| 1 |
------------------------------------------------------------------------------- |
| 2 |
MDP-3.6: |
| 3 |
2020-04-18: NEW: support libsvm for Python 3 libsvm has been updated to be |
| 4 |
compatible with Python >= 3.4. Python 2 support has been dropped, |
| 5 |
so we don't test it anymore. |
| 6 |
2020-04-18: ERF: restore Python 2 compatibility and add a matrix entry for it |
| 7 |
in travis |
| 8 |
2019-18-12: FIX: SciPy 1.4 support. |
| 9 |
2019-13-12: NEW: VartimeSFANode supports SFA on data with non-constant time |
| 10 |
increments. This yields an SFA implementation that is robust |
| 11 |
against missing data. |
| 12 |
2019-01-20: NEW: Added a new family of expansion nodes, including Legendre, |
| 13 |
Hermite and Chebyshev polynomials. A facility for easily including |
| 14 |
recursively defined 1-dimensional expansion functions powers these |
| 15 |
nodes, applying them to N-dimensional data by tensor basis |
| 16 |
construction. The new expansion nodes are numerically more stable |
| 17 |
than ordinary monomials and permit higher expansion degrees. |
| 18 |
Examples where numerically stable expansion is beneficial: |
| 19 |
https://github.com/NiMlr/pyfunctionbases |
| 20 |
https://arxiv.org/abs/1805.08565, section 4.2.1, Figure 11 and A.2. |
| 21 |
2019-01-19: FIX: pytest Python 3.7 compatibility |
| 22 |
2018-07-18: FIX: PEP 479 compatibility |
| 23 |
2018-06-15: NEW: Added SFA-based supervised learning, specifically graph based |
| 24 |
SFA nodes GSFANode and iGSFANode and hierarchical GSFA (HGSFA). |
| 25 |
For further information, see: |
| 26 |
Escalante-B A.-N., Wiskott L, "How to solve classification and |
| 27 |
regression problems on high-dimensional data with a supervised |
| 28 |
extension of Slow Feature Analysis". Journal of Machine Learning |
| 29 |
Research 14:3683-3719, 2013. |
| 30 |
http://www.jmlr.org/papers/volume14/escalante13a/escalante13a.pdf |
| 31 |
https://github.com/AlbertoEsc/cuicuilco |
| 32 |
2018-05-29: DOC: Revamped the documentation for Sphinx-autodoc compatibility. |
| 33 |
MDP documentation is now available on readthedocs: |
| 34 |
https://mdpdocs.readthedocs.io/en/latest |
| 35 |
2018-01-18: NEW: Added solvers to SFA-node that are robust against rank deficit |
| 36 |
in the covariance matrix. When data contained linear redundancies, |
| 37 |
like easily introduced by e.g. insufficient cropping of black bars |
| 38 |
of a video, this usually resulted in |
| 39 |
SymeigException ('Covariance matrices may be singular'). |
| 40 |
Now it provides simple instructions for robust processing of such |
| 41 |
data by setting a new flag in SFANode: |
| 42 |
rank_deficit_method can be 'none' (default), 'reg', 'pca', 'svd', |
| 43 |
leveraging robust solving via regularization, PCA, SVD or LDL |
| 44 |
decomposition respectively. (LDL requires SciPy >= 1.0) |
| 45 |
2017-03-11: DOC: changed reference to new mailing list |
| 46 |
old: [email protected] |
| 47 |
new: [email protected] |
| 48 |
2017-03-06: NEW: Added online mode to enable use of MDP in reinforcement |
| 49 |
learning. This notably includes incremental SFA as described in |
| 50 |
"Varun Raj Kompella, Matthew Luciw, and Juergen Schmidhuber (2012). |
| 51 |
Incremental Slow Feature Analysis: Adaptive Low-Complexity Slow |
| 52 |
Feature Updating from High-Dimensional Input Streams. |
| 53 |
Neural Computation." |
| 54 |
https://www.mitpressjournals.org/doi/abs/10.1162/NECO_a_00344 |
| 55 |
New nodes and flows: |
| 56 |
OnlineNode, OnlineFlow, CircularFlow, OnlineCenteringNode, |
| 57 |
OnlineTimeDiffNode, OnlineFlowNode, CircularOnlineFlowNode, |
| 58 |
OnlineLayer, CloneOnlineLayer, SameInputOnlineLayer, MCANode, |
| 59 |
CCIPCANode, IncSFANode |
| 60 |
|
| 61 |
------------------------------------------------------------------------------- |
| 62 |
MDP-3.5: |
| 63 |
2016-03-08: fix use of str in biflow indexing |
| 64 |
2016-03-08: remove obsolete scripts |
| 65 |
2016-03-07: Add wrapper function for exec Old Python versions don't allow exec |
| 66 |
statement in nested functions: Traceback (most recent call last): |
| 67 |
File "<string>", line 1, in <module> |
| 68 |
File "mdp/__init__.py", line 131, in <module> |
| 69 |
from . import utils |
| 70 |
File "mdp/utils/__init__.py", line 29, in <module> |
| 71 |
from .slideshow import (basic_css, slideshow_css, HTMLSlideShow, |
| 72 |
File "mdp/utils/slideshow.py", line 23, in <module> |
| 73 |
from . import templet |
| 74 |
File "mdp/utils/templet.py", line 196 |
| 75 |
exec(code, globals, kw) SyntaxError: unqualified exec is not |
| 76 |
allowed in function 'expand' it is a nested function |
| 77 |
|
| 78 |
------------------------------------------------------------------------------- |
| 79 |
MDP-3.4: |
| 80 |
2016-03-02: Make use of setuptools instead of distutils While we resisted as |
| 81 |
long as possible, it is now time to move to use setuptools |
| 82 |
officially. This makes MDP installable properly with pip. This |
| 83 |
means that our dependencies are now explicitly declared |
| 84 |
(numpy and future). Optional dependencies are also declared and can |
| 85 |
be installed using the "pip -e MDP[optionaldep]" command, where |
| 86 |
optionaldep can be scipy, or scikit-learn or ... Moreover, |
| 87 |
automatic testing using "python setup.py test" now works, which |
| 88 |
makes MDP finally conform this non-written standard and makes it |
| 89 |
easy to run tests for clueless people. |
| 90 |
2016-03-02: add new modules in sklearn wrappers |
| 91 |
2016-03-02: Remove self-contained py.test script... The use of the |
| 92 |
self-contained script for py.test has been deprecated for a while |
| 93 |
and has stop working. We now require py.test to be installed to run |
| 94 |
the tests. |
| 95 |
2016-03-01: remove obsolete stuff related to py3 conversion with lib2to3 |
| 96 |
2016-03-01: Disable support for shogun > 1 it would be too much work to enable |
| 97 |
support for newer shogun versions... |
| 98 |
2016-03-01: disable failing monkeypatching test for pp... |
| 99 |
.. the pp monkeypatching trick is not needed anymore anyways... |
| 100 |
2016-03-01: futurize stage 2: code is now py2&3 compatible |
| 101 |
2016-03-01: futurize stage1: code should still be python2.6 compatible |
| 102 |
2016-03-01: compatibility fixes for numpy 1.11.., |
| 103 |
...numpy now refuses to index arrays using floats and default dtype |
| 104 |
for zeros and ones has been changed from integers to floats. |
| 105 |
2016-03-01: several new sklearn autogenerated nodes need additional arguments |
| 106 |
to be tested |
| 107 |
2014-04-27: FIX: fixed broken inclusion of CSS data |
| 108 |
2014-03-14: FIX: hacky fix for the unicode issue in bimdp autogen (fix for Bug |
| 109 |
#11) |
| 110 |
2013-09-03: NEW: added extension setup and teardown function decorator |
| 111 |
2013-09-03: DOC: some small updates and fixes in docstrings and comments |
| 112 |
2013-08-31: FIX: cStringIO does not support Unicode |
| 113 |
`autogen` was breaking in nodes with Unicode characters in their |
| 114 |
docstring. We traced this down to cStringIO, which unlike StringIO |
| 115 |
does not support unicode. |
| 116 |
2013-08-31: FIX: wrap sklearn algs in modules ending with "_" For historical |
| 117 |
reasons we were not wrapping sklearn algs in modules with names |
| 118 |
ending with "_". This restriction prevented us from wrapping algs |
| 119 |
that we could have wrapped. We lift it now, since it does not seem |
| 120 |
to have any bad side effects anymore. A proper validation would be |
| 121 |
to really test that we can run the algs. |
| 122 |
2013-05-08: NEW: update local copy of py.test to version 2.3.4 |
| 123 |
2013-04-16: FIX: made code and documentation consistent regarding the parameter |
| 124 |
k in invert_exp_funcs2 |
| 125 |
2013-04-16: DOC: indicated what expansion failed in the tests of |
| 126 |
GeneralExpansionNode |
| 127 |
2013-04-16: DOC: added an example to GeneralExpansionNode and made the |
| 128 |
documentation more explicit |
| 129 |
2012-12-07: FIX: make pip work without explicit URL. now we point directly to |
| 130 |
the tar archive in the download_url field of the setup.py file. |
| 131 |
2012-11-20: DOC: moved coarse_limit description under FastICA and removed |
| 132 |
signature (present elsewhere) |
| 133 |
2012-11-20: ENH: FastICA - test having some coarse_limit |
| 134 |
2012-11-20: ENH: FastICA -- primary_limit -> coarse_limit + default to None |
| 135 |
Introduced to make things a bit clearer since there is already |
| 136 |
"fine tuning" stage where optimization switches to used_g upon |
| 137 |
reaching the limit. Calling the limit "coarse" instead of |
| 138 |
"primary/secondary" would help to preserve original messages/logic |
| 139 |
thus not confusing existing users, while still allowing to kick in |
| 140 |
gFine before reaching the limit |
| 141 |
2012-11-20: ENH: 2 stage optimization for FastICA found in current AFNI |
| 142 |
2012-10-04: DOC: bumped version number after release ;) |
| 143 |
2012-09-21: OTH: updated CHANGES and COPYRIGHT |
| 144 |
2012-09-21: NEW: add support for newest sklearn algorithms |
| 145 |
|
| 146 |
------------------------------------------------------------------------------- |
| 147 |
MDP-3.3: |
| 148 |
2012-09-19: FIX: fix error in automatic testing for MultinomialNB. |
| 149 |
2012-09-19: ERF: make sklearn nodes automatic testing more robust The previous |
| 150 |
solution was actually ignoring the special definitions in NODES. |
| 151 |
|
| 152 |
2012-09-19: FIX: disable pp support if server can not start Being able to |
| 153 |
import pp is not enough to be sure that pp works. For example in a |
| 154 |
low memory situation, the following can happen: |
| 155 |
|
| 156 |
>>> import pp |
| 157 |
>>> server = pp.Server() |
| 158 |
[...] OSError: [Errno 12] Cannot allocate memory |
| 159 |
|
| 160 |
This fix just disables pp support if the server can not be |
| 161 |
started. |
| 162 |
|
| 163 |
2012-06-18: FIX: Fix wrapping of sklearn 0.11 classifiers |
| 164 |
2012-04-17: FIX: make test_SFA2Node even more robust |
| 165 |
2012-04-16: FIX: make FastICANode test more robust |
| 166 |
2012-04-16: FIX: make test_SFA2Node more robust |
| 167 |
2012-04-05: FIX: fix pp_tests when run multiple times. pp tests were failing |
| 168 |
when run twice in a row. hugly work-around, but it seems to |
| 169 |
work... |
| 170 |
|
| 171 |
2012-04-05: FIX: fixed broken test_reload. test_reload was failing when called |
| 172 |
twice in a row. |
| 173 |
|
| 174 |
2012-04-05: FIX: fix random seed tests. The tests were failing when called |
| 175 |
twice in a row: |
| 176 |
>>> import mdp |
| 177 |
>>> mdp.test() |
| 178 |
>>> mdp.test() the first call was working, the second one was |
| 179 |
giving failures. |
| 180 |
|
| 181 |
2012-04-01: ERF: added tests for learning of bias parameters |
| 182 |
2012-03-26: FIX: replace third remaing test for pp_monkeypatch_dirname |
| 183 |
Hopefully this will fix test suite failures. |
| 184 |
|
| 185 |
2012-03-22: FIX: Decrease the noise level in the DiscreteHopfieldClassifier. |
| 186 |
2012-03-22: FIX: honor MDP_DISABLE_SHOGUN env variable |
| 187 |
2012-03-19: FIX: fix left-over directories from testing pp. I do not know why, |
| 188 |
but this simple change fixes the leftover directories problem when |
| 189 |
testig with python-pp and pp monkey-patching. It should have |
| 190 |
worked even as it was before, but apparently some race condition |
| 191 |
happens. |
| 192 |
|
| 193 |
2012-03-06: FIX: fix determinant of random rotation matrix determinant sign |
| 194 |
was wrong if dimensions of rotation matrix were odd. Thanks to |
| 195 |
Philip DeBoer. Actual Fix. |
| 196 |
|
| 197 |
2012-03-06: FIX: fix determinant of random rotation matrix determinant sign |
| 198 |
was wrong if dimensions of rotation matrix were odd. Thanks to |
| 199 |
Philip DeBoer. Failing Test. |
| 200 |
|
| 201 |
2012-02-13: ENH: remove duplicated and overly verbose code |
| 202 |
2012-02-13: FIX: remove FastICA stabilization from tests |
| 203 |
2012-02-13: FIX: remove unused parameter stabilization in FastICA. |
| 204 |
2012-01-19: NEW: added new sklearn algorithms wrapping We now wrap 93 sklearn |
| 205 |
algorithms! |
| 206 |
|
| 207 |
2012-01-19: FIX: fix another imcompatibility with sklearn 0.10 Although |
| 208 |
EllipticEnvelop is derived from sklearn.base.ClassifierMixin, it |
| 209 |
is not a classifier. It is actually a "predictor". Added a check |
| 210 |
in the sklearn wrappers. |
| 211 |
|
| 212 |
2012-01-19: FIX: fix sklearn wrappers for version 0.10 New sklearn version |
| 213 |
introduces classifiers and predictors mixin classes that do not |
| 214 |
have a 'fit' method. Typical error: |
| 215 |
|
| 216 |
AttributeError: type object 'OutlierDetectionMixin' has no |
| 217 |
attribute 'fit' |
| 218 |
|
| 219 |
Just added a check that the method is really present before |
| 220 |
wrapping. |
| 221 |
|
| 222 |
2012-01-12: FIX: fix failing test for no eigenvalues left problem Check that |
| 223 |
PCANode now raises the right exception. |
| 224 |
|
| 225 |
2012-01-12: FIX: add useful exception in case of no eigenvalues left. Check |
| 226 |
for the condition explained in |
| 227 |
b0810d72ce11925e1db6204c3a20bdfc77741a82 and raise a nicely |
| 228 |
formatted exception: |
| 229 |
|
| 230 |
Traceback: |
| 231 |
[...] File ".../mdp/nodes/pca_nodes.py", line 223, in |
| 232 |
_stop_training |
| 233 |
' var_abs=%e!'%self.var_abs) NodeException: No eigenvalues |
| 234 |
larger than var_abs=1.000000e-15! |
| 235 |
|
| 236 |
2012-01-12: OTH: added failing test for no eigenvalues left problem. When |
| 237 |
PCANode is set to use SVD and automatic dimensionality reduction |
| 238 |
it may happen that after removing directions corresponding to |
| 239 |
eigenvalues smaller than var_abs (1e-12 default), nothing is left. |
| 240 |
This happens for example if the data is a matrix of (almost) |
| 241 |
zeros. The error looks like this: Traceback (most recent call |
| 242 |
last): |
| 243 |
[...] File ".../mdp/nodes/pca_nodes.py", line 220, in |
| 244 |
_stop_training |
| 245 |
d = d[ d / d.max() > self.var_rel ] ValueError: zero-size array |
| 246 |
to ufunc.reduce without identity |
| 247 |
|
| 248 |
2012-01-03: FIX: old joblib breaks imports from sklearn.decomposition |
| 249 |
>>> import sklearn.decomposition Traceback (most recent call |
| 250 |
last): |
| 251 |
File "<stdin>", line 1, in <module> |
| 252 |
File |
| 253 |
"/usr/lib/pymodules/python2.6/sklearn/decomposition/__init__.py", |
| 254 |
line 8, in <module> |
| 255 |
from .sparse_pca import SparsePCA, MiniBatchSparsePCA |
| 256 |
File |
| 257 |
"/usr/lib/pymodules/python2.6/sklearn/decomposition/sparse_pca.py", |
| 258 |
line 10, in <module> |
| 259 |
from .dict_learning import dict_learning, dict_learning_online |
| 260 |
File |
| 261 |
"/usr/lib/pymodules/python2.6/sklearn/decomposition/dict_learning.py", |
| 262 |
line 17, in <module> |
| 263 |
from ..externals.joblib import Parallel, delayed, cpu_count |
| 264 |
ImportError: cannot import name cpu_count |
| 265 |
>>> joblib.__version__ |
| 266 |
'0.4.3' |
| 267 |
|
| 268 |
2012-01-02: FIX: py3k compatibility for reload reload() is moved to |
| 269 |
imp.reload(). |
| 270 |
|
| 271 |
It also seems that Python 3 behaves slightly differently wrt. |
| 272 |
reloads. For some reason, mdp.configuration is not imported |
| 273 |
properly on reload. Let's just not remove the name from the |
| 274 |
namespace, as this is the easiest fix. |
| 275 |
|
| 276 |
2011-12-23: ERF: added a failing test for reload MDP does not work with |
| 277 |
reload. See https://github.com/mdp-toolkit/mdp-toolkit/issues/1 |
| 278 |
for details. |
| 279 |
|
| 280 |
Thanks to Yaroslav Halchenko for reporting it. |
| 281 |
|
| 282 |
2011-12-18: FIX: Removed the leftover _global_message_emitter attribute (as |
| 283 |
reported by Tiziano and Hannah). |
| 284 |
2011-12-09: FIX: fix wrong sorting of eigenvectors in degenerate case for SVD. |
| 285 |
Thanks to Yaroslav Halchenko for reporting and fixing! |
| 286 |
|
| 287 |
2011-12-09: NEW: added failing test for wrong sorting of SVDs |
| 288 |
2011-10-24: FIX: updated sklearn wrappers to silence warnings for version >= |
| 289 |
0.9 |
| 290 |
|
| 291 |
------------------------------------------------------------------------------- |
| 292 |
MDP-3.2: |
| 293 |
2011-10-24: FIX: do not complain when failing to remove temporary directory. |
| 294 |
The error warning breaks doctests. On top of that, I do not think |
| 295 |
the error message is really useful. |
| 296 |
|
| 297 |
2011-10-22: FIX: fix previous commit (options can not be None) |
| 298 |
2011-10-22: FIX: fix bug with pytest when MDP is installed on user |
| 299 |
non-writable directory. |
| 300 |
2011-10-22: ENH: make NeuralGasNode automatic tests faster |
| 301 |
2011-10-22: ENH: call scklearn with its name whenever possible without too |
| 302 |
much disruption |
| 303 |
2011-10-22: FIX: workaround for debian patch to pp that disables pp's default |
| 304 |
password. |
| 305 |
2011-10-22: FIX: could not find submodules and classes in old versions of |
| 306 |
sklearn the problem was that we build names of sub-modules from |
| 307 |
strings, and we need to take into account two possible prefixes |
| 308 |
"sklearn" or "scikits.learn" |
| 309 |
|
| 310 |
2011-10-22: FIX: accomodate new name for scikits.learn 0.9 aka sklearn |
| 311 |
2011-10-22: RF: extract two methods and simplify logic of metaclass wrapping |
| 312 |
2011-10-22: OTH: one more metaclass test also cleand up whitespaces |
| 313 |
|
| 314 |
2011-10-21: FIX/python2.5: forgotten __future__ imports |
| 315 |
2011-10-21: ERF: rewrite one format string to make is clearer |
