NIFTy merge requestshttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests2017-05-01T00:35:19Zhttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/74Rename property "symmetric" to "self_adjoint"2017-05-01T00:35:19ZMartin ReineckeRename property "symmetric" to "self_adjoint"https://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/378Renaming2019-11-27T20:07:08ZMartin ReineckeRenaminghttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/502Renamings2020-05-25T13:35:47ZPhilipp Arrasparras@mpa-garching.mpg.deRenamingshttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/881re.optimize_kl: fix typos in docstring2023-11-14T11:18:41ZJakob Rothre.optimize_kl: fix typos in docstring@pfrank fixes two typos in the docstring of re.optimize_kl@pfrank fixes two typos in the docstring of re.optimize_klhttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/147re-organize PowerSpace2017-07-12T08:02:57ZMartin Reineckere-organize PowerSpaceThis branch aims at making the current `PowerSpace` design simpler.
The following changes have been made so far:
- the `PowerSpace` members are always fully computed upon construction of the `PowerSpace`; caching has been removed. Th...This branch aims at making the current `PowerSpace` design simpler.
The following changes have been made so far:
- the `PowerSpace` members are always fully computed upon construction of the `PowerSpace`; caching has been removed. This still has performance drawbacks which I hope to address in the future.
- the parameters `logarithmic` and `nbin` can still be supplied to the constructor, but they are no longer part of the `PowerSpace`'s state. They are used to compute the `binbounds`, and forgotten afterwards. (In the current implementation it is possible to set these attributes to values which contradict the actual `binbounds`, which is inconsistent.)
- `pundex` is currently not supported, but can be reintroduced if needed.
- "natural" binning (which happens if neither `logarithmic`, `nbin` or `binbounds` are specified to the constructor) has been improved to avoid spurious extra bins due to numerical noise.
- index computation has been accelerated significantly
- the `config` property has been removed; `binbounds` is now a direct property of the class.
I'm sure that a lot will have to change before this can be merged; please consider this merge request mainly as a starting point for discussion!
As a quick demonstration of the advantages of the new implementation, run the following commands in both "master" and "index_games" branch and watch CPU time and memory consumption:
```
a=RGSpace((10000,10000),harmonic=True)
p=PowerSpace(a,binbounds=sqrt(2.)*np.arange(10000.))
```https://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/17Replace all occurances of np.prod shape to it's faster alternative.2016-10-25T02:13:07ZGhost UserReplace all occurances of np.prod shape to it's faster alternative.Theo SteiningerTheo Steiningerhttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/96Replace direct smoothing code by pure Python2017-05-15T15:11:03ZMartin ReineckeReplace direct smoothing code by pure PythonThis gets rid of smooth_util.pyx and replaces it by a pure Python implementation that seems to be slightly faster for the tests I have done so far.
Since smooth_util.pyx was the last Cython code in Nifty, I have also removed all menti...This gets rid of smooth_util.pyx and replaces it by a pure Python implementation that seems to be slightly faster for the tests I have done so far.
Since smooth_util.pyx was the last Cython code in Nifty, I have also removed all mentions of Cython from documentation, requirements files etc.https://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/237Replace InverseOperator and AdjointOperator with OperatorAdapter, and more2018-04-03T13:50:37ZMartin ReineckeReplace InverseOperator and AdjointOperator with OperatorAdapter, and moreThis patch introduces the `OperatorAdapter` class, which does the work so far done by `InverseOperator` and `AdjointOperator`, but it is smaller and more flexible.
It also removes the method `inverse_draw_sample()` and adds a `fromInver...This patch introduces the `OperatorAdapter` class, which does the work so far done by `InverseOperator` and `AdjointOperator`, but it is smaller and more flexible.
It also removes the method `inverse_draw_sample()` and adds a `fromInverse` flag to `draw_sample()` instead, which also makes many things easier to write.
As a completely unrelated bonus, there is an interface to SciPy's conjugate gradient solver, which can now be used as an alternative to Nifty's `ConjugateGradient`.https://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/276replace makro artifact in references2018-10-17T11:52:49ZChristoph Lienhardreplace makro artifact in referencesThere was a left-over user-defined makro in the references section on the NIFTy homepage.
