068
- scipy.special.hermitenorm exception fix
- #1070
- test_lambertw.test_values segfaults
- #1074
- newton(lambda x: x-1, 0) return 0, not 1
- #1077
- kmeans chokes on data with lots of features
- #1088
- signal.convolve wrong variable names
- #1093
- signal.hilbert incorrect axis handling for 2d input
- #1105
- Logarithmic chirp frequency sweep incorrect
- #1106
- Cannot import cKDTree under windows.
- #1110
- Mysterious 'end' in ss2tf in ltisys.py
- #1112
- scipy.signal.lsim2: pass additional keyword args to odeint.
- #1115
- Bugfix in leastsq
- #1116
- Spelling error in signal/filter_design.py
- #1121
- odeint docstring does not give the default values for atol and rtol
-