| 316 |
2011-10-21: FIX: extend infodict with **kwargs name, to avoid guessing later |
| 317 |
2011-10-19: ENH: metaclass: use dict to replace two lists and avoid infodict |
| 318 |
No functional changes. |
| 319 |
|
| 320 |
2011-10-21: FIX: finally fig bug in metaclass signature overwriting |
| 321 |
2011-10-21: OTH: make test_metaclass more stringent |
| 322 |
2011-10-21: OTH: added self contained test for metaclass and extensions |
| 323 |
issues. |
| 324 |
2011-10-19: OTH: added more detailed failing test for metaclass |
| 325 |
2011-10-19: OTH: another bug in signature overwriting. commited a failing |
| 326 |
test. Thanks to Alberto Escalante! |
| 327 |
|
| 328 |
2011-10-16: ENH: make tests for NeuralGasNode shorter |
| 329 |
2011-10-15: FIX: chmod -x mdp/test/test_NeuralGasNode.py |
| 330 |
2011-10-15: DOC: tiny changes in comments |
| 331 |
2011-10-15: FIX: tests for neural gas node |
| 332 |
2011-10-15: FIX: when start positions for the nodes was set, the number of |
| 333 |
nodes default is discarded |
| 334 |
2011-10-14: DOC+RF: slight improvements to documentation, comments, and |
| 335 |
spacing |
| 336 |
2011-10-14: FIX: NeuralGas fixes by Michael Schmucker |
| 337 |
2011-08-03: ENH: remove tuple unpacking init: it took twice as much lines |
| 338 |
2011-08-03: FIX: NeuralGas fixed by Michael Schmuker |
| 339 |
2011-08-03: FIX: remove \ line breaks |
| 340 |
2011-07-25: FIX: remove \ line breaks |
| 341 |
2011-07-25: NEW, FIX: tests for NeuralGasNode and two fixes |
| 342 |
2011-07-25: FIX: _remove_old_edges makes use of its argument |
| 343 |
2011-07-25: DOC, FIX: Typos in docstrings and whitespaces |
| 344 |
2011-07-25: NEW: NeuralGasNode by Michael Schmuker |
| 345 |
2011-10-14: FIX: remove unused variable |
| 346 |
2011-10-14: FIX: py3k bytes/str issue in subprocess input |
| 347 |
2011-10-14: FIX: remove unused variable |
| 348 |
2011-10-14: ENH: use a context manager for sys.stdout stealing |
| 349 |
2011-10-14: FIX: add pytest ini file to manifest |
| 350 |
2011-10-14: NEW: really check if pp needs monkey-patching |
| 351 |
2011-10-14: NEW: be compatible with shogun 1.0. Drop support for older |
| 352 |
releases. |
| 353 |
2011-10-14: FIX: fix handling of MDP_DISABLE_SHOGUN variable |
| 354 |
2011-10-14: Fix KeyError with NSDEBUG=1 Commit 7345d4a changed old-style |
| 355 |
formatting to new-style formatting, or at least the format string, |
| 356 |
but .format was not replaced with %! |
| 357 |
|
| 358 |
2011-10-13: DOC: make it a bit easier to see that you can use pp on debian too |
| 359 |
2011-10-13: FIX: remove (now) useless duplicated test in bimdp |
| 360 |
2011-10-13: FIX: windows does not allow me to remove a file |
| 361 |
2011-10-13: FIX: no with statement in tests {flow,node}.save Apparently |
| 362 |
windows does not allow one process to open a file multiple times. |
| 363 |
As {flow,node}.save use internally the with statement to open the |
| 364 |
dump file, we can not use the with statement in the test of these |
| 365 |
methods: nesting with statement means opening the same file |
| 366 |
multiple times. |
| 367 |
|
| 368 |
2011-10-12: NEW: allow for pretty printing of mdp.config. now config.info() is |
| 369 |
printed on stdout when you type |
| 370 |
|
| 371 |
>>> print mdp.config |
| 372 |
|
| 373 |
or |
| 374 |
|
| 375 |
>>> mdp.config |
| 376 |
|
| 377 |
don't think about the fact that this is only achievable by a |
| 378 |
metaclass ;-) |
| 379 |
|
| 380 |
2011-10-12: ERF: also add mdp.__version__ in config.info |
| 381 |
__revision__ is not enough, when mdp is installed and does not |
| 382 |
load from a git repository that string is empty (by the way: we |
| 383 |
should fix it!) |
| 384 |
|
| 385 |
2011-10-11: FIX: remove forgotten print statement |
| 386 |
2011-10-11: NEW: add tempfile test to help diagnose win7 problems |
| 387 |
2011-10-11: FIX: make py.test.__version__ check more robust |
| 388 |
2011-10-11: Add mdp.__revision__ to mdp.config.info() output This changes the |
| 389 |
order of initialization a bit: first mdp.configuration is fully |
| 390 |
imported, then mdp.utils. Because of this change, |
| 391 |
mdp.utils.repo_revision was renamed to mdp.repo_revision. It is |
| 392 |
imported only in mdp.configuration and then deleted from the |
| 393 |
visible namespace. get_git_revision was available as |
| 394 |
utils.get_git_revision, and now it's gone, but mdp.__revision__ |
| 395 |
contains the same value, so there little need to export the |
| 396 |
function. I hope nobody used it. |
| 397 |
|
| 398 |
>>> print mdp.config.info() |
| 399 |
python: 2.7.2.final.0 |
| 400 |
mdp: MDP-3.1-74-gbaca2a8+ |
| 401 |
parallel python: 1.6.0 |
| 402 |
shogun: NOT AVAILABLE: No module named shogun |
| 403 |
libsvm: libsvm.so.3 |
| 404 |
joblib: 0.4.6 |
| 405 |
scikits: 0.8.1 |
| 406 |
numx: scipy 0.9.0 |
| 407 |
symeig: scipy.linalg.eigh |
| 408 |
|
| 409 |
2011-10-11: NEW: append + to mdp.__revision__ in modified repo |
| 410 |
2011-10-11: FIX: half of an assert message was lost |
| 411 |
2011-10-11: ERF: use py.test 2 ini-file configuration. we can now check for |
| 412 |
py.test version, and get finally rid of the redundant conftest.py |
| 413 |
files. we now have only one conftest.py file for bimdp and mdp. |
| 414 |
|
| 415 |
2011-10-11: FIX: check if we have ENOENT in TemporaryDir.__del__ |
| 416 |
2011-10-11: FIX: stop complaining if dir is gone already in TemporaryDirectory |
| 417 |
2011-10-11: ERF: get rid of the pp monkey patch dirs the hard way. |
| 418 |
TemporaryDirectory does not manage to delete itself when created |
| 419 |
in a subprocess |
| 420 |
|
| 421 |
2011-10-11: ERF: make the pp monkeypatch dir prefix more explicit |
| 422 |
2011-10-11: ERF: change the interpretation of MDP_MONKEYPATCH_PP now you can |
| 423 |
specify a container directory for the patching. it defaults to |
| 424 |
tempfile.gettempdir(), i.e. /tmp on unix and something else on |
| 425 |
windows |
| 426 |
2011-10-11: FIX: only allow py.test version >= 2.1.2 py.test version > 2 are |
| 427 |
subtly incompatible with versions < 2. just settle on the new |
| 428 |
ones. problem: debian still does not package a reasonably recent |
| 429 |
version of py.test. in this case, just use the standalone script. |
| 430 |
thanks to 8a5b2e272de4230f15d09f85b7d35b0aeee3078e it's as |
| 431 |
flexible as py.test itself. |
| 432 |
2011-10-11: ERF: pass options to py.test in mdp.test. this way using the |
| 433 |
standalone script run_tests.py is as flexible as using py.test |
| 434 |
directly |
| 435 |
|
| 436 |
2011-10-11: ERF: updated standalone py.test script to version 2.1.2 |
| 437 |
2011-10-11: DOC: added pointer to win64 binaries distributor |
| 438 |
2011-10-10: ENH: use with statement for better readability in a couple of |
| 439 |
tests No functional changes. |
| 440 |
|
| 441 |
2011-10-10: ENH: make pyflakes happy by not defining unused names No |
| 442 |
functional changes. |
| 443 |
|
| 444 |
2011-10-10: FIX: make monkeypatch_pp compatible with both pp 1.6 and 1.6.1 |
| 445 |
>>> import pp |
| 446 |
>>> pp.version |
| 447 |
'1.6.0' |
| 448 |
>>> pp._Worker.command |
| 449 |
'"/usr/bin/python" -u "/usr/lib/pymodules/python2.6/ppworker.py" > |
| 450 |
2>/dev/null' |
| 451 |
|
| 452 |
>>> import pp |
| 453 |
>>> pp.version |
| 454 |
'1.6.1' |
| 455 |
>>> pp._Worker.command |
| 456 |
['/usr/bin/python2.6', '-u', |
| 457 |
'/usr/lib/pymodules/python2.6/ppworker.py', '2>/dev/null'] |
| 458 |
|
| 459 |
2011-10-10: FIX: temporary directory for tests was not managed properly sed |
| 460 |
's/mdp_toolkit_reporting_configured/mdp_configured/g' because the |
| 461 |
existence of this variable guards not only reporting |
| 462 |
configuration, but also the temporary directory stuff (and also |
| 463 |
the name was kind of long). |
| 464 |
|
| 465 |
sed 's/tempdirname/mdp_tempdirname/' because it goes in the |
| 466 |
py.test namespace and should have a prefix like the other |
| 467 |
variable. |
| 468 |
|
| 469 |
sed -r 's+bimdp/+TMP\/+g; s+mdp/+bimdp/+g; s+TMP\/+mdp/+g' |
| 470 |
<mdp/test/conftest.py >bimdp/test/conftest.py because bimdp should |
| 471 |
be kept in sync. In my installation bimdp is actually configured |
| 472 |
first, so tests in the temporary directory were failing with |
| 473 |
AttributeError on py.test.tempdirname. |
| 474 |
|
| 475 |
2011-10-10: FIX: tempdirs in cacheing tests were shortlived A temporary |
| 476 |
directory created with TemporaryDirectory can go at any time, so a |
| 477 |
construct like |
| 478 |
dirname = TemporaryDirectory().name is broken. |
| 479 |
|
| 480 |
Since the directories are created in the master temporary |
| 481 |
directory anyway, they don't have to be deleted separately. |
| 482 |
|
| 483 |
2011-10-10: FIX: do not set the cachedir by default on import mdp.caching |
| 484 |
setting it by default was creating a lot of bogus temporary |
| 485 |
directories when parallel processes where started using |
| 486 |
process_scheduler. I see no side effects for not setting it |
| 487 |
|
| 488 |
2011-10-10: FIX: better name for global temp dir |
| 489 |
2011-10-10: FIX: fix pp monkey patching. pp._Worker.command.replace(ppworker, |
| 490 |
ppworker3) did not work, as pp._Worker.command is a list and not a |
| 491 |
string! |
| 492 |
|
| 493 |
2011-10-10: FIX: put joblib caching stuff in the new global test tempdir |
| 494 |
2011-10-10: ERF: added a container directory for temporary test files. the |
| 495 |
directory gets deleted (no matter its contents) at the end of the |
| 496 |
test run. This should fix stale temp files. |
| 497 |
|
| 498 |
2011-10-05: FIX: partially revert 569cbe7 Changing 'except Exception:' clauses |
| 499 |
to bare 'except:' would catch and swallow ^C and the like. |
| 500 |
|
| 501 |
2011-10-02: ENH: revert the condition for pp monkeypatching It's better to |
| 502 |
only enable it when explicitely requested. The warning is also |
| 503 |
removed, because it doesn't make sense to warn about something |
| 504 |
that was explicitely requested. If the option to monkey patch is |
| 505 |
not given, check if pp is likely to be broken, and if so, disable |
| 506 |
it at configuration time. |
| 507 |
|
| 508 |
2011-10-02: FIX: remove some unused imports and one allocation |
| 509 |
2011-10-02: ENH: make temporary directory names more readable tmpxayPZEpp4mdp/ |
| 510 |
=> pp4mdp.tmpxayPZE/ |
| 511 |
|
| 512 |
2011-09-08: FIX: new module structure in scikits.learn 0.8 |
| 513 |
2011-09-07: BF: Exclude float16 dtypes in numpy.linalg.eigh. |
| 514 |
2011-08-24: Fixes to FastICA node Fine tuning and stabilisation are now used |
| 515 |
correctly, and the result closely resembles the original matlab |
| 516 |
version for the deflation/gaussian case with prewhitened data. The |
| 517 |
results are not numerically identical because there are still some |
| 518 |
differences in stopping criteria, but it's close. To verify, see |
| 519 |
original Matlab implementation. |
| 520 |
|
| 521 |
2011-08-10: DOC: added documentation item to TODO list |
| 522 |
2011-06-30: ERF: PEP8 fix (da pignolo) |
| 523 |
2011-06-30: ERF: simplify metaclass wrapping code |
| 524 |
2011-06-30: FIX: Fixed the node metaclass to correctly handle signatures and |
| 525 |
docstrings for inheritance chains. |
| 526 |
2011-06-30: ERF: Some small pylint-motivated updates. |
| 527 |
2011-06-30: ERF: wrap private methods in meteclass even if no docstring is |
| 528 |
defined. |
| 529 |
2011-06-30: ERF: Made the _get_required_train_args method more robust. |
| 530 |
2011-06-30: OTH: refactored metaclass tests, again. |
| 531 |
2011-06-30: OTH: refactored metaclass tests |
| 532 |
2011-06-16: DOC: added proper git format for changelog |
| 533 |
2011-06-16: OTH: commit failing test for docstring overwriting Thanks to |
| 534 |
Alberto Escalante to point this bug out. |
| 535 |
|
| 536 |
2011-04-07: FIX: verbose argument in set_cachedir was ignored. |
| 537 |
2011-06-12: FIX: Fixed a regression that broke the DBN example. |
| 538 |
2011-06-11: ERF: Major refactoring of the binet inspection code. This follows |
| 539 |
the refactoring of the flow-to-HTML conversion. Smaller |
| 540 |
refactorings in realted parts (slideshow) were also done. |
| 541 |
2011-06-11: ERF: Prepare large refactoring by renaming the files first. |
| 542 |
2011-05-31: ERF: Refactored the flow-to-HTML conversion. Threw out useless |
| 543 |
code, fixed style issues and now use the proper vocabulary |
| 544 |
(visitor pattern). The bimdp inspection has not yet been updated |
| 545 |
and is therefore broken in this commit. |
| 546 |
2011-04-18: FIX: monkeypatch_pp was broken I made some changes without |
| 547 |
actually testing if stuff works... |
| 548 |
|
| 549 |
2011-04-15: FIX: delete main joblib cache directory at mdp exit The directory |
| 550 |
is of course only deleted if we created it ourselves, i.e. if it |
| 551 |
is a temporary directory. |
| 552 |
|
| 553 |
2011-04-15: FIX: delete joblib cache directory after tests |
| 554 |
2011-04-15: DOC: describe MDPNUMX and MDPNSDEBUG env. vars. |
| 555 |
2011-04-15: FIX: bump supported scikits-learn version number Setting >=0.6 is |
| 556 |
pretty conservative, but Debian's 0.5-1 is bad for sure. |
| 557 |
|
| 558 |
2011-04-14: DOC: document MDP_DISABLE_MONKEYPATCH_PP |
| 559 |
2011-04-14: DOC: document MDP_DISABLE_* variables |
| 560 |
2011-01-12: DOC: improve rst in mdp/extension.py |
| 561 |
2011-03-21: FIX: drop some unused variables & imports Wanted to test pyflakes |
| 562 |
on something. This is the result. |
| 563 |
|
| 564 |
2011-04-14: FIX: apply a temporary workaround to the pp import problem For |
| 565 |
details see |
| 566 |
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=620551. |
| 567 |
|
| 568 |
ppworker.py must be run from some safe place. A temporary dir |
| 569 |
seems a good option. Unfortunately a temporary file is not enough, |
| 570 |
because parent directory should be empty. TemporaryDirectory is so |
| 571 |
complicated, because the code is run during Python shutdown |
| 572 |
procedure and some modules are not available. |
| 573 |
|
| 574 |
2011-04-05: FIX: make short description fit within 200 char limit Tiziano |
| 575 |
proposed the text, nobody protested... I guess that this signifies |
| 576 |
a consensus. :) |
| 577 |
|
| 578 |
Some small formatting fixes in source code and output were |
| 579 |
applied. |
| 580 |
|
| 581 |
2011-04-01: OTH: remove unnecessary parentheses around assert |
| 582 |
2011-04-04: FIX: restore python2.4, 2.5 support in setup.py Module ast is not |
| 583 |
available in python < 2.6. |
| 584 |
|
| 585 |
The format of docstrings is slightly changed: removed empty lines |
| 586 |
at the top. |
| 587 |
|
| 588 |
Tested with python 2.4, 2.5, 2.6, 2.7, 3.1, 3.2. Output looks |
| 589 |
identical. |
| 590 |
|
| 591 |
2011-04-01: FIX: [pp] Result container was shared between invocations |
| 592 |
2011-04-01: ERF: improved scikits.learn detection code. |
| 593 |
2011-03-31: ERF: Some small code improvements, no functional changes. |
| 594 |
2011-03-31: FIX: Fixed source path handling. |
| 595 |
|
| 596 |
------------------------------------------------------------------------------- |
| 597 |
MDP-3.1: |
| 598 |
|
| 599 |
2011-03-29: FIX: fixed shogun version detection code. As I expected, explicit |
| 600 |
version checking is bound to fail. I would prefer to just remove |
| 601 |
it and rely on functional checks, i.e. is the feature we depend on |
| 602 |
present or not? |
| 603 |
2011-03-24: ERF: CovarianceMatrix has the option of not centering the data; |
| 604 |
SFANode modified accordingly |
| 605 |
2011-03-22: FIX: fix wrong normalization in SFANode SFANode was using the |
| 606 |
covariance matrix of the derivatives, while the SFA algorithm |
| 607 |
requires the non-centered second moment matrix of the derivatives. |
| 608 |
This bug is probably irrelevant in most cases, but in some corner |
| 609 |
cases the error scales with 1/sqrt(N) where N is the number of |
| 610 |
samples. Many thanks to Sven Dähne for fishing this out ;-) |
| 611 |
2011-03-22: ERF: added new tests for SFANode Those tests are currently |
| 612 |
failing. |
| 613 |
2011-03-22: ERF: really use the debug flag in SFANode.stop_training |
| 614 |
2011-03-02: FIX: old svm versions would break the import process Bug reported |
| 615 |
by Philipp Meier. |
| 616 |
2011-03-03: DOC: updated TODO |
| 617 |
2011-03-03: DOC: updated TODO |
| 618 |
2011-02-23: DOC: updated CHECKLIST and TODO |
| 619 |
2011-02-23: ERF: performance and memory improvements for KNNClassifier. |
| 620 |
Transform array of n_samples^2 * n_features array into |
| 621 |
n_samples^2. Thanks to Fabian Pedregosa! |
| 622 |
2011-02-17: FIX: Small cosmetic fix in flow inspection for the case when |
| 623 |
stop_training doesn't receive any argument. |
| 624 |
2011-02-17: FIX: Do not wrap abstract scikits.learn transformers |
| 625 |
2011-02-17: Robustify wrap_scikits_transformer. Fix for the case that a |
| 626 |
transform does not implement some of the methods. |
| 627 |
2011-01-31: ERF: running self-contained test script is now blocking the prompt |
| 628 |
2011-01-27: FIX: Made the target checks more permissive when execution ends or |
| 629 |
is aborted. |
| 630 |
2011-01-27: ERF: Added additional checks for copy flag in BiLayer. |
| 631 |
2011-01-27: FIX: Fixed regression in inspection method output. |
| 632 |
2011-01-27: FIX: BiLayer was not compatible with all return value options. |
| 633 |
2011-01-27: FIX: Fixed some leftover inconsistencies from BiMDP stop_training |
| 634 |
interface changes. |
| 635 |
2011-01-26: DOC: with_extension doesn't work for coroutines, tell epydoc about |
| 636 |
extension |
| 637 |