This prevented copy-pasting the refs into my bib file because obviously this makro was not defined in my tex project.
It should be fixed now.
(I re...There was a left-over user-defined makro in the references section on the NIFTy homepage.
This prevented copy-pasting the refs into my bib file because obviously this makro was not defined in my tex project.
It should be fixed now.
(I replaced the relevant line with one from a bibtex entry generated by scholar.google.com)https://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/149replace ** with clearer operators where possible2017-06-05T00:49:54ZTheo Steiningerreplace ** with clearer operators where possiblehttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/520__repr__ methods for _FunctionApplier, _Amplitude and _LognormalMomentMatching2020-05-30T15:40:22ZLukas Platz__repr__ methods for _FunctionApplier, _Amplitude and _LognormalMomentMatchingThose operators obfuscated their actions, making it hard to debug overflows. Fixed it.Those operators obfuscated their actions, making it hard to debug overflows. Fixed it.Martin ReineckeMartin Reineckehttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/643Reproduce stdnormal samples2021-06-11T14:27:57ZPhilipp Arrasparras@mpa-garching.mpg.deReproduce stdnormal samplesMartin ReineckeMartin Reineckehttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/719ResidualSampleList: Fail early during loading2021-12-08T10:32:37ZPhilipp Arrasparras@mpa-garching.mpg.deResidualSampleList: Fail early during loading`ResidualSampleList`s save a `*.mean.pickle` file, whereas `SampleList`s do not.`ResidualSampleList`s save a `*.mean.pickle` file, whereas `SampleList`s do not.Martin ReineckeMartin Reineckehttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/906Resolve "Implement Evidence Calculation"2024-03-25T09:30:43ZMatteo GuardianiResolve "Implement Evidence Calculation"Closes #370.Closes #370.Matteo GuardianiMatteo Guardianihttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/474Resolve "Streamline arguments position"2020-05-19T13:04:45ZRouven LemmerzResolve "Streamline arguments position"Closes #300Closes #300https://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/757Restore matplotlib as optional dependency and improve `exec_time`2022-04-01T09:15:59ZPhilipp Arrasparras@mpa-garching.mpg.deRestore matplotlib as optional dependency and improve `exec_time`Martin ReineckeMartin Reineckehttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/752restoring plot of setup in getting_started_32022-03-06T17:33:20ZTorsten Ensslinrestoring plot of setup in getting_started_3Restoring the setup plot in getting_started_3.py so that its output is consistent with the figures shown in my "IFT & AI" paper.Restoring the setup plot in getting_started_3.py so that its output is consistent with the figures shown in my "IFT & AI" paper.Philipp Arrasparras@mpa-garching.mpg.dePhilipp Arrasparras@mpa-garching.mpg.dehttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/761Restructure Docs2022-04-14T16:46:50ZVincent EberleRestructure DocsChanges in Readme
=================
I fixed some broken links, e.g.: the mpi4py link and one link the a code section on git were broken)
Changes in the generated docs
=============================
- how to generate the docs was stated...Changes in Readme
=================
I fixed some broken links, e.g.: the mpi4py link and one link the a code section on git were broken)
Changes in the generated docs
=============================
- how to generate the docs was stated twice (user/installation and dev/index) and once the dependencies were incomplete
- IMHO installation for developers should be part of dev-section (to keep it ease to find the installation guide I put in a link from installation to dec/index
- plotting, mpi4py, etc are also used by non developers, so I moved the section to the user installation
Further suggestions
===================
- [x] moving the section on installation of mpi4py and matplotlib in the README file as well
- [x] installation guide for astropy, h5py (or a link), in order to keep it the way we do it for other opt. dependencies
- [x] putting in or changing the developer section to the way most of us install nifty etc. (cloning repo, pip install --user -e .,) since also new group members use these instructions. (I think this would decrease some overhead)
@parras @mtr what do you think?Martin ReineckeMartin Reineckehttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/542Restructure operator tests2020-06-19T10:44:20ZPhilipp Arrasparras@mpa-garching.mpg.deRestructure operator testsMartin ReineckeMartin Reineckehttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/583Restructure tensors2020-12-27T11:00:23ZPhilipp FrankRestructure tensors