2011-01-25: ERF: remove unittest-style wrapper class from test_extension.py. |
| 638 |
2011-01-25: OTH: use context managers to open/close file for pickling |
| 639 |
2011-01-25: ERF: Check for empty slides in execution as well. |
| 640 |
2011-01-25: ERF: Improved the behavior of show_training if an already trained |
| 641 |
flow is given (a special expection is now raised). |
| 642 |
2011-01-17: FIX: add trove classifiers for Python 2 and 3. |
| 643 |
2011-01-14: ERF: Updated the Parallel Python support, especially the |
| 644 |
NetworkPPScheduler. It now works with pp 1.6, and there are some |
| 645 |
refactoring and robustnes improvements as well. |
| 646 |
2011-01-13: FIX: really update CHECKLIST |
| 647 |
2011-01-13: NEW: mdp descriptions now defined in the module. They new |
| 648 |
descriptions are used in the setup.py script and in generating the |
| 649 |
web site. |
| 650 |
2011-01-13: ERF: get version and description in a robust way. |
| 651 |
|
| 652 |
------------------------------------------------------------------------------- |
| 653 |
MDP-3.0: |
| 654 |
|
| 655 |
2011-01-12: FIX: import NormalizeNode in mdp.nodes |
| 656 |
2011-01-12: FIX: no return in IdentityNode method |
| 657 |
2011-01-12: FIX: test_scikits should be run only if scipy is available |
| 658 |
2011-01-12: FIX: tell epydoc that nodes are documented in rst |
| 659 |
2011-01-11: NEW: convert scikits docstrings to rst This way I'm only getting |
| 660 |
missing target warnings and the text is parse as rst. No monospace |
| 661 |
font :) |
| 662 |
|
| 663 |
2011-01-11: FIX: fix some epydoc warnings, add some links and a forgotten |
| 664 |
export |
| 665 |
2011-01-07: MISC: duplicate some useful code among the three wrappers last |
| 666 |
merge from scikits branch. |
| 667 |
|
| 668 |
2011-01-11: ERF: make scikits default docs rst-compatible |
| 669 |
2011-01-10: Merge branch 'namespace_fixup' Conflicts: |
| 670 |
mdp/__init__.py |
| 671 |
mdp/nodes/__init__.py |
| 672 |
mdp/utils/__init__.py |
| 673 |
|
| 674 |
Add missing comma in __all__ in mdp/__init__.py. |
| 675 |
|
| 676 |
2011-01-10: OTH: shorten test name |
| 677 |
(Looked awkward in py.test output.) |
| 678 |
|
| 679 |
2011-01-10: NEW: introduce MDPNSDEBUG environment variable This variable |
| 680 |
controls namespace fixup debugging messages. |
| 681 |
|
| 682 |
Also remove 'advanced string formatting' in favour of printf-style |
| 683 |
statements (for python 2.5 compatibility). |
| 684 |
|
| 685 |
2011-01-10: FIX: made the one scikits test a bit more useful |
| 686 |
2011-01-10: FIX: there is no problem with output_dim in scikits classifiers |
| 687 |
because they are identity nodes |
| 688 |
|
| 689 |
2011-01-10: DOC: documentation of output_dim problem in scikits.learn wrappers |
| 690 |
2011-01-10: FIX: check for existence of im_func in classifier wrapper |
| 691 |
2011-01-10: FIX: fail meaningfully when trying to set output_dim |
| 692 |
2011-01-10: FIX: added docstrings to scikits wrappers |
| 693 |
2011-01-10: FIX: Cache mechanism called pre_execution_check with too much |
| 694 |
information |
| 695 |
2011-01-10: FIX: broken variable for disabling scikits.learn |
| 696 |
2011-01-10: FIX: fix wrapper string in scikits nodes |
| 697 |
2011-01-10: FIX: fixed hard-coded scikits test. |
| 698 |
2011-01-10: FIX: fixed a bit pp_remote support. I commented out the test. |
| 699 |
There's no point in having a test that requires a working ssh |
| 700 |
setup. we should move that code in the examples. TODO: I still did |
| 701 |
not manage to test the remote support properly: it does not work |
| 702 |
if ssh spits some session information on startup :-( |
| 703 |
|
| 704 |
2011-01-10: FIX: fix namespace issues with scikits nodes. |
| 705 |
2011-01-10: FIX: Made the straight forward fixes for pp 1.6 support, not sure |
| 706 |
if this is working now. |
| 707 |
2011-01-10: Merge remote branch 'origin/master' |
| 708 |
2011-01-10: Merge a branch with some scikits work I'm doing this as a merge |
| 709 |
instead of rebasing, because the work was really done in parallel, |
| 710 |
and e.g. Tiziano removed the file, which I moved and updated. |
| 711 |
|
| 712 |
2011-01-10: FIX: make scikits support depend on version 0.5 Version 0.4 |
| 713 |
breaks... Version 0.5 at least does learn & execute. |
| 714 |
|
| 715 |
2011-01-10: FIX: move scikits test into mdp/tests/ |
| 716 |
2011-01-10: Merge remote branch 'origin/master' Conflicts: |
| 717 |
mdp/test/test_pp_local.py |
| 718 |
|
| 719 |
2011-01-10: FIX: Updated local scheduler to work with latest pp version 1.6. |
| 720 |
2011-01-10: FIX: fix name leaking in nodes/__init__.py |
| 721 |
2011-01-10: FIX: add scikits to list of dependencies in testall.py |
| 722 |
2011-01-10: FIX: Fixed local Parallel Python tests. |
| 723 |
2011-01-10: FIX: make test_ISFANode more robust. |
| 724 |
2011-01-10: FIX: remove stale test_scikits.py file |
| 725 |
2011-01-10: NEW: merged constant_expansion_node branch. Adds new |
| 726 |
GeneralExpansionNode thanks to Alberto Escalante. |
| 727 |
|
| 728 |
Tests and node had to be adjusted to merge cleanly and to follow |
| 729 |
new conventions. TODO: pseudo_inverse method is quite fragile, |
| 730 |
why? |
| 731 |
|
| 732 |
2011-01-10: ERF: Added native traceback flag to IDE test run template. |
| 733 |
2011-01-09: FIX: remove bitrot in test_pp_*.py. This makes the local tests |
| 734 |
pass with parallel python 1.5.4. Remote tests still fail because |
| 735 |
amongst other things the server list must be passed explicitely. |
| 736 |
|
| 737 |
The tests fail horribly with newer pp version. |
| 738 |
|
| 739 |
2011-01-09: FIX: kill epydoc warnings |
| 740 |
2011-01-09: Revert "DOC: migrated class.__doc__ to rst for nerual_gas_nodes" |
| 741 |
This reverts commit 2c2a4cd72ecd7d2dbb2032dd7db4f0754b3ba846. |
| 742 |
|
| 743 |
2011-01-07: FIX: make joblib version checking more robust e.g. |
| 744 |
joblib.__version__='1.0.0-gitABCEDFG' should pass. |
| 745 |
|
| 746 |
2011-01-07: FIX: be more tolerant against missing modules in scikits.learn. |
| 747 |
depending on the installation procedure, user may or may not have |
| 748 |
some of the scikits.learn modules. we just wrap the modules we |
| 749 |
find. |
| 750 |
|
| 751 |
2011-01-07: FIX: add wrongly removed 'scikits' namespace from mdp.nodes. |
| 752 |
2011-01-07: FIX: wrong version in scikits.learn info item. |
| 753 |
2011-01-07: ERF: Merged "scikits" branch into "master". Conflicts: |
| 754 |
mdp/__init__.py. Fixed new config syntax and generic_tests. |
| 755 |
|
| 756 |
2011-01-06: FIX: fixed testall script. now it works. the script should be |
| 757 |
ported to use subprocess, so that it runs OS agnostic. |
| 758 |
|
| 759 |
2011-01-05: ERF: improved testall.py |
| 760 |
2011-01-05: DOC: updated CHECKLIST for release. |
| 761 |
2011-01-05: FIX: Slightly changed the include_last_sample argument to fix the |
| 762 |
problem with the standard parallel fork implementation. |
| 763 |
2011-01-05: FIX: fix test broken due to new keyword in SFANode. |
| 764 |
2011-01-05: FIX: added include_last_sample to SFA2Node too. |
| 765 |
2011-01-05: ERF: added include_last_sample switch to SFANode. Motivation and |
| 766 |
discussion can be found in this thread: |
| 767 |
http://sourceforge.net/mailarchive/forum.php?thread_name=20100826183424.GF26995%40tulpenbaum.cognition.tu-berlin.de&forum_name=mdp-toolkit-users |
| 768 |
|
| 769 |
2011-01-05: FIX: clean up __all__ in mdp/__init__.py numx* should not be in |
| 770 |
all. Quoting Tiziano: |
| 771 |
|
| 772 |
> numx was not in __all__, and therefore was not documented by |
| 773 |
epydoc. |
| 774 |
> Once it is added to __all__, indeed things break heavily. well, |
| 775 |
I think even it was by mistake, it is a good thing that numx is |
| 776 |
not in __all__. mdp.numx is an *internal* thing. people doing from |
| 777 |
mdp import * should not get numx. so, leave it like this and we |
| 778 |
don't even need a workaround for epydoc. |
| 779 |
|
| 780 |
While at it, remove whitening which was already gone and make the |
| 781 |
list alphabetical. |
| 782 |
|
| 783 |
2011-01-04: Merge branch 'cache' Conflicts: |
| 784 |
mdp/__init__.py |
| 785 |
|
| 786 |
2011-01-04: FIX: make setup.py run on also on python3 The code must run on all |
| 787 |
supported python versions. |
| 788 |
|
| 789 |
try: ... except Type, v: ... |
| 790 |
|
| 791 |
file() |
| 792 |
|
| 793 |
cannot be used. |
| 794 |
|
| 795 |
While at it, I'm changing exec() to re.match. It's simpler. |
| 796 |
|
| 797 |
2011-01-04: FIX: python2.5 has no itertools.product |
| 798 |
2011-01-04: ERf: added script to test all python versions and dependencies. |
| 799 |
2011-01-04: FIX: fix name error in configuration.py |
| 800 |
2011-01-04: ERF: added way of disabling optional dependencies optional |
| 801 |
dependencies can be disabled setting the env variable |
| 802 |
MDP_DISABLE_DEPNAME |
| 803 |
|
| 804 |
2011-01-04: ERF: remove long deprecated helper functions. pca and fastica |
| 805 |
survive because we use them as ad in the the site and because most |
| 806 |
probably they are the only one in real use out there. |
| 807 |
|
| 808 |
2011-01-04: FIX: fix test failures due to previous commit. |
| 809 |
2011-01-04: ERF: cleaned up mdp/__init__.py. configuration is now a separate |
| 810 |
file. the logic of numx loading has been changed (again!). |
| 811 |
numpt/scipy can not be treated as an external dependency. |
| 812 |
|
| 813 |
2011-01-04: FIX: fix generic node test failures for |
| 814 |
TimeFrameSlidingWindowNode. |
| 815 |
2011-01-04: ERF: support iterables in generic node tests. |
| 816 |
2011-01-03: FIX: MDP depends on joblib >= 0.4.3 |
| 817 |
2011-01-03: FIX; fixed git log format in CHECKLIST |
| 818 |
2011-01-03: DOC: add __homepage__ variable in __init__.py |
| 819 |
2011-01-03: FIX: removed useless file. |
| 820 |
2011-01-03: ERF: updated author and maintainers, put version in one single |
| 821 |
place. |
| 822 |
2011-01-03: DOC: changed reference to installation instructions |
| 823 |
2011-01-03: ERF: removed our names from COPYRIGHT. It makes little sense to |
| 824 |
keep up-to-date the list of maintainers in this file. we have one |
| 825 |
authoritative source: the development.rst page. |
| 826 |
|
| 827 |
2010-12-31: FIX: fixed TimeDelayNode docstring rst formatting |
| 828 |
2010-12-31: NEW: new TimeDelay and TimeDelaySlidingWindow nodes. Thanks to |
| 829 |
Sebastian Höfer! |
| 830 |
|
| 831 |
2010-12-31: DOC: specify dependency in svm nodes. |
| 832 |
2010-12-31: FIX: change order of nodes in nodes/__init__.py __all__ order is |
| 833 |
somehow arbitrary, but better than before. it is needed for the |
| 834 |
automatic node list generation in the web site. |
| 835 |
|
| 836 |
2010-12-31: DOC: migrated class.__doc__ to rst for svm family |
| 837 |
2010-12-31: FIX: logic and wording of the symeig dependency test were wrong. I |
| 838 |
don't have a 64bit machine to test that the function in |
| 839 |
mdp/utils/routines.py still does the right thing. if it does not, |
| 840 |
just change mdp.config.has_symeig == '...' to |
| 841 |
mdp.config.has_symeig != '...' |
| 842 |
;-) |
| 843 |
|
| 844 |
2010-12-31: DOC: migrated class.__doc__ to rst for misc_nodes |
| 845 |
2010-12-31: DOC: migrated class.__doc__ to rst for classifier_nodes |
| 846 |
2010-12-31: DOC: migrated class.__doc__ to rst for expansion_nodes |
| 847 |
2010-12-31: DOC: migrated class.__doc__ to rst for nerual_gas_nodes |
| 848 |
2010-12-31: DOC: migrated class.__doc__ to rst for regression_nodes |
| 849 |
2010-12-31: DOC: migrated class.__doc__ to rst for rbm_nodes |
| 850 |
2010-12-31: DOC: migrated class.__doc__ to rst for lle_nodes |
| 851 |
2010-12-31: DOC: migrated class.__doc__ to rst for em_nodes |
| 852 |
2010-12-31: DOC: migrated class.__doc__ to rst for fda_nodes |
| 853 |
2010-12-31: DOC: migrated class.__doc__ to rst for sfa family |
| 854 |
2010-12-31: DOC: migrated class.__doc__ to rst for ica family |
| 855 |
2010-12-31: DOC: migrated class.__doc__ to rst for nipals |
| 856 |
2010-12-31: DOC: migrated class.__doc__ to rst for pca_nodes |
| 857 |
2010-12-31: DOC: migrate convolution nodes to rst docstring |
| 858 |
2010-12-31: DOC: fixed migration to rst docstrings for xsfa_nodes. |
| 859 |
2010-12-30: Merge branch 'feature/libsvm_291_support' Conflicts: |
| 860 |
mdp/__init__.py |
| 861 |
mdp/test/test_contrib.py |
| 862 |
|
| 863 |
2010-12-30: FIX: failing dtype_consistency_tests with numpy as backend The |
| 864 |
config.has_symeig logic was that config.has_symeig is False when |
| 865 |
using the _fake_symeig compatibility wrapper. Dtypes were selected |
| 866 |
based on that. |
| 867 |
|
| 868 |
2010-12-30: FIX: make sure mdp.config.has_{numpy,scipy} are always set, really |
| 869 |
2010-12-30: FIX: make sure mdp.config.has_{numpy,scipy} are always set |
| 870 |
2010-12-30: FIX: Don’t try to fetch version of numx_version twice. |
| 871 |
2010-12-29: ERF: Put SVM tests into a separate file. Also: Fix |
| 872 |
skip_on_condition on class. |
| 873 |
2010-12-27: DOC: rename ExternalDepFail to -Failed and document stuff |
| 874 |
2010-12-27: FIX: test_seed failured when testing running bimdp.test() after |
| 875 |
mdp.test() The reason was that when the test functions were |
| 876 |
imported from mdp.test into bimdp.test, they still referenced |
| 877 |
mdp.test._SEED global variable. But the --seed option in bimdp is |
| 878 |
independent, so it was set to a different random seed. By |
| 879 |
duplicating test_seed.py, a second independent global is created. |
| 880 |
|
| 881 |
2010-12-27: Add a test for separable data. |
| 882 |
2010-12-27: NEW: check if joblib is >= 0.4.6 |
| 883 |
2010-12-27: ERF: Restructured the libsvm tests. |
| 884 |
2010-12-27: OTH: make COPYRIGHT valid rst This makes is easier to include in |
| 885 |
the webpage. |
| 886 |
|
| 887 |
2010-12-27: Merge branch 'config_simplification' Conflicts: |
| 888 |
mdp/__init__.py |
| 889 |
|
| 890 |
This merge commit contains an adaptation of caching to the new |
| 891 |
framework. |
| 892 |
|
| 893 |
2010-12-27: NEW: convert caching_extention.py to rst |
| 894 |
2010-12-27: Merge branch 'rst_docstrings' Conflicts: |
| 895 |
mdp/signal_node.py |
| 896 |
|
| 897 |
2010-12-27: Merge branch 'cache' Conflicts: |
| 898 |
mdp/caching/caching_extension.py |
| 899 |
|
| 900 |
2010-12-27: Merge commit '06007d3e8987a6c8574fd63c7608b37e9e657e66' |
| 901 |
2010-12-27: ERF: Added test for execute_method in classifiers. |
| 902 |
2010-12-27: ERF: Add validation for libsvm parameters. |
| 903 |
2010-12-27: ERF: Get rid of superfluous parameter handling. Use svm_parameter |
| 904 |
directly. |
| 905 |
2010-12-27: FIX: Get the libsvm tests running for versions >=2.91. TODO: * The |
| 906 |
new interface still needs some polishing. |
| 907 |
* We need to test the RBF classifier. It won’t work right now. |
| 908 |
2010-12-24: FIX: fix test of PCANode for degenerate case, it was failing |
| 909 |
occasionally. |
| 910 |
2010-12-24: removed now useless copy method overwriting in NoiseNode |
| 911 |
2010-12-24: Merge branches 'deepcopy' and 'master' |
| 912 |
2010-12-24: ERF: added additional tests for node and flow deepcopy This tests |
| 913 |
fail in master branch righ now, because lambdas can not be |
| 914 |
pickled! |
| 915 |
|
| 916 |
2010-12-24: ERF: added additional tests for node and flow deepcopy |
| 917 |
2010-12-24: Merge branch 'master' of |
| 918 |
ssh+git://mdp-toolkit.git.sourceforge.net/gitroot/mdp-toolkit/mdp-toolkit |
| 919 |
2010-12-24: FIX: fixex bimdp.test() under windows. The FIX introduced with |
| 920 |
dfa8ec5a0d367941e4c297325a43361d579f8bbd was not working on |
| 921 |
windows, because git does not support symlinks there. It has been |
| 922 |
fixed differently now, and it should work reliably under any |
| 923 |
system, I hope. |
| 924 |
|
| 925 |
2010-12-22: ERF: Improved the exception handling for failing training argument |
| 926 |
check (user now gets a better error message). |
| 927 |
2010-12-22: changed my email address in copyright and setup |
| 928 |
2010-12-21: Merge remote branch 'new_guidelines' |
| 929 |
2010-12-21: FIX: now bimdp.test() really works as advertised! |
| 930 |
2010-12-21: FIX: regenerate run_tests.py script for running tests without |
| 931 |
py.test |
| 932 |
2010-12-03: OTH: remove uses of deprecated dict.has_key() |
| 933 |
2010-12-03: OTH: use with_statement in signal_node |
| 934 |
2010-12-18: DOC: improved docs of caching mechanism |
| 935 |
2010-12-18: FIX: possible to switch caching path while extension active |
| 936 |
2010-12-18: FIX: caching started at second execution because of automatic |
| 937 |
setting of dtype, input_dim |
| 938 |
2010-12-18: NEW: caching extension can now cache individual classes or |
| 939 |
instances |
| 940 |
2010-12-18: DOC: a few fixes to the cache extension docs |
| 941 |
2010-12-09: NEW: Added parallelization for NearestMean and KNN classifier. |
| 942 |
Also updated NearestMeanClassifier attribute names to avoid |
| 943 |
confusion. |
| 944 |
2010-12-09: DOC: Added explanation in _default_fork docstring. |
| 945 |
2010-12-09: DOC: Improved code documentation. |
| 946 |
2010-12-03: FIX: copy with deepcopy everywhere, raise DeprecationWarning on |
| 947 |
old API Since Node.copy() is a public function, changing the |
| 948 |
allowed parameters might be bad. So just ignore the parameter and |
| 949 |
raise a DeprecationWarning if the parameter has a non-None value. |
| 950 |
|
| 951 |
>>> import mdp |
| 952 |
>>> node = mdp.Node() |
| 953 |
>>> node2 = node.copy(protocol=1) |
| 954 |
__main__:1: MDPDeprecationWarning: protocol parameter to copy() is |
| 955 |
ignored |
| 956 |
|
| 957 |
2010-12-03: FIX: Patch for using deepcopy instead of pickle for Node and Flow |
| 958 |
Hi all, |
| 959 |
|
| 960 |
here is a patch and test script for replacing the pickle with |
| 961 |
deepcopy in Node.copy() and Flow.copy(). The signature of the copy |
| 962 |
function changes (the protocol keyword is no longer necessary), I |
| 963 |
don't know what the proper way to handle this is (deprecation |
| 964 |
error?), so I'll leave it up to you guys. |
| 965 |
|
| 966 |
Thanks for all the efforts, we get very positive feedback from all |
| 967 |
users of mdp and Oger! |
| 968 |
|
| 969 |
David |
| 970 |
|
| 971 |
2010-12-02: OTH: remove redundant shogun version checking |
| 972 |
2010-12-02: NEW: add tests for the new config architecture |
| 973 |
2010-12-01: OTH: Convert config.has* to an attribute I have suspicions that |
| 974 |
using a lambda function was causing problems with the process |
| 975 |
scheduler. |
| 976 |
|
| 977 |
git grep -l -E 'config.has_'|xargs sed -r -i |
| 978 |
's/(config.has_[a-z_]+)\(\)/\1/g' |
| 979 |
|
| 980 |
2010-12-01: OTH: simplify mdp config discovery Create |
| 981 |
ExternalDepFound/ExternalDepFail objects immediately after |
| 982 |
attempting import. This way there's no need to reverse engineer |
| 983 |
where the symeig implementation comes from. |
| 984 |
|
| 985 |
TODO: test_process_schedule fails! |
| 986 |
|
| 987 |
2010-12-01: OTH: simplify numpy/scipy imports |
| 988 |
2010-12-01: OTH: make ExternalDep objects have boolean value |
| 989 |
2010-12-02: OTH: replace 'map()' with 'for:' following 2to3's advice |
| 990 |
2010-12-02: OTH: remove redundant OrderedDict implementation I'm also moving |
| 991 |
the OrderedDict recipe to a separate file, because it's kind of |
| 992 |
logically separate and will not be used in newer versions of |
| 993 |
Python anyway. |
| 994 |
|
| 995 |
2010-12-01: OTH: mdp.test() should not cause SystemExit Function |
| 996 |
py.cmdline.pytest calls py.test.cmdline.main, but also raises |
| 997 |
SystemExit. Just call the latter one, so the interactive session |
| 998 |
can continue. |
| 999 |
|
| 1000 |
2010-12-01: Change __license__ to 'Modified BSD' in source code The license |
| 1001 |
text was already changed in the documentation, but not in the |
| 1002 |
source code. |
| 1003 |
|
| 1004 |
2010-11-30: NEW: Added the execute_method argument and attribute to the |
| 1005 |
Classifier class. This makes it possible to get the classification |
| 1006 |
results from a normal flow without resorting to monkey patching |
| 1007 |
(which makes nodes unpickelable). |
| 1008 |
2010-11-30: ERF: Improved code readibility. |
| 1009 |
2010-11-30: NEW: Added iadd method to Flow. Also fixed the error strings. |
| 1010 |
2010-11-29: NEW: Added parallel version of Gaussian classifier. |
| 1011 |
2010-11-29: ERF: Added helper function to combine CovarianceMatrix instances. |
| 1012 |
2010-11-24: NEW: Added KNN classifier. |
| 1013 |
2010-11-24: FIX: Fixed wrongly created classifier BiNodes. |
| 1014 |
2010-11-24: NEW: Added nearest-mean classifier node. |
| 1015 |
2010-11-24: FIX: Updated test name. |
| 1016 |
2010-11-24: ERF: Some cleanup in the GaussianClassifier. Added an exception if |
| 1017 |
any class covariance matrix is singular. |
| 1018 |
2010-11-23: BRK: Renamed GaussianClassifierNode to GaussianClassifier to |
| 1019 |
follow the general classifier convention (otherwise a special case |
| 1020 |
would be needed in the BiMDP autogenerated code). |
| 1021 |
2010-11-23: FIX: Updated BiMDP test for the new siganture of FDANode. |
| 1022 |
2010-11-23: BRK: Updated FDANode siganture to follow the newer conventions |
| 1023 |
from ClassifierNode (labels instead of cl) and the PCA/SFANode (n |
| 1024 |
instead of range). |
| 1025 |
2010-11-09: OTH: move PROJECT_GUIDELINES and NEW_DEVELOPER_INFO to |
| 1026 |
development.rst in docs |
| 1027 |
2010-11-01: NEW: convert neural_gas_nodes.py and rbm_nodes.py docstrings to |
| 1028 |
rst |
| 1029 |
2010-11-01: NEW: add reference to JoMLR article |
| 1030 |
2010-11-01: NEW: convert mdp/nodes/xsfa_nodes.py docstrings to rst |
| 1031 |
2010-11-01: NEW: convert mdp/__init__.py and mdp/nodes/signal_node.py |
| 1032 |
docstrings to rst |
| 1033 |
2010-11-01: Merge commit 'c3bdc14c092b92c214f8bd76002be565706f07ca' |
| 1034 |
2010-11-01: FIX: Fixed bug in node join, leading to inconsistent state under |
| 1035 |
certain conditions with FlowNode. |
| 1036 |
2010-10-31: ERF: Added copy_callable option for thread scheduler. |
| 1037 |
2010-10-16: FIX: Updated failing unittest. |
| 1038 |
2010-10-16: FIX: Fixed the parameter names in the switchboard factory. |
| 1039 |
2010-10-15: BRK: Updated the remaining switchboards. |
| 1040 |
2010-10-15: BRK: Started to replace the x_ and y_ arguments in 2d switchboards |
| 1041 |
with a single _xy argument. So far only the |
| 1042 |
Rectangular2dSwitchboard is done. |
| 1043 |
2010-09-26: FIX: Correctly distinguish Flow and BiFlow in trace inspection |
| 1044 |
execute. |
| 1045 |
2010-09-25: FIX: Fixed bug in ParallelFlowNode and added unittest. |
| 1046 |
2010-09-14: DOC: Fixed wrong docstring. |
| 1047 |
2010-09-14: BRK: Improved extension context manager and decorator to only |
| 1048 |
deactivate extensions that were activated by it. This prevents |
| 1049 |
unintended side effects. Corresponsing tests were added. |
| 1050 |
2010-09-14: FIX: Fixed outdated import. |
| 1051 |
2010-09-14: Merge commit '4671fc2511b3446e57afcc1974d46f15ac92cb5f' |
| 1052 |
2010-09-14: BRK: Improved the dimension consistency handling in FlowNode, also |
| 1053 |
making the dimension setting stricter. This is to prevent really |
| 1054 |
strange errors in some corner cases (e.g. when the output_dim is |
| 1055 |
automatically set in execute). One generic test had to be dropped, |
| 1056 |
but explicit test for the dimension setting were added. |
| 1057 |
2010-09-14: ERF: Streamlined the parallel fork template method. |
| 1058 |
2010-09-14: FIX: Fixed the test node in the test_execute_fork unittest. |
| 1059 |
2010-09-14: ERF: Removed the ParallelBiNode, since it is no longer needed. |
| 1060 |
2010-09-14: Revert "FIX: Fixed small bug in dimension consistency check." This |
| 1061 |
reverts commit 0b37100894f08faf3653bab09bc4b5fb21705d72. |
| 1062 |
|
| 1063 |
2010-09-13: ERF: Added more unittests for parallel. One still fails because an |
| 1064 |
improvement in FlowNode is needed. |
| 1065 |
2010-09-13: DOC: Added comment on necessary extension context improvement. |
| 1066 |
2010-09-13: FIX: Fixed small bug in dimension consistency check. |
| 1067 |
2010-09-13: ERF: Improved the threaded scheduler (better exception handling). |
| 1068 |
2010-09-13: FIX: Fixed parallel bug, missing argument in callable. |
| 1069 |
2010-09-13: ERF: Turned the use_execute_fork property into a normal method. |
| 1070 |
2010-09-12: FIX: Added the missing property decorators (this was dangerous, |
| 1071 |
since a function attribute is cast as True).. |
| 1072 |
2010-09-12: FIX: Fixed bug in ParallelNode. |
| 1073 |
2010-09-12: FIX: Updated ParallelFlow to use correct execution result |
| 1074 |
container. |
| 1075 |
2010-09-12: FIX: Updated __all__ list. |
| 1076 |
2010-09-10: OTH: python2.5: remove PEP 3101 formatting |
| 1077 |
2010-07-30: OTH: python2.5: 'with' and 'as' |
| 1078 |
2010-09-09: Remove next-to-last import numpy from mdp codebase numx rules! |
| 1079 |
|
| 1080 |
2010-09-09: ERF: Changed the node purging mechanism in parallel to no longer |
| 1081 |
use a special extension method in FlowNode. Also some smaller |
| 1082 |
comment updates. |
| 1083 |
2010-09-09: NEW: MDP migrates to BSD license! |
| 1084 |
2010-09-09: BRK: Clarfied some subtle aspects of the parallel API and updated |
| 1085 |
the code to improve consistency. |
| 1086 |
2010-09-09: Make git display python diffs better |
| 1087 |
2010-09-08: FIX: Some smaller fixes in the updated parallel code. |
| 1088 |
2010-09-08: FIX: The parallel BiMDP tests are running again. Also some more |
| 1089 |
cleanup in MDP parallel. |
| 1090 |
2010-09-08: FIX: The standard MDP tests now all run again after the changes in |
| 1091 |
parallel. |
| 1092 |
2010-09-07: BRK: Continued work on the parallelization update, not working |
| 1093 |
yet. |
| 1094 |
2010-09-06: BRK: Started to modify the parallel node forking. The |
| 1095 |
is_bi_training in BiNode is replaced with a use_execute_fork in |
| 1096 |
the ParallelExtensionNode base class. This is more logical, less |
| 1097 |
complicated and a first step towards enabling more parallelization |
| 1098 |
patterns. |
| 1099 |
2010-09-02: OTH: the caching mechanism could be simplified with the latest |
| 1100 |
joblib |
| 1101 |
2010-08-31: DOC: Improve docs of caching mechanism, keywork argument for |
| 1102 |
verbosity of cache |
| 1103 |
2010-08-26: FIX: SFANode.train fails with a meaningful error if trained with |
| 1104 |
x.shape = (1, n) |
| 1105 |
2010-08-26: Fix --seed option for py.test >= 1.3 Py.test in version 1.3 added |
| 1106 |
a feature to skip conftest.py files when they are exactly the |
| 1107 |
same. This helps when a directory is copied to a subdirectory of |
| 1108 |
itself and the config files are not idempotent. Unfortunately, |
| 1109 |
this runs afoul of our logic to add the --seed option everywhere, |
| 1110 |
irrespective of wheter py.test is run implicitely for the whole |
| 1111 |
mdp-toolkit or for mdp or for bimdp or for both. The solution is |
| 1112 |
to modify one of the conftest.py files to be unidentical. |
| 1113 |
|
| 1114 |
Tested with with py.test 1.2.1 and 1.3.3. |
| 1115 |
|
| 1116 |
2010-08-18: OTH: Simplify docstring formatting. |
| 1117 |
2010-08-25: NEW: FANode detects singluar matrices |
| 1118 |
2010-08-24: ERF: fftpack does not support float >64; those types are removed |
| 1119 |
from the supported list for ConvolutionNode |
| 1120 |
2010-08-24: FIX: Remove forgotten print statements |
| 1121 |
2010-08-24: ERF: ISFA supports float>64 |
| 1122 |
2010-08-23: ERF: added python version to mdp.config |
| 1123 |
2010-08-19: ERF: Add complex256 to list of unsafe dtypes |
| 1124 |
2010-08-19: FIX: Generic tests failed for MDPNUMX='numpy' When symeig and |
| 1125 |
scipy.linalg.eigh are not present, MDP falls back into using |
| 1126 |
numpy.linalg.eig . This does not support large floating point |
| 1127 |
types, for some reasons, which caused the tests to fail. |
| 1128 |
|
| 1129 |
2010-08-17: ERF: Updated the classifier nodes that were overlooked in the big |
| 1130 |
PreserveDimNode cleanup. |
| 1131 |
2010-08-17: Merge branch 'supporteddtypes' Conflicts: |
| 1132 |
mdp/nodes/misc_nodes.py |
| 1133 |
|
| 1134 |
2010-08-17: ERF: Clarify questions about support for float64 only |
| 1135 |
2010-08-17: FIX: Fixed generic tests, now check that PreserveDimNode raises |
| 1136 |
right exceptions. |
| 1137 |
2010-08-17: ERF: Added tests for PreserveDimNode. |
| 1138 |
2010-08-17: ERF: Wrapped up the PreserveDimNode changes. Unable to fix the |
| 1139 |
generic unittests, somebody has to fix those. |
| 1140 |
2010-08-16: ERF: Refined the choice of supported dtypes |
| 1141 |
2010-08-16: ERF: Redefine default supported dtypes to be 'Floats', and modify |
| 1142 |
all nodes accordingly |
| 1143 |
2010-08-16: ERF: Further simplified Cumulator mechanism; added test After |
| 1144 |
digging into numpy's C code, we decided that at present |
| 1145 |
numpy.concatenate does the right thing, i.e. it creates a buffer |
| 1146 |
large enough and uses memcpy to copy the data in each array. |
| 1147 |
|
| 1148 |
2010-08-16: FIX: py.test failed with an error reporting a duplicated 'seed' |
| 1149 |
option In py.test 1.3.3 they changed the sequence in which the |
| 1150 |
conftest.py |
| 1151 |
files are loaded. In our case, the /mdp/test/conftest.py file |
| 1152 |
was |
| 1153 |
laoded before the /conftest.py file, which tried to redefine |
| 1154 |
the |
| 1155 |
optione 'seed', causing an error. |
| 1156 |
|
| 1157 |
2010-08-14: ERF: Simplified concatenation mechanism in Cumulator The new |
| 1158 |
solution is much faster and still avoids direct calls to |
| 1159 |
numpy.concatenate, that performs pairwise concatenations and |
| 1160 |
wastes a lot of memory. |
| 1161 |
|
| 1162 |
2010-08-14: FIX: Cumulator had no docstring |
| 1163 |
2010-08-10: FIX: Added missing exit target check in BiFlow. Changed the |
| 1164 |
unittest to better test this case. |
| 1165 |
2010-08-10: ERF: Changed BiFlow.execute to always return a tuple (x,msg), even |
| 1166 |
if the msg is empty. The previous changing return signature was |
| 1167 |
too likely to cause errors. |
| 1168 |
2010-08-10: DOC: Small fix in docstring. |
| 1169 |
2010-08-10: DOC: Small fix in comment. |
| 1170 |
2010-08-09: ERF: Added check for case when both x and msg are None to produce |
| 1171 |
better error message. |
| 1172 |
2010-08-09: ERF: Removed duplicate verbose print. |
| 1173 |
2010-08-08: FIX: Fixed the missing import for PreserveDimNode and fixed the |
| 1174 |
broken generic tests for these nodes by excluding them. This is |
| 1175 |
only a provisional fix. |
| 1176 |
2010-08-08: ERF: Rewrote the creation of the autogenerated BiNode classes. |
| 1177 |
Went back to dynamically creating the classes at runtime. Now use |
| 1178 |
a node blacklist instead of the previous whitelist. These two |
| 1179 |
changes allow the import of nodes with dependencies, like the SVM |
| 1180 |
nodes. |
| 1181 |
2010-08-08: NEW: Added new PreserveDimNode base class for input_dim == |
| 1182 |
output_dim. This should avoid duplication of the set_dim methods. |
| 1183 |
2010-08-05: ERF: Tweaked the hinet CSS. |
| 1184 |
2010-08-05: DOC: Improved the ParallelFlow docstrings. |
| 1185 |
2010-07-30: DOC: Small improvements to the corutine decorator docs. |
| 1186 |
2010-07-30: FIX: Removed some left over stop_message references. |
| 1187 |
2010-07-29: NEW: extend namespace_fixup to bimdp |
| 1188 |
2010-07-29: ERF: rework fixup_namespace function and add tests |
| 1189 |
2010-07-27: FIX: NodeMetaclass and ExtensionException where imported but not |
| 1190 |
in __all__ |
| 1191 |
2010-07-27: FIX: remove helper_funcs module from mdp namespace, they are |
| 1192 |
imported into mdp |
| 1193 |
2010-07-27: NEW: Added a simple run script to run the tests from a .py file |
| 1194 |
(e.g. to use the Eclipse debugger). |
| 1195 |
2010-07-27: ERF: Optimized the switchboard gradient implementation (about 10x |
| 1196 |
faster). |
| 1197 |
2010-07-27: ERF: Changed one more unittest to use scheduler context manager. |
| 1198 |
2010-07-27: ERF: Added context manager interface to Scheduler class and added |
| 1199 |
unittests. |
| 1200 |
2010-07-27: ERF: Renamed callable_ to task_callable and made some small |
| 1201 |
docstring updates. |
| 1202 |
2010-07-27: ERF: Removed unused imprts and updated docs. |
| 1203 |
2010-07-27: FIX: Fix double delete |
| 1204 |
2010-07-26: NEW: use sphinx links in mdp.parallel docstring |
| 1205 |
2010-07-25: FIX: remove del's of nonexistent symbols |
| 1206 |
2010-07-25: NEW: use fixup_namespace to clean up __import__ |
| 1207 |
2010-07-25: FIX: missing comma in __all__ was concatentating names |
| 1208 |
2010-07-25: OTH: add import of fixup_namespace without actually using it This |
| 1209 |
is seperate to check if there are no circular imports |
| 1210 |
|
| 1211 |
2010-07-24: FIX: remove nonexistent name from __all__ |
| 1212 |
2010-07-23: FIX: remove unused import of hotshot.log |
| 1213 |
2010-07-23: FIX: fix broken name delete |
| 1214 |
2010-07-23: FIX: fixed setup.py (remove refs to demo & contrib) |
| 1215 |
2010-07-23: FIX: convolution nodes are conditional on scipy.signal |
| 1216 |
2010-07-23: FIX: remove del of nonexistent symbol |
| 1217 |
2010-07-23: FIX: move pp_simple_slave_test.py to examples |
| 1218 |
2010-07-23: FIX: regenerate bimdp wrappers for classifiers |
| 1219 |
2010-07-23: ERF: convert parallel python tests to py.test and disable them |
| 1220 |
2010-07-23: FIX: TODO had the wrond line ending. Sorry everybody :-( |
| 1221 |
2010-07-23: FIX: Fix tests in test_nodes_generic with svm classifier nodes. |
| 1222 |
Add arguments to ShogunSVMClassifier and LibSVMClassifier to |
| 1223 |
automatically set kernel and classifier on __init__. Add |
| 1224 |
parameters to test_nodes_generic to use these arguments. |
| 1225 |
|
| 1226 |
2010-07-23: DOC: updated TODO list |
| 1227 |
2010-07-23: NEW: Added new tests for Convolution2DNode |
| 1228 |
2010-07-23: FIX: removed contrib directory |
| 1229 |
2010-07-23: FIX: duplicate conftest.py twice, so py.test can be run without |
| 1230 |
options, try 3 Previous version worked only under cygwin when on |
| 1231 |
windows. |
| 1232 |
|
| 1233 |
2010-07-23: ERF: PEP8 |
| 1234 |
2010-07-23: DOC: The reason of the strange transformations in CumulatorNode |
| 1235 |
2010-07-23: FIX: Automatic test now cover everything, fixed a lot of stuff in |
| 1236 |
the meanwhile! |
| 1237 |
2010-07-23: FIX: NormalNoise did not function properly |
| 1238 |
2010-07-23: FIX: KMeansNode failed on some special cases |
| 1239 |
2010-07-23: FIX: duplicate conftest.py twice, so py.test can be run without |
| 1240 |
options, try 2 Previous version worked only when PYTHONPATH |
| 1241 |
included mdp-toolkit/. |
| 1242 |
|
| 1243 |
2010-07-23: ERF: Migrated the demos to the examples. Removed the tutorial |
| 1244 |
demo. |
| 1245 |
2010-07-23: FIX: duplicate conftest.py twice, so py.test can be run without |
| 1246 |
options Note: bimdp/test/conftest.py is a symbolic link to |
| 1247 |
bindmp/test/conftest.py. |
| 1248 |
OTOH, conftest.py in the top directory is a little different. |
| 1249 |
|
| 1250 |
2010-07-23: ERF: remove old run_coverage script The new way is: |
| 1251 |
py.test --figleaf and the result go into html/. |
| 1252 |
|
| 1253 |
2010-07-23: OTH: Renamed NEW_DEVELOPER_INFO.txt → NEW_DEVELOPER_INFO. |
| 1254 |
2010-07-23: ERF: Made tests for SVM classifiers, Convolution2D conditional on |
| 1255 |
presence of nodes |
| 1256 |
2010-07-23: Merge branch 'test_cleanup' Conflicts: |
| 1257 |
mdp/test/_tools.py |
| 1258 |
|
| 1259 |
2010-07-23: FIX: fix broken conditional test for caching |
| 1260 |
2010-07-23: ERF: simplify testing by providing commonly used imports in |
| 1261 |
test/_tools All testing files can just say |
| 1262 |
from _tools import * and the get numx, numx_rand, mult, ... |
| 1263 |
|
| 1264 |
Py.test is not imported into _tools, so that it is easier to |
| 1265 |
possible to import files which don't actually depend on py.test |
| 1266 |
when it is not available. |
| 1267 |
|
| 1268 |
Also includes some tiny naming and space corrections. |
| 1269 |
|
| 1270 |
2010-07-23: NEW: Conditional tests decorator; decorated test_caching tests |
| 1271 |
2010-07-23: FIX: use the right TESTDECIMALS in test_utils_generic |
| 1272 |
2010-07-23: ERF: convert utils tests to py.test |
| 1273 |
2010-07-23: NEW: add QuadraticFormException for QuadraticForm errors |
| 1274 |
2010-07-23: ERF: Caching now imported conditional on joblib |
| 1275 |
2010-07-23: Merge branch 'master' of |
| 1276 |
ssh+git://mdp-toolkit.git.sourceforge.net/gitroot/mdp-toolkit/mdp-toolkit |
| 1277 |
2010-07-23: Merge branch 'master' into HEAD Conflicts: |
| 1278 |
mdp/__init__.py |
| 1279 |
|
| 1280 |
2010-07-23: ERF: Added *Cumulator and ClassifierNode to __all__ |
| 1281 |
2010-07-23: OTH: Trying to push changes |
| 1282 |
2010-07-23: ERF: Caching tests moved to own file, updated to py.test style |
| 1283 |
2010-07-23: ERF: de-obfuscated code in config object |
| 1284 |
2010-07-23: Merge branch 'sprint_cache' Conflicts: |
| 1285 |
mdp/__init__.py |
| 1286 |
mdp/extension.py |
| 1287 |
mdp/nodes/__init__.py |
| 1288 |
mdp/nodes/convolution_nodes.py |
| 1289 |
mdp/test/test_extension.py |
| 1290 |
mdp/test/test_nodes.py |
| 1291 |
mdp/utils/__init__.py |
| 1292 |
mdp/utils/routines.py |
| 1293 |
|
| 1294 |
2010-07-23: FIX: Windows specific fix. |
| 1295 |
2010-07-23: OTH: whitespaces |
| 1296 |
2010-07-23: ERF: make generic_test_factory a public function It is now even |
| 1297 |
documented. |
| 1298 |
|
| 1299 |
2010-07-23: OTH: Fixed whitespaces in convolution nodes |
| 1300 |
2010-07-23: DOC: describe the complicated big_nodes argument of |
| 1301 |
_generic_test_factory |
| 1302 |
2010-07-23: NEW: tests can be run within python shell. You can run the tests |
| 1303 |
with import mdp mdp.test() |
| 1304 |
|
| 1305 |
you don't need py.test installed! |
| 1306 |
|
| 1307 |
2010-07-23: Merge branch 'sprint_conv2' |
| 1308 |
2010-07-23: ERF: Made import of Convolution2DNode conditional on presence of |
| 1309 |
scipy |
| 1310 |
2010-07-23: ERF: Added method in config object to check for existence of |
| 1311 |
arbitrary module by name |
| 1312 |
2010-07-22: FIX: remove nonexistent import from __all__ |
| 1313 |
2010-07-22: ERF: Adding numpy and scipy to config. |
| 1314 |
2010-07-22: Merge branch 'sprint_variadic_cumulator' |
| 1315 |
2010-07-22: FIX: fix broken test_gradient. It was not useing the numx, |
| 1316 |
numx_rand convention!!! |
| 1317 |
>:-( |
| 1318 |
|
| 1319 |
2010-07-22: FIX: Fixing an error in contrib tests. |
| 1320 |
2010-07-22: ERF: migrated bimdp tests to py.test |
| 1321 |
2010-07-22: FIX: remove unexistent test name |
| 1322 |
2010-07-22: NEW: Added convolution class and tests |
| 1323 |
2010-07-22: ERF: Merged master to convolution branch and removed vestigial |
| 1324 |
file |
| 1325 |
2010-07-22: DOC: Update docs |
| 1326 |
2010-07-22: ERF: Cache extension migrated to own directory, tests updated |
| 1327 |
2010-07-22: NEW: Init for caching module |
| 1328 |
2010-07-22: OTH: Attempted refactoring cache extension in own directory, got |
| 1329 |
metaclass conflict |
| 1330 |
2010-07-22: ERF: update gitignore for py.test gnerated files |
| 1331 |
2010-07-22: Merge branch 'test_framework' Conflicts: |
| 1332 |
bimdp/test/test_binode.py |
| 1333 |
mdp/nodes/misc_nodes.py |
| 1334 |
|
| 1335 |
2010-07-22: NEW: Context manager for caching extension |
| 1336 |
2010-07-22: FIX: Fixed documentation and call with wrong argument |
| 1337 |
2010-07-22: DOC: Fixed error in documentation of context manager |
| 1338 |
2010-07-22: ERF: Deleted forgotten print statement |
| 1339 |
2010-07-22: NEW: activate_caching, deactivate_caching |
| 1340 |
2010-07-22: ERF: Simplified extension context manager |
| 1341 |
2010-07-22: ERF: port contrib and hinet to py.test |
| 1342 |
2010-07-22: ERF: Cache directory can be changed or picked at random (default) |
| 1343 |
2010-07-22: Merge branch 'sprint_info_refactoring' |
| 1344 |
2010-07-22: ERF: Renamed ‘Requirements’ → ‘MDPConfiguration’. Changed the API |
| 1345 |
style a little. Needs more love with version checking. |
| 1346 |
2010-07-22: ERF: Caching extension moved to own file |
| 1347 |
2010-07-22: ERF: Tests for cache extension |
| 1348 |
2010-07-22: DOC: add note about where the links must be changed in sphinx rst |
| 1349 |
2010-07-22: FIX: Exception missing mdp module. |
| 1350 |
2010-07-22: ERF: ClassifyCumulator also inherits from VariadicCumulator. A |
| 1351 |
little bit useless atm, since many methods are overwritten. |
| 1352 |
2010-07-22: FIX: UTF-8 problem. I want to have Python 3 only… |
| 1353 |
2010-07-22: FIX: broken linear regression test Use non-random input value, so |
| 1354 |
that it doesn't fail randomly. |
| 1355 |
|
| 1356 |
2010-07-22: NEW: added hooks and a command-line option in py.test Use --seed |
| 1357 |
to set the random seed. |
| 1358 |
|
| 1359 |
Added report of extended configuration infos before and after |
| 1360 |
testing reports. Added unit test for random seed [should only |
| 1361 |
fail in case of a bug in py.test and/or numpy]. |
| 1362 |
|
| 1363 |
2010-07-22: FIX: make generated hinet tests pass |
| 1364 |
2010-07-22: ERF: Cache extension now based on joblib |
| 1365 |
2010-07-22: NEW: Added VariadicCumulator with mdp.Cumulator being a special |
| 1366 |
case of it. VariadicCumulator adds as many automatic fields as are |
| 1367 |
specified in the initialisation function. |
| 1368 |
2010-07-22: ERF: use py.test.raises for exception testing |
| 1369 |
2010-07-22: FIX: correct PCANode generation |
| 1370 |
2010-07-22: ERF: extract PCANode generation to a helper function |
| 1371 |
2010-07-22: ERF: migrate tests for hinet to py.test |
| 1372 |
2010-07-22: FIX: Switchboard.is_invertible() cannot be a staticmethod |
| 1373 |
2010-07-22: ERF: remove previously converted tests from test_nodes.py |
| 1374 |
2010-07-22: ERF: migrate tests for CuBICA- and TDSEPNodes to py.test |
| 1375 |
2010-07-22: ERF: migrate tests for HistParade- and TimeFrameNodes to py.test |
| 1376 |
2010-07-22: ERF: migrate tests for EtaComputerNode to py.test |
| 1377 |
2010-07-22: ERF: migrate tests for GrowingNeuralGasNode to py.test |
| 1378 |
2010-07-22: ERF: migrate tests for NoiseNode to py.test |
| 1379 |
2010-07-22: ERF: migrate tests for FDANode to py.test |
| 1380 |
2010-07-22: ERF: migrate tests for GaussianClassifier to py.test |
| 1381 |
2010-07-22: ERF: migrate tests for FANode to py.test |
| 1382 |
2010-07-22: ERF: migrate tests for ISFANode to py.test |
| 1383 |
2010-07-22: ERF: migrate tests for RMB*Node to py.test |
| 1384 |
2010-07-22: ERF: move spinner to _tools |
| 1385 |
2010-07-22: ERF: migrate tests for LinearRegressionNode to py.test |
| 1386 |
2010-07-22: ERF: migrate tests for CutoffNode to py.test |
| 1387 |
2010-07-22: ERF: migrate tests for HistogramNode to py.test |
| 1388 |
2010-07-22: ERF: migrate tests for RBFExpansionNode to py.test |
| 1389 |
2010-07-22: ERF: migrate tests for AdaptiveCutoffNode to py.test |
| 1390 |
2010-07-22: ERF: migrate tests for SFA2Node to py.test |
| 1391 |
2010-07-22: ERF: migrate tests for SFANode to py.test |
| 1392 |
2010-07-22: ERF: migrate tests for WhiteningNode to py.test |
| 1393 |
2010-07-22: ERF: migrate tests for PCANode to py.test |
| 1394 |
2010-07-22: ERF: migrate tests for PolynomialExpansionNode to py.test |
| 1395 |
2010-07-22: ERF: migrate covariance Node tests to py.test |
| 1396 |
2010-07-22: ERF: add node copying, saving and training tests |
| 1397 |
2010-07-22: ERF: move BogusNode* to _tools |
| 1398 |
2010-07-22: FIX: fix python2.5 compatibility |
| 1399 |
2010-07-21: FIX: Don’t include utils until it is needed. |
| 1400 |
2010-07-21: ERF: Drop Python 2.4, use built-in all(). |
| 1401 |
2010-07-21: ERF: Fix some of the issues with contrib.__all__ and dependencies. |
| 1402 |
2010-07-21: FIX: Fixing bugs with scheduling. Avoid printing while mdp is |
| 1403 |
imported. |
| 1404 |
|
| 1405 |
Rename _info() to req.info() in MDPVersionCallable. req.info() is |
| 1406 |
a different object but it’s exact layout is not important for |
| 1407 |
scheduling. |
| 1408 |
2010-07-21: ERF: add tests for RBFExpansionNode |
| 1409 |
2010-07-21: FIX: don't use self in staticmethod |
| 1410 |
2010-07-21: ERF: add generated tests for FastICA |
| 1411 |
2010-07-21: ERF: generate tests for reversing |
| 1412 |
2010-07-21: ERF: use any and all to check conditions |
| 1413 |
2010-07-21: BRK: make is_trainable staticmethod where possible |
| 1414 |
2010-07-21: BRK: make is_invertible staticmethod where possible |
| 1415 |
2010-07-21: ERF: Everybody loves nicer output |
| 1416 |
2010-07-21: OTH: implemented obvious thing of decorating the execute method |
| 1417 |
with joblib, it does not work; breaks the caching extension |
| 1418 |
2010-07-21: ERF: Remove all the cool logic and decorators because __dict__ is |
| 1419 |
not ordered and all the methods are executed in arbitrary order. |
| 1420 |
2010-07-21: Merge commit '38664d80675ea7304ee28b460ba88c547b55629e' |
| 1421 |
2010-07-20: NEW: Added a new Requirements object which checks for available |
| 1422 |
features. Current version needs some discussion/rework because |
| 1423 |
class methods are not initialised in order... |
| 1424 |
|
| 1425 |
2010-07-21: ERF: generate dimdtypeset consistency tests |
| 1426 |
2010-07-21: FIX: Fixed the remaining broken unittests. |
| 1427 |
2010-07-21: ERF: generate outputdim_consistency tests |
| 1428 |
2010-07-21: NEW: Adding the ‘official’ backport for OrderedDict to our |
| 1429 |
routines.py. |
| 1430 |
2010-07-21: FIX: Updated the bimdp.hinet package to work with the new |
| 1431 |
stop_training. |
| 1432 |
2010-07-21: ERF: world's first automatic test generation |
| 1433 |
2010-07-21: OTH: Merged cache branch with convolution branch |
| 1434 |
2010-07-21: FIX: Updated the BiFlowNode to work with new stop_training. |
| 1435 |
2010-07-21: DOC: Improved documentation for Convolution2DNode |
| 1436 |
2010-07-21: FIX: Fixed the failing BiFlow unittest. |
| 1437 |
2010-07-21: ERF: Convolution2DNode now supports FFT convolution, checks for |
| 1438 |
the validity of all of its argments; the new tests are in a file |
| 1439 |
outside the repository, waiting for the new test framework |
| 1440 |
2010-07-21: FIX: Fixed the inspection by removing the stop_message reference. |
| 1441 |
2010-07-21: BRK: Simplified the coroutine decorator according the stop_message |
| 1442 |
removal. |
| 1443 |
2010-07-21: BRK: Updated and simplified BiFlow for the new stop_training |
| 1444 |
specification. Not tested yet. |
| 1445 |
2010-07-21: ERF: Added output dimension check to IdentityNode. |
| 1446 |
2010-07-21: ERF: ported test_flows to py.test |
| 1447 |
2010-07-21: ERf: ported test_graph.py to py.test |
| 1448 |
2010-07-21: ERF: ported test_parallelflows to py.test |
| 1449 |
2010-07-21: ERF: ported test_parallelhinet.py to py.test |
| 1450 |
2010-07-21: ERF: ported test_parallelnodes.py to py.test |
| 1451 |
2010-07-21: ERF: ported test_process_schedule to py.test |
| 1452 |
2010-07-21: ERF: ported test_extension to py.test |
| 1453 |
2010-07-21: FIX: caching extension now works with non-contiguous arrays |
| 1454 |
2010-07-21: OTH: Updated bimdp todo list. |
| 1455 |
2010-07-21: BRK: Started with change of the stop_training signature in bimdp |
| 1456 |
and removed stop_message. The binode file is basically done. |
| 1457 |
2010-07-21: NEW: New utility function to generate Gabor wavelets |
| 1458 |
2010-07-21: ERF: ported test_classifier to py.test |
| 1459 |
2010-07-21: ERF: added new testing tools file. |
| 1460 |
2010-07-21: ERF: ported test_schedule to py.test |
| 1461 |
2010-07-20: FIX: globally remove trailing whitespace Used |
| 1462 |
http://svn.python.org/projects/python/trunk/Tools/scripts/reindent.py |
| 1463 |
. |
| 1464 |
|
| 1465 |
2010-07-20: FIX: broken import |
| 1466 |
2010-07-20: Merge branch 'py3k', remote branch 'origin/sprint_gradient' into |
| 1467 |
master Conflicts: |
| 1468 |
bimdp/inspection/trace_inspection.py |
| 1469 |
|
| 1470 |
2010-07-20: FIX: Removed use of new.instancemethod from inspection. |
| 1471 |
2010-07-20: FIX: typo in Flow.__setitem__ |
| 1472 |
2010-07-20: ERF: use context manager for file access |
| 1473 |
2010-07-20: FIX: don't create a list that is ignored RefactoringTool: |
| 1474 |
Warnings/messages while refactoring: RefactoringTool: ### In file |
| 1475 |
/home/zbyszek/mdp/mdp-toolkit/build/py3k/mdp/linear_flows.py ### |
| 1476 |
RefactoringTool: Line 481: You should use a for loop here |
| 1477 |
|
| 1478 |
2010-07-20: FIX: Fixed the integer division in bimdp. |
| 1479 |
2010-07-20: FIX: remove bogus import in test/__init__.py |
| 1480 |
2010-07-20: FIX: get rid of the "callable" variable and fix the scheduler |
| 1481 |
shutdown An explicit flush was needed. |
| 1482 |
|
| 1483 |
2010-07-20: NEW convolution2d node |
| 1484 |
2010-07-20: ERF: esthetic change: activate heart beat spinner |
| 1485 |
2010-07-20: FIX: another integer division problem |
| 1486 |
2010-07-20: FIX: don't use None in comparisons |
| 1487 |
2010-07-20: FIX: integer division again |
| 1488 |
2010-07-20: FIX: fixed handling of binary mode for stdin and stdout for pickle |
| 1489 |
2010-07-20: FIX: remove unused import |
| 1490 |
2010-07-20: FIX: don't compare float with NoneType In Python2.6 None is |
| 1491 |
smaller than any float, but in Python3 the comparison fails with |
| 1492 |
TypeError. |
| 1493 |
|
| 1494 |
2010-07-20: FIX: whitespace |
| 1495 |
2010-07-20: FIX: correct relative imports in process_schedule.py |
| 1496 |
2010-07-20: FIX: another integer division problem |
| 1497 |
2010-07-20: FIX: whitespace |
| 1498 |
2010-07-20: FIX: add x before range git grep -l 'for.*[^x]range' |
| 1499 |
mdp/test/|xargs sed -i -r 's/for(.*[^x])in range/for\1in xrange/g' |
| 1500 |
|
| 1501 |
2010-07-20: FIX: matlab→python and py3k cleanups a = a + 1 becomes a += 1 |
| 1502 |
/ becomes // range() becomes xrange() useless cast removed |
| 1503 |
exception variable doesn't leak out after the catch block anymore |
| 1504 |
|
| 1505 |
2010-07-19: FIX: whitespace |
| 1506 |
2010-07-20: FIX: zip returns a generator in py3k, we need a list |
| 1507 |
2010-07-19: FIX: fixed integer division |
| 1508 |
2010-07-19: FIX: fixed calling unbound class methods in extensions. unbound |
| 1509 |
methods have been removed in python3. |
| 1510 |
2010-07-19: ERF: refactored py3tool |
| 1511 |
2010-07-19: FIX: use sets not lists in testing GraphNode |
| 1512 |
2010-07-19: FIX: open files in binary mode when loading a pickle |
| 1513 |
2010-07-19: MISC: gitignore build stuff |
| 1514 |
2010-07-19: ERF: use 'open' function instead of the deprecated 'file' in |
| 1515 |
mdp/demo |
| 1516 |
2010-07-19: FIX: missing import in setup.py [2] |
| 1517 |
2010-07-19: FIX: missing import in setup.py |
| 1518 |
2010-07-19: NEW: python3 setup logic (from numpy) |
| 1519 |
2010-07-19: ERF: use 'open' function instead of the deprecated 'file'. |
| 1520 |
2010-07-20: ERF: Restricted the gradient default implementation to the |
| 1521 |
IdentityNode, now raise an exception as the default. |
| 1522 |
2010-07-20: ERF: Removed the explicitly defined IdentityBiNode, use the |
| 1523 |
autogenerated one instead. |
| 1524 |
2010-07-20: NEW context manager for extensions |
| 1525 |
2010-07-20: NEW basic implementation of the cache execute extension |
| 1526 |
2010-07-19: ERF: Added a gradient unittest with a small network. |
| 1527 |
2010-07-19: ERF: Improved one layer gradient unittest into a functional test. |
| 1528 |
Insignificant cleanup in gradient module. |
| 1529 |
2010-07-19: FIX: Fixed the switchboard gradient and added unittest. |
| 1530 |
2010-07-19: NEW: Added the switchboard gradient implementation with a test |
| 1531 |
(more unittests are needed). |
| 1532 |
2010-07-19: NEW: Added gradient implementation for layer with some unittests |
| 1533 |
(more are needed). Added check in gradient base node that node has |
| 1534 |
finished training. |
| 1535 |
2010-07-19: DOC: Added the new project guidelines. |
| 1536 |
2010-07-19: ERF: use integer division |
| 1537 |
2010-07-19: NEW: Added gradient extension with some implementations for simple |
| 1538 |
nodes (no hinet yet). |
| 1539 |
2010-07-02: Added kwargs to SenderBiNode init to enable multiple Inheritance. |
| 1540 |
2010-07-02: Fixed bug in inspection: If nodes were added in later training |
| 1541 |
phases the clickability broke for all nodes. Again this affected |
| 1542 |
XSFA. Changed text in inspection from layer to node. |
| 1543 |
2010-07-02: Updated the inspection example (more training phases for better |
| 1544 |
illustration). Tiny doc update in SenderBiNode. |
| 1545 |
2010-07-02: Fixed bug for normal node in BiFlowNode (this broke the XSFA |
| 1546 |
inspection example). |
| 1547 |
2010-07-01: Simplified the inspection msg printing. |
| 1548 |
2010-07-01: Improved the inspection message printing (alphabetic sorting and |
| 1549 |
highlighted keywords). |
| 1550 |
2010-06-29: Improved the key order in the inspection msg HTML view. |
| 1551 |
2010-06-29: Improved the array HTML display in the inspection. |
| 1552 |
2010-06-29: Fixed one corner case bug in coroutine decorator and added more |
| 1553 |
unittests. Updated bimdp todo list. Small docstring update in |
| 1554 |
switchboard. |
| 1555 |
2010-06-16: Merged the coroutine mixin class into the BiNode base class. Added |
| 1556 |
_bi_reset hook method that should be overwritten instead of |
| 1557 |
bi_reset. This is more consistent with the standard MDP design. |
| 1558 |
The bi_reset template method now clears the coroutine table. |
| 1559 |
Dervied BiNode classes should rename their bi_reset to _bi_reset, |
| 1560 |
but unless they use the coroutine decorator there is currently no |
| 1561 |
negative effect for not doing so. Updated the BiMDP todo list. |
| 1562 |
2010-06-15: Fixed duplicate classifier nodes in autogenereated code. |
| 1563 |
2010-06-14: Added check and unittest for corner case of a directly terminating |
| 1564 |
coroutine. |
| 1565 |
2010-06-13: Reverted my misguided fixes to the metaclass wrapping helpers. |
| 1566 |
2010-06-11: Small comments update. |
| 1567 |
2010-06-11: Updated the codecorator to work with the updated NodeMetaclass |
| 1568 |
wrapping helper functions. Add new feature to specify argument |
| 1569 |
default values. Changed the _coroutine_instances to be initially |
| 1570 |
None for better efficiency. |
| 1571 |
2010-06-11: Rewrote the NodeMetaclass wrapping helper functions. The old |
| 1572 |
versions contained a bug (default values would prevent user |
| 1573 |
defined values from being passed to the function). The new version |
| 1574 |
is also arguably more elegant and flexible. A unittest was added |
| 1575 |
for parts of the new functionality. |
| 1576 |
2010-06-11: Fixed the n argument in PCANode inverse. |
| 1577 |
2010-06-09: Added new coroutine decorator and mixin for BiMDP, allowing easy |
| 1578 |
continuations (without boilerplate code or explicit state |
| 1579 |
machines). |
| 1580 |
2010-06-09: Fixed bug in SenderBiNode. |
| 1581 |
2010-06-09: Fixed typo. |
| 1582 |
2010-06-09: Fixed broken SFA execute range argument and renamed it to n. Added |
| 1583 |
unittests for PCA and SFA n arguments. Added some missing SFANode |
| 1584 |
attributes to init. |
| 1585 |
2010-06-09: Fixed bug in BiFlowNode. |
| 1586 |
2010-05-31: Added loop check for previous button in slideshow. |
| 1587 |
2010-05-27: Numpy still uses the deprecated string exceptions, which were |
| 1588 |
naturally not caught during inspection. I now changed inspection |
| 1589 |
to catch everything. |
| 1590 |
2010-05-19: Added check for loop setting in slideshow next(). |
| 1591 |
2010-05-17: Very small cosmetic updates. |
| 1592 |
2010-05-17: fixed other two python2.4 incompatibilities in demos. |
| 1593 |
2010-05-15: fixed bug with ProcessScheduler when mdp installed by debian |
| 1594 |
package. Now it works correctly even when multiple versions of mdp |
| 1595 |
are installed, for example a system version via debian and a local |
| 1596 |
git version. A unit test has been added to make sure parent and |
| 1597 |
child processes are running the very same mdp. Thanks to Yaroslav |
| 1598 |
Halchenko for pointing out the problem! |
| 1599 |
|
| 1600 |
------------------------------------------------------------------------------- |
| 1601 |
MDP-2.6: |
| 1602 |
|
| 1603 |
2010-05-14: fix spurious shogun warning. |
| 1604 |
2010-05-14: applied patch from Rike to re-enable some of the working Shogun |
| 1605 |
tests. |
| 1606 |
2010-05-14: fixed bug when using most recent numpy (1.4.1). utils.get_dtype |
| 1607 |
was failing when called with 'All' argument. most of MDP unittests |
| 1608 |
were failing! |
| 1609 |
2010-05-14: added stylesheet files to setup script. only *.py files are |
| 1610 |
installed by default. |
| 1611 |
2010-05-14: modified mechanism to get git revision in order to properly handle |
| 1612 |
the case when git fails. added stylesheets to MANIFEST template |
| 1613 |
2010-05-14: several additions to distribution files |
| 1614 |
2010-05-14: removed NaiveBayesClassifer from automatically generated nodes in |
| 1615 |
bimdp. regenerated node wrappers |
| 1616 |
2010-05-14: fixed bug in GaussianClassifierNode and FDANode when training with |
| 1617 |
a list of labels. Removed obsolete 'uniq' function in utils. we |
| 1618 |
can use builtin set instead. |
| 1619 |
2010-05-14: disabled Shogun tests. they often fail withtou apparent reasons. |
| 1620 |
try for example: python2.5 -c 'import mdp; |
| 1621 |
mdp.test(seed=257416710)' |
| 1622 |
2010-05-14: Added a maximum iteration constraint to the k-means classifier. |
| 1623 |
2010-05-14: Moved the NaiveBayesClassifier to the demo. |
| 1624 |
2010-05-14: Fixed a bug where the Classifier rank() method returned items in |
| 1625 |
the reverse order. |
| 1626 |
2010-05-13: LLENode was not fixed ;-) disabled 2D test and added reminder to |
| 1627 |
TODO |
| 1628 |
2010-05-13: fixed ever failing LLENode test. |
| 1629 |
2010-05-13: added info about LibSVM and Shogun in mdp.info() |
| 1630 |
2010-05-13: first take at disabling SVM tests when libraries are not available |
| 1631 |
2010-05-12: Updated the todo list to reflect the latest developments. |
| 1632 |
2010-05-12: Adjusted the BiClassifier node according to the renamings. |
| 1633 |
2010-05-12: Renamings in the classifier nodes (classify->label, cl->labels). |
| 1634 |
2010-05-11: removed code duplication from test/__init__.py |
| 1635 |
2010-05-11: added info function to display infos about MDP and dependencies |
| 1636 |
2010-04-29: Added ability to exit BiFlow with -1 target value, making it |
| 1637 |
easier to use inverse execution. |
| 1638 |
2010-04-22: Inserted the new sys.path behavior also in the main clause. |
| 1639 |
2010-04-22: Changed the sys.path behavior in subprocesses to fix problems with |
| 1640 |
multiple MDP versions on a system. |
| 1641 |
2010-04-14: Fall back on default if specified browser is not available. |
| 1642 |
2010-04-14: Added tracebacks to exception handling during debug inspection. |
| 1643 |
2010-04-12: Added ability to specify a specific browser for slideshows, also |
| 1644 |
changed the argument name. |
| 1645 |
2010-04-09: Rewrote K-Means training to be aligned with the Cumulator |
| 1646 |
training. |
| 1647 |
2010-04-09: Adapted the SVM Classifiers to use the new ClassifierCumulator |
| 1648 |
class. |
| 1649 |
2010-04-03: Made a Cumulator version for simple Classifiers. |
| 1650 |
2010-04-06: Made GaussianClassifierNode inherit from ClassifierNode. |
| 1651 |
2010-04-04: Simplified the inheritance names in classifier nodes’ |
| 1652 |
declarations. |
| 1653 |
2010-03-29: Patch from Benjamin Schrauwen to fix the method polution issue. |
| 1654 |
2010-03-29: Fixed the dysfunctional unittest, it now breaks, currently no |
| 1655 |
solution. |
| 1656 |
2010-03-29: Added test for pp scheduler with real flow. |
| 1657 |
2010-03-28: Improved the extension verbose printing. |
| 1658 |
2010-03-28: Fixed one more issue (method polution) and added a test. |
| 1659 |
2010-03-28: Added new failing test, caused by polution. |
| 1660 |
2010-03-28: Added unittest for the special behavior when multiple extensions |
| 1661 |
extend the same method. |
| 1662 |
2010-03-28: Patch from Benjamin Schrauwen for extension mechanism. |
| 1663 |
2010-03-27: Slightly changed the behavior of targeted message keys: They are |
| 1664 |
now removed even if they are not args of the method. Updated the |
| 1665 |
unittests accordingly. |
| 1666 |
2010-03-27: Fixed bug for layer containing a single node. |
| 1667 |
2010-03-26: Added BiClassifier and automatic creation of BiMDP versions for |
| 1668 |
MDP classifiers (currently except the SVN based ones). |
| 1669 |
2010-03-25: Applied patch from Benjamin Schrauwen, the BiMDP tests now all |
| 1670 |
pass. |
| 1671 |
2010-03-25: Changed the BiMDP msg key seperator from "=>" to "->" to allow for |
| 1672 |
more consistent convention in the future, when more options might |
| 1673 |
be added. Updated the todos accordingly. |
| 1674 |
2010-03-25: Undid the changes to the extension mechanism, they are now |
| 1675 |
available in a seperate branch. |
| 1676 |
2010-03-25: Enable inheritance in extensions. Patch thanks to Benjamin |
| 1677 |
Schrauwen! (cherry picking from commit |
| 1678 |
aacbc906be0f79e66124085e1c51c35a0aee731d) Enable inheritance in |
| 1679 |
extensions. Test thanks to Benjamin Schrauwen! (cherry picking |
| 1680 |
from commit 89b7facb223aff09d4581ce5aa68d07a8ef47b1b) |
| 1681 |
2010-03-25: Enable inheritance in extensions. Test thanks to Benjamin |
| 1682 |
Schrauwen! |
| 1683 |
2010-03-25: Enable inheritance in extensions. Patch thanks to Benjamin |
| 1684 |
Schrauwen! |
| 1685 |
2010-03-24: Python fixes for compatibility with 2.4. |
| 1686 |
2010-03-24: Reverted some of the Python 2.6 (and Numpy 1.4) only changes. |
| 1687 |
2010-03-24: Added verbose argument to mdp.activate_extension. |
| 1688 |
2010-03-23: Added a section about dealing with testing branches. |
| 1689 |
2010-03-22: Added unittest for bimdp parallel stuff and updated another. |
| 1690 |
Update the todos. |
| 1691 |
2010-03-22: Improved the CloneBiLayer message data splitting and added |
| 1692 |
corresponding unittest. |
| 1693 |
2010-03-22: Slightly changed the message handing in BiSwitchboard and added |
| 1694 |
corresponding unittests. Small rename in BiNode. Small todo |
| 1695 |
update. |
| 1696 |
2010-03-16: Added a simple k-means clustering classifier. |
| 1697 |
2010-03-14: Added node_id support to BiFlowNode. |
| 1698 |
2010-03-13: Added utility function izip_stretched. |
| 1699 |
2010-03-12: Added test for adding node to flow. Moved one BiNode test to |
| 1700 |
correct class. |
| 1701 |
2010-03-11: Fixed bug in flow in-place adding method. |
| 1702 |
2010-03-11: Added a refcast to classify and prob methods. |
| 1703 |
2010-03-10: Fixed the remaining failing tests from Tiziano. It was caused by |
| 1704 |
an ambiguity of the BiFlow.train docstring, so this is now |
| 1705 |
explained in more detail. An additional check was added to BiFlow |
| 1706 |
to detect this kind of error. |
| 1707 |
2010-03-10: Refactored the additional argument check into seperate method, so |
| 1708 |
that it can be reused by derived flow classes (especially BiFlow). |
| 1709 |
2010-03-10: Fixed the test for FDA in a biflow. |
| 1710 |
2010-03-10: Added other failing tests for BiFlow. |
| 1711 |
2010-03-10: Added failing tests for BiFlows. |
| 1712 |
2010-03-10: Added special __add__ method to BiNode, creating a BiFlow. Also |
| 1713 |
added corresponding unittest. |
| 1714 |
2010-03-09: Fixed an assertion bug in Shogun test. |
| 1715 |
2010-03-09: Added unittests for recent stop_message magic. Also made some |
| 1716 |
small improvements and added for documentation for this. Made |
| 1717 |
small updates to the general bimdp docstring. |
| 1718 |
2010-03-09: Added a section on using git with Windows or Eclipse. Added a |
| 1719 |
section on following PEP 8. |
| 1720 |
2010-03-09: Added checks for input_dim. |
| 1721 |
2010-03-09: We need to have shogun 0.9 for our SVM Classifier. |
| 1722 |
2010-03-08: Updated NEW_DEVELOPER_INFO to ‘The git situation’. |
| 1723 |
2010-03-08: Fixed Py2.6 DeprecationWarning in RBMLearning test. |
| 1724 |
2010-03-08: Fixed a new bug in BiNode. Moved the SVN nodes to the classifier |
| 1725 |
part in the autogen module. |
| 1726 |
2010-03-08: Made a larger set of additions to the SVM tests. |
| 1727 |
2010-03-08: Updated the docstrings for the SVM classifiers. |
| 1728 |
2010-03-06: Fix: Forgot to set the classifier_type in libsvm. |
| 1729 |
2010-03-06: Added simple support for probability estimations in LibSVM through |
| 1730 |
our _prob() method. |
| 1731 |
2010-03-04: Renamed SVM_Node -> SVM_Classifier. |
| 1732 |
2010-03-02: Supplying arguments to the shogun classifier works now. Improved |
| 1733 |
the test suite. |
| 1734 |
2010-03-01: Lots of updates and refinements to shogun integration. |
| 1735 |
2010-02-28: Added some recognition for KernelMachines in shogun. |
| 1736 |
2010-02-27: Applied some fixes and optimisations to the SVM code. |
| 1737 |
2010-02-26: Added class _LabelNormalizer to ease the handling of label |
| 1738 |
mapping. |
| 1739 |
2010-02-26: ClassifierNode-Migration: Now use ClassifierNode._classify in |
| 1740 |
SVMNodes. |
| 1741 |
2010-02-26: Last commit before migration of SVM nodes to ClassifierNode. |
| 1742 |
2009-09-16: Made some improvements in SVM testing and overall structure. |
| 1743 |
2010-03-06: Improved the algorithm for _randomly_filled_hyperball(). Scaling |
| 1744 |
is much better now. |
| 1745 |
2010-03-06: Fixed a bug in ClassifierNode.rank(). It now expects a list of |
| 1746 |
feature vectors to function. |
| 1747 |
2010-03-06: Fixed one potential stop_message issue. |
| 1748 |
2010-03-06: Added magic in BiNode to call execute or inverse from |
| 1749 |
stop_message. |
| 1750 |
2010-02-27: Made the product helper method take advantage of Python 2.6. |
| 1751 |
2010-03-04: Moved release files in root of master branch. |
| 1752 |
2010-03-03: Added automatic detection of number of CPU cores. This is used in |
| 1753 |
the ProcessScheduler and the ThreadScheduler as the default value. |
| 1754 |
2010-03-02: Fixed bug left over from inspection cleanup. |
| 1755 |
2010-02-27: get_svn_revision --> get_git_revision |
| 1756 |
2010-02-25: Made a couple of smaller improvements in the inspections stuff. |
| 1757 |
Added a nice demo for customized inspections. |
| 1758 |
2010-02-22: Introducing a node for a simple discrete Hopfield model. |
| 1759 |
2010-02-22: Stupid bug in bool_to_sign fixed. |
| 1760 |
2010-02-23: Simplified the improved import precautions. |
| 1761 |
2010-02-23: Enabled calling bimdp.test(). Removed redundant code from autogen. |
| 1762 |
2010-02-23: Made the import process in the worker processes more robust, to |
| 1763 |
fix the problem with calling bimdp.test() from the command line. |
| 1764 |
2010-02-23: Small updated in the autogen code. |
| 1765 |
2010-02-23: Changed the automatic creation of BiNode versions for MDP nodes. |
| 1766 |
Now a module file is used that is created by a special script. |
| 1767 |
While less elegant than the original dynamic solution this is now |
| 1768 |
compatible with code checkers and is more transparent for users. |
| 1769 |
2010-02-23: Added general purpose routines bool_to_sign and sign_to_bool. |
| 1770 |
2010-02-22: Fixed mdp import bug in process scheduler (caused by the occurance |
| 1771 |
of mdp in the repo name). |
| 1772 |
2010-02-19: Delete mdp_dist. All future updates should be made in |
| 1773 |
release-manager branch. |
| 1774 |
2010-02-19: Removed remaining binet references. Renamed some inspection |
| 1775 |
classes to get the MDP out of the name. Updated the todo list. |
| 1776 |
2010-02-19: Need to check against numpy.all when comparing array elements. |
| 1777 |
2010-02-18: Corrected the common assert-statement-with-parentheses mistake. |
| 1778 |
2010-02-18: Restructured bimdp to follow the mdp structure. |
| 1779 |
2010-02-17: Renamed binet to bimdp. |
| 1780 |
2009-08-01: Typo fix in progressbar script and clarification for OS X (the |
| 1781 |
automatic size works there as well). |
| 1782 |
2010-02-17: Added .gitignore file for .pyc files and some generic patterns. |
| 1783 |
2010-02-14: Added option to automatically handle the source_paths in the |
| 1784 |
worker processes. |
| 1785 |
2010-02-13: Added more unittests for BiFlow. Fixed one bug in inspection. |
| 1786 |
Moved the JumpBiNode to the tests. |
| 1787 |
2010-02-11: Updated the parallel demo. |
| 1788 |
2010-02-11: Added a new thread based scheduler. Some very tiny updates to |
| 1789 |
ProcessScheduler. |
| 1790 |
2010-02-09: Added more BiNode unittests. Updated BiNode documentation. Updated |
| 1791 |
BiFlow unittests (not yet complete). |
| 1792 |
2010-01-30: Some small fixes in the inspection. Tried to use cgitd to show |
| 1793 |
nice HTML tracebacks, but so far this failed (FlowExceptionCR only |
| 1794 |
preserves a traceback string). |
| 1795 |
2010-01-29: Removed the updown module (this should be no longer needed). |
| 1796 |
2010-01-26: Added check in get_quadratic_form to make it more flexible (e.g. |
| 1797 |
when a derived class wants to use it in _stop_training). Also made |
| 1798 |
some small PEP8 updates. |
| 1799 |
2010-01-26: Updated the imports for the parallel package. Fixed some other |
| 1800 |
small issues. |
| 1801 |
2010-01-22: Fixed one issue in rectangular switchboards (the data in each |
| 1802 |
patch was transposed). Added corresponding unittests. |
| 1803 |
2010-01-11: Seriously beefed up the automatic BiNode creation to preserve the |
| 1804 |
init signature. This makes it compatible with the introspection in |
| 1805 |
the parallel default fork implementation. Also added a special |
| 1806 |
ParallelBiNode and made a tiny style improvement in BiNode. |
| 1807 |
2010-01-11: Added a default implementation for forking nodes. |
| 1808 |
2010-01-06: Updated the todos. Also one small improvement to the metaclass |
| 1809 |
modification. |
| 1810 |
2010-01-05: Some small fixes and cosmetic correction. |
| 1811 |
2010-01-05: Fixed one bug from hell in FlowNode (copy was broken for nested |
| 1812 |
FlowNodes when the copy happened after training). |
| 1813 |
2010-01-05: Fixed a bug in BiFlowNode. |
| 1814 |
2010-01-04: Modified the automatic node creation, now is compatible with |
| 1815 |
pickle and should be 100% transparent. Also fixed the switchboard |
| 1816 |
factory extension support. |
| 1817 |
2010-01-04: Put the automatic node creation into a function to avoid namespace |
| 1818 |
polution. Added automatic node creation for switchboards. There is |
| 1819 |
one remaining issue with pickling. |
| 1820 |
2010-01-04: Fixed bugs from refactoring. |
| 1821 |
2010-01-04: Small refactoring in switchboard factory extension (to allow |
| 1822 |
easier modification by derived classes). |
| 1823 |
2010-01-04: Added automatic creation of BiNode versions of all nodes in |
| 1824 |
mdp.nodes. |
| 1825 |
2010-01-04: Removed leftover code. |
| 1826 |
2010-01-04: Updated todos and docstring in init. Small cosmetic updates |
| 1827 |
elsewhere. |
| 1828 |
2010-01-03: Fixed bug of the new inspection training node highlighting |
| 1829 |
feature. |
| 1830 |
2010-01-03: Fixed one issue with stop messages in ParallelBiFlow, added a |
| 1831 |
related convenience check to CloneBiLayer. |
| 1832 |
2010-01-03: Small improvements and bugfixes with respect to stop_message. |
| 1833 |
2010-01-03: Added highlighting of the currently training node in the |
| 1834 |
inspection. |
| 1835 |
2010-01-02: Fixed a couple of bugs. The DBN example is now working again. |
| 1836 |
2010-01-02: Inspection was simplified, parallel did not need any changes. The |
| 1837 |
binet simplification should now be complete (but I have not run |
| 1838 |
real-world tests yet). |
| 1839 |
2010-01-02: Updated bilayer and biswitchboard (with some huge simplifications |
| 1840 |
in biswitchboard). Now parallel and inspection remain to be |
| 1841 |
updated. |
| 1842 |
2010-01-02: Added a new switchboard. |
| 1843 |
2010-01-01: Fixed some issues in BiFlow after simplification. Updated |
| 1844 |
BiFlowNode, this is hopefully done. |
| 1845 |
2009-12-30: Finished simplification of BiFlow. bihinet still needs to be |
| 1846 |
updated as well. |
| 1847 |
2009-12-25: Added /usr/share/dict/words to the list of dictionaries. |
| 1848 |
2009-12-23: Started with the binet simplification. BiNode should be more or |
| 1849 |
less finished, most of the rest is not yet done. In principle |
| 1850 |
everything should already work, there is just a large amount of |
| 1851 |
dead code around. |
| 1852 |
2009-12-22: Fixed one bug in the trace inspection (added check for msg being |
| 1853 |
None). |
| 1854 |
2009-12-22: Slightly modified the BiLayer data splitting behavior. |
| 1855 |
2009-12-22: Fixed bug in iterator / argument check for train: default argument |
| 1856 |
values were not considered. |
| 1857 |
2009-12-16: Renamed the 'html_representation' extension to 'html'. |
| 1858 |
2009-12-15: Added a copy method to FlowNode to enable delegation to internal |
| 1859 |
nodes. Added corresponding tests. |
| 1860 |
2009-12-13: Prettified the argument representations in the slides. |
| 1861 |
2009-12-12: Small update to SenderBiNode. |
| 1862 |
2009-12-12: Improved the exception display in the inspection slideshows. Fixed |
| 1863 |
missing reset in fascade. |
| 1864 |
2009-12-12: Fixed bug in BiFlow (same bug that was previously fixed in |
| 1865 |
BiFlowNode, just forgot to commit this). |
| 1866 |
2009-12-12: Fixed bug in BiFlowNode (relative target index was not translated |
| 1867 |
to absolute index during training). |
| 1868 |
2009-12-12: Increased the robustnes of the training inspection in debug mode |
| 1869 |
(it can now deal with exceptions before the stop training hook is |
| 1870 |
reached). |
| 1871 |
2009-12-11: clean up a couple of lines in RBM nodes |
| 1872 |
2009-12-09: Added a small parallelization demo to demonstrate and explore the |
| 1873 |
speedup. |
| 1874 |
2009-12-05: Fixed outdated default value in slideshow (could cause JS error in |
| 1875 |
special circumstances). |
| 1876 |
2009-12-04: added documentation about the convention for the train methods of |
| 1877 |
multiple training phase nodes. |
| 1878 |
2009-12-04: added Node.has_multiple_training_phases method. |
| 1879 |
2009-12-03: Modified check for missing argument in node training to also check |
| 1880 |
for too many arguments. Added a corresponding test. |
| 1881 |
2009-12-03: Added check for missing training arguments in Flow (which should |
| 1882 |
be provided by the iterator). Also edited the corresponding |
| 1883 |
unittest. |
| 1884 |
2009-12-03: Fixed small error in docstring for slideshow. |
| 1885 |
2009-12-02: well, NoiseNode *can* be pickled if you use pickle instead of |
| 1886 |
cPickle. I hope this bug is fixed in python2.6 or python3 ;-) |
| 1887 |
2009-12-02: raise error if trying to copying or saving the NoiseNode error |
| 1888 |
(bug in cPickle) |
| 1889 |
2009-12-02: added failing test for iterables mess in flow. |
| 1890 |
2009-12-01: Added missing super call. |
| 1891 |
2009-12-01: Fixed one evil bug: When the last training phase of a node was not |
| 1892 |
parallelizable in a ParallelCheckpointFlow then the checkpoint |
| 1893 |
function was not called. The fix required the addition of another |
| 1894 |
hook method in ParallelFlow, but is overall much cleaner than the |
| 1895 |
previous solution. Also improved the status messages for local |
| 1896 |
training phases. |
| 1897 |
2009-12-01: gaussian classifier still had overwritten train just to offer a |
| 1898 |
doc-string. putting the doc-string in _train is sufficient. |
| 1899 |
2009-11-29: Added new version of SenderBiNode (doesn't use branching). |
| 1900 |
2009-11-26: Fixed one biflow issue when in inverse mode (especially inside a |
| 1901 |
BiFlowNode). |
| 1902 |
2009-11-26: Updated some docstrings for the extension mechanism (they were |
| 1903 |
only refering to methods and not attributes). |
| 1904 |
2009-11-26: Fixed a bug in the inspection facade. |
| 1905 |
2009-11-22: Improved the extension unittests (now testing class attributes as |
| 1906 |
well). |
| 1907 |
2009-11-21: Improved the slideshow button text hints (they now also contain |
| 1908 |
the keyboard shortcuts). |
| 1909 |
2009-11-21: Fixed one embarrassing mistake in the node metaclasses. |
| 1910 |
2009-11-21: Added support for kwargs in training and execution inspection. |
| 1911 |
2009-11-16: Added a SimpleMarkovClassifier node. There is duplicated |
| 1912 |
functionality compared to the NaiveBayesClassifier, so the |
| 1913 |
NaiveBayesClassifier could be refactored to be based on the |
| 1914 |
SimpleMarkovClassifier. |
| 1915 |
2009-11-15: Fixed the signature overwriting in NodeMetaclass. Also renamed |
| 1916 |
some occurances of "kargs" to "kwargs" in the Node class |
| 1917 |
(previously both names were used in different methods). |
| 1918 |
2009-11-15: Fixed a bug in rank(). |
| 1919 |
2009-11-14: Moved ClassifierNode up one directory. |
| 1920 |
2009-11-14: added file with info for new developers |
| 1921 |
2009-11-13: Added a ClassifierNode with basic methods classify, rank and prob. |
| 1922 |
2009-11-13: Fixed one bug in the message container, thanks to Sven Dähne. |
| 1923 |
2009-11-09: Fixed one small bug in switchboard factory. Added switchboard |
| 1924 |
factory extension in binet. Updated todos. |
| 1925 |
2009-11-07: Added new switchboard factory node extension. This has been |
| 1926 |
factored out of the hinet planer and I will now use it in my |
| 1927 |
simulation software as well (therefore I wanted to integrate it |
| 1928 |
into MDP). Unittests and documentation are currently missing |
| 1929 |
(unittests will be somewhat tricky, not sure how deeply this |
| 1930 |
should be covered in the tutorial anyway). |
| 1931 |
2009-11-06: Rewrote some parts of the extension mechanism. It now works for |
| 1932 |
arbitrary attributes. One nice things is that this actually |
| 1933 |
reduced the number of lines of code. The tutorial has not yet been |
| 1934 |
updated. Attribute specific unittests are missing as well. |
| 1935 |
2009-11-06: Updated the todo list. |
| 1936 |
2009-11-05: Slightly improved the API for activating multiple extensions. |
| 1937 |
Discovered that the extension mechanism does not work for |
| 1938 |
classmethods/staticmethods, since they appear as non-callable |
| 1939 |
descriptors. I will change the extension mechanism to work for |
| 1940 |
arbitrary attributes ASAP. |
| 1941 |
2009-11-04: Small improvements in extension mechanism, added more comments. |
| 1942 |
Added missing Exception import in hinet init. |
| 1943 |
2009-11-04: Fixed a small slideshow issue. |
| 1944 |
2009-11-03: Completed the clickable inspection nodes. One can now click on a |
| 1945 |
node to jump to the next slide where it is reached. This does not |
| 1946 |
work in IE yet (the workaround would be simple, but don't want to |
| 1947 |
clutter the code with IE specific garbage). Also removed the AJAX |
| 1948 |
storage div (caused trouble due to duplicate ids and is not needed |
| 1949 |
anyway). |
| 1950 |
2009-11-03: Added support for active node ids in binet inspection. This will |
| 1951 |
make it possible to click on a node and directly jump to the next |
| 1952 |
time it is reached. The JS part of this is not yet implemented. |
| 1953 |
2009-11-03: Fixed the slideshow image CSS and added IE support. |
| 1954 |
2009-10-30: Cleaned up the JavaScript code in the binet inspection. |
| 1955 |
2009-10-27: Added checks in channel switchboard. |
| 1956 |
2009-10-21: Sorry, forgot to fix the webbrowser.open issue in binet. |
| 1957 |
2009-10-21: Fixed the webbrowser.open issue on MacOS as suggested by Rike. |
| 1958 |
2009-10-20: Tweaked hinet css. Extracted slideshow css into speparate file. |
| 1959 |
2009-10-20: Extracted CSS code into separate file. |
| 1960 |
2009-10-18: Fixed error in rhombic switchboard, added corresponding unittest. |
| 1961 |
2009-10-17: Added new features to the ChannelSwitchboard base class to get the |
| 1962 |
effective receptive fields in hierarchical networks. Also added |
| 1963 |
corresponding tests. Minor cosmetic updates in hinet translator. |
| 1964 |
2009-10-16: Added hinet translatation adapter for valid XHTML (instead of |
| 1965 |
HTML). Also added a corresponding unittest, but this does not test |
| 1966 |
the validity of the XHTML. |
| 1967 |
2009-10-16: Fixed two small issues in the hinet HTML representation. |
| 1968 |
2009-10-14: fixed bug in TimeFramesNode when input_dim was set explicitly. |
| 1969 |
thanks to anonymous sourceforge user [Bug ID: 2865106] |
| 1970 |
2009-10-14: Some CSS fixes and a fix in the demo. |
| 1971 |
2009-10-14: Added slideshow CSS for crisp rescaling, in anticipation of the |
| 1972 |
Firefox 3.6 beta release. |
| 1973 |
2009-10-12: Added feature that nicely line wraps the section ids in a |
| 1974 |
slideshow. |
| 1975 |
2009-10-07: Added more checks to DoubleRect2dSwitchboard. Some small |
| 1976 |
improvements. Updated the package init file. |
| 1977 |
2009-10-07: Reverted output_channels renaming in switchboard, to not create |
| 1978 |
compatibility issues. |
| 1979 |
2009-10-07: Many switchboard updates. A new ChannelSwitchbard base class was |
| 1980 |
introduced, many code improvements, HTML representations were |
| 1981 |
added for the new switchboard types. |
| 1982 |
2009-10-06: removed spurious print statement in QuadraticForm tests. |
| 1983 |
2009-10-04: Added a section on the extension mechanism in the tutorial (note |
| 1984 |
that this is a first draft). Also added a description of the |
| 1985 |
slideshow stuff (plus a minor update in the slideshow code). |
| 1986 |
2009-10-04: Migrated extension stuff to separate module. |
| 1987 |
2009-10-01: Fixed some issues in the new DoubleRhomb2dSwitchboard, added more |
| 1988 |
tests. |
| 1989 |
2009-09-29: Added a new switchboard class, DoubleRhomb2dSwitchboard. The tests |
| 1990 |
are currently incomplete and there are probably bugs in there. |
| 1991 |
2009-09-29: Some simplifications in the switchboard classes. |
| 1992 |
2009-09-29: Added a new switchboard class, DoubleRect2dSwitchboard. This will |
| 1993 |
probably be complemented in the future by a Rhomb2d switchboard |
| 1994 |
for multilayer networks. |
| 1995 |
2009-09-24: Fixed support for multiple slideshows in one file. Added multiple |
| 1996 |
slideshow demo. Allowed keyboard shortcuts for slideshows to be |
| 1997 |
turned off. |
| 1998 |
2009-09-23: Updated todos. |
| 1999 |
2009-09-23: Fixed the template indentation handling. Updated the binet |
| 2000 |
inspection to work with the new slideshow. Moved the manual tests |
| 2001 |
for slideshow and hinet to the demos folder. |
| 2002 |
2009-09-23: Some more cleanup for the slideshow stuff. |
| 2003 |
2009-09-23: Small update in binet for moving the hinet basic CSS to utils. |
| 2004 |
2009-09-23: Cleaned up the slideshow JavaScript code. Most importantly the use |
| 2005 |
of global variables is now minimized by using an object oriented |
| 2006 |
design with a closure. Some additional cleanup in the slideshow |
| 2007 |
classes was done as well. Added one helper function to directly |
| 2008 |
create an HTML slideshow. Moved the default CSS from hinet to |
| 2009 |
utils, so it can be used by both slideshow and hinet. Moved the |
| 2010 |
hinet HTML test/demo to the test folder and added a similar demo |
| 2011 |
for slideshow (using the MDP logo animation). |
| 2012 |
2009-09-19: added script to analyze code coverage during tests and output html |
| 2013 |
report; needs figleaf, |
| 2014 |
http://darcs.idyll.org/~t/projects/figleaf/doc/; mostly, we don't |
| 2015 |
test exception, but there are some more worrying parts, like we |
| 2016 |
don't test for for flow inversion... |
| 2017 |
2009-09-17: Added new utility function orthogonal_permutations. The function |
| 2018 |
helps avoiding deeply nested statements when several of several |
| 2019 |
arguments need to be tested against each other. |
| 2020 |
2009-09-14: Many thanks to Christian Hinze for pointing out a bug in |
| 2021 |
QuadraticForm.get_invariances. Removed warning about bug in |
| 2022 |
linalg.qr in numpy 1.0.1. |
| 2023 |
2009-09-14: Mainly added basic svm tests and did some cleanup and bug fixes. |
| 2024 |
Also, the requested comments have been investigated. |
| 2025 |
2009-09-10: fixed bug in QuadraticForm.get_invariances: second derivative |
| 2026 |
value was wrong when linear term significantly large; second |
| 2027 |
derivative directions were ok |
| 2028 |
2009-08-30: small cleanups of svm_nodes and comments/requests for comments. |
| 2029 |
rike? |
| 2030 |
2009-08-30: changed import order in nodes/__init__.py to clean up import in |
| 2031 |
expansion_nodes.py for GrowingNeuralGasExpansionNode. |
| 2032 |
2009-08-23: Made extension method attributes public. Improved the extension |
| 2033 |
unittests. |
| 2034 |
2009-08-22: Small code improvement in binet inspector. Updated todo for |
| 2035 |
extensions. |
| 2036 |
2009-08-22: Modified the NodeMetaclass wrapping procedure to use super instead |
| 2037 |
of direct method references. Hopefully this has no negative side |
| 2038 |
effects. The advantage is that it is compatible with runtime |
| 2039 |
monkey patching, so it enables more extension stuff. |
| 2040 |
2009-08-21: Extension decorator now preserves function docstring and |
| 2041 |
signature. |
| 2042 |
2009-08-17: Fixed extension related bug in ParallelBiFlow. Improved the |
| 2043 |
related unittests. Added some comments. |
| 2044 |
2009-08-16: Added little comment. |
| 2045 |
2009-08-16: Removed use of parallel FlowNodes in parallel flows. Small update |
| 2046 |
to extensions. |
| 2047 |
2009-08-12: Added option to extension decorator to specify the method name. |
| 2048 |
Added extension decorator unittest (now we have 400 unittests, |
| 2049 |
wheee!). |
| 2050 |
2009-08-12: Fixed left-over bugs from yesterdays mess, fixed unittests. |
| 2051 |
2009-08-11: Added more unittests for the extension mechanism. Changes in the |
| 2052 |
extension documentation and comments, and one small fix. |
| 2053 |
2009-08-11: Use set for list of active extensions. |
| 2054 |
2009-08-11: Added safety mechanisms to extensions, added one corresponding |
| 2055 |
unittest. |
| 2056 |
2009-08-10: Re-enabled Shogun nodes. |
| 2057 |
2009-08-10: Added Henning Sprekeler to the list of developers and removed him |
| 2058 |
from the list of contributors. Updated the various copyright |
| 2059 |
statements all over the place, and the home page. |
| 2060 |
2009-08-10: Corrected the docstring of the GrowingNeuralGasExpansionNode to be |
| 2061 |
more precise on the relation between max_nodes and the dimension |
| 2062 |
of the expansion. |
| 2063 |
2009-08-10: added standard tests for GworingNeuralGasExpansionNode and cleaned |
| 2064 |
up its setting of output_dim. |
| 2065 |
2009-08-07: Bug removed in GrowingNeuralGasExpansionNode: scipy -> numx |
| 2066 |
2009-08-07: missing file added; UpDownNode modified so it can pass down the |
| 2067 |
results of computations at he top |
| 2068 |
2009-08-07: Added GrowingNeuralGasExpansionNode to expansion_nodes. Tests are |
| 2069 |
still missing. |
| 2070 |
2009-08-07: ShogunSVMNode only gets imported when libraries are present. |
| 2071 |
Stylistic changes (not all done yet). Some parameter issues and |
| 2072 |
docstring clarified. Non-kernel classification. |
| 2073 |
self._classification_type added. Parameter mapping for kernels |
| 2074 |
introduced. |
| 2075 |
2009-08-06: remove import of svm_nodes in contrib. the automatic shogun import |
| 2076 |
is not ready for prime-time. |
| 2077 |
2009-08-02: Initial SVM commit. Using only shogun for now. Still needs a lot |
| 2078 |
of work and a better API. But it is already working somehow. |
| 2079 |
2009-08-02: uploaded UpDownBiNode, mother class for DBN, backpropagation; |
| 2080 |
corrected small bug in binet; I've got loads of questions about |
| 2081 |
binet |
| 2082 |
2009-07-31: GaussianClassifierNode has now explicit output_dim method (which |
| 2083 |
fails). |
| 2084 |
2009-07-25: Added RBF node, utilities to replicate array on an additional |
| 2085 |
dimension (rrep, lrep, irep) -> very useful |
| 2086 |
2009-07-19: Updated BiNet todo list. |
| 2087 |
2009-07-16: Fixed bug in rectangular switchboard that was introduced in the |
| 2088 |
previous update. |
| 2089 |
2009-07-15: Added unused channel information to rectangular switchboard, this |
| 2090 |
is also shown in the HTML representation. Added comments in |
| 2091 |
slideshow. |
| 2092 |
|
| 2093 |
------------------------------------------------------------------------------- |
| 2094 |
MDP-2.5: |
| 2095 |
|
| 2096 |
2009-06-30: Added online detection of numerical backend, parallel |
| 2097 |
python support, symeig backend and numerical backend to the output of |
| 2098 |
unit tests. Should help in debugging. |
| 2099 |
2009-06-12: Integration of the cutoff and histogram nodes. |
| 2100 |
2009-06-12: Fixed bug in parallel flow (exception handling). |
| 2101 |
2009-06-09: Fixed bug in LLENode when output_dim is a float. Thanks to Konrad |
| 2102 |
Hinsen. |
| 2103 |
2009-06-05: Fixed bugs in parallel flow for multiple schedulers. |
| 2104 |
2009-06-05: Fixed a bug in layer inverse, thanks to Alberto Escalante. |
| 2105 |
2009-04-29: Added a LinearRegressionNode. |
| 2106 |
2009-03-31: PCANode does not complain anymore when covariance matrix has negative |
| 2107 |
eigenvalues iff svd==True or reduce==True. If output_dim has been |
| 2108 |
specified has a desired variance, negative eigenvalues are ignored. |
| 2109 |
Improved error message for SFANode in case of negative eigenvalues, |
| 2110 |
we now suggest to prepend the node with a PCANode(svd=True) or |
| 2111 |
PCANode(reduce=True). |
| 2112 |
2009-03-26: Migrated from old thread package to the new threading one. |
| 2113 |
Added flag to disable caching in process scheduler. There are some |
| 2114 |
breaking changes for custom schedulers (parallel flow training or |
| 2115 |
execution is not affected). |
| 2116 |
2009-03-25: Added svn revision tracking support. |
| 2117 |
2009-03-25: Removed the copy_callable flag for scheduler, this is now completely |
| 2118 |
replaced by forking the TaskCallable. This has no effect for the |
| 2119 |
convenient ParallelFlow interface, but custom schedulers get broken. |
| 2120 |
2009-03-22: Implemented caching in the ProcessScheduler. |
| 2121 |
2009-02-22: make_parallel now works completely in-place to save memory. |
| 2122 |
2009-02-12: Added container methods to FlowNode. |
| 2123 |
2009-03-03: Added CrossCovarianceMatrix with tests. |
| 2124 |
2009-02-03: Added IdentityNode. |
| 2125 |
2009-01-30: Added a helper function in hinet to directly display a flow HTML |
| 2126 |
representation. |
| 2127 |
2009-01-22: Allow output_dim in Layer to be set lazily. |
| 2128 |
2008-12-23: Added total_variance to the nipals node. |
| 2129 |
2008-12-23: Always set explained_variance and total_variance after training in |
| 2130 |
PCANode. |
| 2131 |
2008-12-12: Modified symrand to really return symmetric matrices (and not only |
| 2132 |
positive definite). Adapted GaussianClassifierNode to account for |
| 2133 |
that. Adapted symrand to return also complex hermitian matrices. |
| 2134 |
2008-12-11: Fixed one problem in PCANode (when output_dim was set to input_dim |
| 2135 |
the total variance was treated as unknown). |
| 2136 |
Fixed var_part parameter in ParallelPCANode. |
| 2137 |
2008-12-11: Added var_part feature to PCANode (filter according to variance |
| 2138 |
relative to absoute variance). |
| 2139 |
2008-12-04: Fixed missing axis arg in amax call in tutorial. |
| 2140 |
Thanks to Samuel John! |
| 2141 |
2008-12-04: Fixed the empty data iterator handling in ParallelFlow. |
| 2142 |
Also added empty iterator checks in the normal Flow (raise an |
| 2143 |
exception if the iterator is empty). |
| 2144 |
2008-11-19: Modified pca and sfa nodes to check for negative eigenvalues |
| 2145 |
in the cov matrices |
| 2146 |
2008-11-19: symeig integrated in scipy, mdp can use it from there now. |
| 2147 |
2008-11-18: Added ParallelFDANode. |
| 2148 |
2008-11-18: Updated the train callable for ParallelFlow to support additional |
| 2149 |
arguments. |
| 2150 |
2008-11-05: Rewrite of the make parallel code, now supports hinet structures. |
| 2151 |
2008-11-03: Rewrite of the hinet HTML repesentation creator. Unfortunately this |
| 2152 |
also breaks the public interface, but the changes are pretty simple. |
| 2153 |
2008-10-29: Shut off warnings coming from remote processes in ProcessScheduler |
| 2154 |
2008-10-27: Fixed problem with overwriting kwargs in the init method |
| 2155 |
of ParallelFlow. |
| 2156 |
2008-10-24: Fixed pretrained nodes bug in hinet.FlowNode. |
| 2157 |
2008-10-20: Fixed critical import bug in parallel package |
| 2158 |
when pp (parallel python library) is installed. |
| 2159 |
|
| 2160 |
------------------------------------------------------------------------------- |
| 2161 |
MDP-2.4: |
| 2162 |
|
| 2163 |
2008-10-16: added interface to BLAS's "gemm" matrix multiplication |
| 2164 |
function. |
| 2165 |
2008-10-15: removed obsolete helper functions. |
| 2166 |
2008-10-15: added new feature. Now: |
| 2167 |
output = XXXNode(output_dim=10)(x) |
| 2168 |
trains and executes the node. This makes helper_functions obsolete! |
| 2169 |
It even works for multiple training phases (only if the training |
| 2170 |
phase has not started yet). |
| 2171 |
2008-10-15: removed use of deprecated features with python2.6 -3. |
| 2172 |
2008-10-14: removed dangerous list default argument in rotate and permute |
| 2173 |
functions. A tuple is now used. |
| 2174 |
2008-10-13: PEP8 code restyling (pylint). |
| 2175 |
2008-10-07: Removed workarounds for pickle bug in numpy < 1.1.x (see |
| 2176 |
numpy ticket 551). |
| 2177 |
2008-09-24: Implemeted metaclass trick for automatic inheritance of |
| 2178 |
documentation for private methods. Node's subclass authors |
| 2179 |
can now directly document "_execute", and let users see those |
| 2180 |
docs as documenting "execute": |
| 2181 |
|
| 2182 |
class MyNode(Node): |
| 2183 |
def _execute(self): |
| 2184 |
"""blah""" |
| 2185 |
|
| 2186 |
>>> print MyNode.execute.__doc__ |
| 2187 |
blah. |
| 2188 |
|
| 2189 |
Just defined set of methods allow overwriting of the docstring. |
| 2190 |
the current list is: |
| 2191 |
['_train', '_stop_training', '_execute', '_inverse'] |
| 2192 |
2008-09-22: Added new functionality to nodes and flows. |
| 2193 |
Node1+Node2 now returns a flow and Flow1+Node1 appends |
| 2194 |
Node1 to Flow1. |
| 2195 |
2008-09-07: New node for Locally Linear Embedding |
| 2196 |
2008-08-28: The docstring of Flow.train now mentions that instead of x the |
| 2197 |
iterators can also return a tuple of x and additional args. |
| 2198 |
2008-08-28: Fixed bug in PCANode: |
| 2199 |
when setting output_dim to a float number after instantiation |
| 2200 |
but before stop_training, i.e: |
| 2201 |
pca = PCANode() |
| 2202 |
pca.train(x) |
| 2203 |
pca.output_dim = 0.9 |
| 2204 |
pca.stop_training() |
| 2205 |
an exception was thrown: Output dim are set already (0) (1 given) |
| 2206 |
2008-08-28: Fixed bug in PCANode: |
| 2207 |
when setting output_dim to float number, after stop_training |
| 2208 |
pca.output_dim was a float number instead of integer. |
| 2209 |
2008-08-21: Added inverse for Switchboard, including unittests. |
| 2210 |
2008-08-19: Added parallel package. |
| 2211 |
2008-08-19: Added new NormalNoiseNode (with small unittest), |
| 2212 |
which can be safely pickled. |
| 2213 |
2008-08-07: fixed bug in rbm_nodes (TrainingFinishedException). |
| 2214 |
Thanks to Mathias Franzius! |
| 2215 |
2008-07-03: Node.__call__ method now takes *args and **kwargs to call |
| 2216 |
the execute method. Thanks to Jake VanderPlas! |
| 2217 |
2008-06-30: Fix debian bug 487939: |
| 2218 |
Modifed the fix for the debian bug so that white_parm defaults to |
| 2219 |
None instead of {}. With default {} a sublcass updating the |
| 2220 |
dictionary would have updated it for all instances... |
| 2221 |
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=487939 |
| 2222 |
Applied patch from Michael Hanke. Thank you! |
| 2223 |
|
| 2224 |
------------------------------------------------------------------------------- |
| 2225 |
MDP-2.3: |
| 2226 |
2008-05-08: fixed bug in the raising of the switchboard exception. |
| 2227 |
2008-05-02: added exception for hinet.Rectangular2dSwitchboard |
| 2228 |
and corresponding unittest. |
| 2229 |
2008-04-25: added new TDSEPNode, updated ISFANode. |
| 2230 |
2008-04-16: added an HTML visualisation tool for hinet flows to the hinet |
| 2231 |
package. Added a hinet section in the tutorial. |
| 2232 |
2008-03-22: released MDP 2.2, this release is intended for internal use only |
| 2233 |
2008-03-21: added hinet package |
| 2234 |
2008-03-19: added RBM nodes |
| 2235 |
2008-03-18: fixed bug in PCANode when output_dim was specified as float. |
| 2236 |
2008-03-16: HitParadeNode now supports integer dtypes. |
| 2237 |
2008-03-13: created contrib subpackage and test_contrib test suites. |
| 2238 |
JADENode moved to contrib. Tests are now sorted by test_suite. |
| 2239 |
Added the NIPALSNode for iterative PCA. |
| 2240 |
2008-03-11: added JADENode with tests |
| 2241 |
2008-03-10: added test for dimensions settings (nodes should have |
| 2242 |
output_dim and input_dim set after training) |
| 2243 |
2008-01-22: removed utils.comb scipy code and obsolete SciPy copyright |
| 2244 |
notice from utils/routines.py |
| 2245 |
2008-01-03: fixed bug in SFA2Node when input_dim was set on instantiation, |
| 2246 |
fixed bug in SFA2Node._stop_training when output_dim was set and |
| 2247 |
larger than input_dim, |
| 2248 |
fixed bug in QuadraticForm when H was indefinite. |
| 2249 |
2007-09-14: FastICANode has been completely rewritten. It is now up-to-date |
| 2250 |
with the original Matlab version (2.5) published on 19.10.2005. |
| 2251 |
Highlights: |
| 2252 |
- fine tuning is implemented |
| 2253 |
- the stabilized version of the algorithm is implemented |
| 2254 |
- the new 'skew' non linearity for skewed input data is implemented |
| 2255 |
- bug fix in 'tanh' non linearity |
| 2256 |
- bug fix in 'gaus' non linearity in 'symm' approach |
| 2257 |
- all combinations of input parameters are being tested |
| 2258 |
2007-09-12: Added new funcionality for PCA and Whitening. The nodes can now |
| 2259 |
be instantiated with svd=True, to use SVD instead of the standard |
| 2260 |
eigenvalue problem solver. Setting reduce=True and var_abs and |
| 2261 |
var_rel it is possible to automatically discard irrelevant |
| 2262 |
principal components. See the nodes documentation for more info |
| 2263 |
2007-09-11: added check for positive eigenvalues in symeg and symeig_fake |
| 2264 |
2007-08-17: added pre_inversion_checks, inversion equivalent to |
| 2265 |
pre_execution_checks |
| 2266 |
2007-07-20: enable iteration on items and time completion estimation method in |
| 2267 |
progress_bar |
| 2268 |
2007-07-09: fixed bug in SFA2Node: when instantiated with |
| 2269 |
input_dim==output_dim, the range in symeig was not set and |
| 2270 |
output_dim was wrong |
| 2271 |
2007-07-04: - fixed bug in mdp.utils.__init__.py about SymeigException not |
| 2272 |
caught |
| 2273 |
- fixed bug in ISFANode when perturbing eps_contrast was not |
| 2274 |
respected (it was perturing indefinitely) |
| 2275 |
- added a 'debug' kwarg to stop_training in PCANode, WhiteningNode, |
| 2276 |
and SFANode. When stop_training fails because of singular |
| 2277 |
matrices, the matrices itselves are not deleted (as before) but |
| 2278 |
kept in self.cov_mtx and self.dcov_mtx for later inspection |
| 2279 |
2007-07-03: fixed bug in ISFANode when lags was array. |
| 2280 |
2007-06-25: Added get_{proj,rec}matrix method to the ICANodes. |
| 2281 |
'get_recmatrix' returns an estimate of the mixing matrix. |
| 2282 |
|
| 2283 |
------------------------------------------------------------------------------- |
| 2284 |
MDP-2.1: |
| 2285 |
|
| 2286 |
2007-03-23: Updated tutorial and web site. |
| 2287 |
2007-03-23: Use get_limits.finfo to set precision limits |
| 2288 |
2007-03-23: Use symeig if found |
| 2289 |
2007-03-23: Use scipy instead of numpy when possible |
| 2290 |
2007-03-23: Numpy functions substituted with array method where possible |
| 2291 |
2007-03-22: Implemented invariances in Quadraticforms |
| 2292 |
2007-03-21: MDP now fully compatible with numpy 1.0 |
| 2293 |
2007-03-15: Added ISFANode |
| 2294 |
2007-03-13: Added MultipleCovarianceMatrices utility class |
| 2295 |
2006-11-04: Added 'save' method to Node and Flow |
| 2296 |
2006-11-04: Node's train and stop_training methods now accept *args and **kwargs |
| 2297 |
|
| 2298 |
------------------------------------------------------------------------------- |
| 2299 |
MDP-2.0RC: |
| 2300 |
|
| 2301 |
29.06.2006: Updated tutorial and web site. |
| 2302 |
28.06.2006: New random_rot. Added SFA2Node, QuadraticForm, moved graph into the mdp tree. |
| 2303 |
27.06.2006: Converted typecode handling to numpy.dtype style; supported_typecodes |
| 2304 |
is now a property. Added bias in covariance. |
| 2305 |
26.06.2006: Converted to numpy. Scipy, Numeric, and numarray are not supported |
| 2306 |
anymore |
| 2307 |
06.12.2005: pca_nodes: added get_explained_variance public method. |
| 2308 |
02.12.2005: New introspection utils in mdp.utils. |
| 2309 |
01.11.2005: New nodes: FANode, FDANode, GaussianClassifierNode |
| 2310 |
31.10.2005: Non back-compatible changes. Node has got three new 'properties': |
| 2311 |
output_dim, input_dim, typecode. They are accessible through their |
| 2312 |
getters (get_PROPERTY). They can be set using the default setters |
| 2313 |
(set_PROPERTY). Subclasses can customize the setters, overriding |
| 2314 |
the _set_PROEPRTY private methods. All nodes had to be changed to |
| 2315 |
conform with the new structure. All tests pass. |
| 2316 |
26.10.2005: To force MDP to use a particular numerical extension, you |
| 2317 |
can now set the enviroment variable MDPNUMX. Supported |
| 2318 |
values are 'symeig', 'scipy', 'Numeric', 'numarray'. |
| 2319 |
Mainly useful for testing purposes. |
| 2320 |
06.10.2005: The SfaNode, CuBICA, and FastICA aliases have been deleted. |
| 2321 |
06.10.2005: Node supports multiple and even an infinite number of training |
| 2322 |
phases. FiniteNode makes the implementation of a class with |
| 2323 |
a finite number of phases easy and one with just one phase |
| 2324 |
trivial. |
| 2325 |
06.10.2005: Flow supports the new nodes and even nodes requiring a |
| 2326 |
supervision signal during training. |
| 2327 |
06.10.2005: SignalNode, SignalNodeException, and SimpleFlow are now |
| 2328 |
deprecated aliases. Use Node, NodeException and Flow instead. |
| 2329 |
06.10.2005: some bug fixes. |
| 2330 |
05.10.2005: fixed failing 'import mdp' in test_symeig when mdp is not installed |
| 2331 |
12.07.2005: bug in utils.symrand (global name "mult" is not defined) |
| 2332 |
06.07.2005: changed round off error checking in FastICANode (it was 1e-15 and |
| 2333 |
it is now 1e-5). it failed on a MacOSX machine. |
| 2334 |
23.06.2005: _check_roundoff in lcov.py issues a MDPWarning using warnings.warn. |
| 2335 |
it was using 'raise' before, and it was caught as a FlowException. |
| 2336 |
21.06.2005: node consistency is assured at flow instantiation. |
| 2337 |
------------------------------------------------------------------------------- |
| 2338 |
MDP-1.1.0: |
| 2339 |
|
| 2340 |
13.06.2005: MDP 1.1.0 released. |
| 2341 |
01.06.2005: Crash_recovery is now off by default. |
| 2342 |
To switch it on do: flow.set_crash_recovery(1) |
| 2343 |
30.05.2005: New NoiseNode. |
| 2344 |
30.05.2005: SimpleFlow and SignalNode now have a 'copy' method. |
| 2345 |
30.05.2005: SimpleFlow is now a mutable sequence type and implements |
| 2346 |
many of the 'list' methods. |
| 2347 |
30.05.2005: removed scipy dependency. Now mdp runs with either Numeric, |
| 2348 |
numarray or scipy. |
| 2349 |
24.05.2005: symeig removed from mdp. |
| 2350 |
23.05.2005: all classes are now new-style. |
| 2351 |
------------------------------------------------------------------------------- |
| 2352 |
MDP-1.0.0: |
| 2353 |
|
| 2354 |
15.11.2004: MDP 1.0.0 released |
| 2355 |
09.11.2004: Added crash recovery capabilities to SimpleFlow (on by default). |
| 2356 |
05.11.2004: New GrowingNeuralGasNode. New graph module. |
| 2357 |
04.11.2004: New IdentityNode subclass added. All analysis nodes are |
| 2358 |
now subclasses of IdentityNode. |
| 2359 |
Use of input_dim and output_dim in the nodes' constructor |
| 2360 |
is now consistent. |
| 2361 |
02.11.2004: Now symeig works reliably also for complex matrices. |
| 2362 |
symeig now can be distributed as an independent package |
| 2363 |
It is still contained in mdp.utils for convenience. |
| 2364 |
Default value for option overwrite changed from 1 to 0. |
| 2365 |
14.10.2004: With the release of Scypy 0.3.2, the installation of MDP got much |
| 2366 |
simpler. |
| 2367 |
06.10.2004: Fixed a bug in symeig (when B=None and overwrite=0, |
| 2368 |
symeig raised an exception) |
| 2369 |
04.09.2004: Fixed Windows-xspecific problems. |
| 2370 |
------------------------------------------------------------------------------- |
| 2371 |
MDP-0.9.0: |
| 2372 |
|
| 2373 |
24.08.2004: MDP 0.9.0 released. First public release. |
| 2374 |
16.08.2004: MDP project registered at SourceForge. |
| 2375 |
------------------------------------------------------------------------------- |