NIFTy merge requestshttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests2017-08-21T06:02:58Zhttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/170More cg work2017-08-21T06:02:58ZTheo SteiningerMore cg workhttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/169Added dtype option to prober.2017-07-18T13:05:35ZTheo SteiningerAdded dtype option to prober.https://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/168WIP: Fixing wiener filter demo2017-08-02T21:13:16ZJakob KnollmuellerWIP: Fixing wiener filter demoI added some distribution_strategies in wiener_filter_via_hamiltonian.py, the WienerFilterEnergy now also uses the given inverter and distribution_strategies are now set in InvertibleOperatorMixin if no x0 is givenI added some distribution_strategies in wiener_filter_via_hamiltonian.py, the WienerFilterEnergy now also uses the given inverter and distribution_strategies are now set in InvertibleOperatorMixin if no x0 is givenhttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/167Revert "added some distribution_strategy's for mpirun, fixed some errors in W...2017-07-17T14:29:00ZJakob KnollmuellerRevert "added some distribution_strategy's for mpirun, fixed some errors in WienerFilterEnergy"This reverts commit 03cc7f7e8edcc381eca505dbfc9c81feac394c1dThis reverts commit 03cc7f7e8edcc381eca505dbfc9c81feac394c1dhttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/166Line search2017-07-26T08:28:50ZTheo SteiningerLine searchhttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/165avoid exec statements; they cause problems with Python 32017-07-14T10:52:51ZMartin Reineckeavoid exec statements; they cause problems with Python 3This small change replaces exec() calls by calls to ast.literal_eval().
I don't know why exactly, but Python 3 appears to forget newly created variables after returning from exec(), so the current approach does not work with Python 3.This small change replaces exec() calls by calls to ast.literal_eval().
I don't know why exactly, but Python 3 appears to forget newly created variables after returning from exec(), so the current approach does not work with Python 3.https://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/164Branch master2017-07-14T00:09:08ZTheo SteiningerBranch masterhttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/163Index games22017-07-13T12:08:14ZMartin ReineckeIndex games2For some strange reason I was not able to merge "master" back onto "index_games" ... it simply overwrote all my changes.
So I have put the necessary changes on the "index_games2" branch.For some strange reason I was not able to merge "master" back onto "index_games" ... it simply overwrote all my changes.
So I have put the necessary changes on the "index_games2" branch.https://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/162fix demos2017-07-12T09:30:13ZMartin Reineckefix demosAfter the merge of "working_on_demos", some demos stopped working ...
This merge request fixes most of them.
wiener_filter_easy.py is probably obsolete (it relies on PropagatorOperator, which was removed). If I get confirmation for thi...After the merge of "working_on_demos", some demos stopped working ...
This merge request fixes most of them.
wiener_filter_easy.py is probably obsolete (it relies on PropagatorOperator, which was removed). If I get confirmation for this, I'll delete the file as well.https://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/161suggested fix for stopping criterion2017-07-11T08:27:33ZTheo Steiningersuggested fix for stopping criterionhttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/160fix field.vdot2017-07-11T08:18:03ZMartin Reineckefix field.vdothttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/159Vl bfgs2017-07-09T23:04:21ZTheo SteiningerVl bfgshttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/158various cleanups2017-07-08T11:45:05ZMartin Reineckevarious cleanupsA few classes have setters that are not really needed. Immutability should be treasured, so I propose to remove them.A few classes have setters that are not really needed. Immutability should be treasured, so I propose to remove them.https://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/157Fix hermitianizer2017-07-08T16:25:52ZTheo SteiningerFix hermitianizerhttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/156Add Python3 compatibility2017-08-22T10:23:29ZMartin ReineckeAdd Python3 compatibilityThis branch adds Python3 compatibility to NIFTy, while also staying compatible to Python2.
At the current state, the code compiles in both language versions and passes almost all tests.
Exception: some HDF5 tests, (1 fail with Pyth...This branch adds Python3 compatibility to NIFTy, while also staying compatible to Python2.
At the current state, the code compiles in both language versions and passes almost all tests.
Exception: some HDF5 tests, (1 fail with Python2, 5 fails with Python3). I need help with debugging those. **[now fixed]**
NOTE: An update to the "mpi_dummy" package is necessary to make it Python3 compatible as well. A merge request has been opened.
Problems encountered:
- the "hack" in the __new__ method of SmoothingOperator, which returns a DirectSmoothingOperator or an FFTSmoothingOperator depending on the arguments, stopped working with Python3. I have put together a crude workaround, but this could still be improved.
Things to do:
- enhance continuous integration to test Nifty with Python3 **[done now]**https://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/155Tweak CG2017-07-02T11:48:33ZMartin ReineckeTweak CGThe conjugate gradient solver had a few problems:
- if an initial guess `x0` was given to the solver, this guess was silently changed during the solver run.
- some quantities like alpha, beta, gamma and delta must be real-valued by cons...The conjugate gradient solver had a few problems:
- if an initial guess `x0` was given to the solver, this guess was silently changed during the solver run.
- some quantities like alpha, beta, gamma and delta must be real-valued by construction, but were computed as complex numbers. This could have unwanted effects due to accumulation of numerical noise, so I cast the values to real type manually.https://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/154Fix for RG smoothing width?2017-06-23T23:55:25ZTheo SteiningerFix for RG smoothing width?https://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/153Field.dot->Field.vdot2017-06-20T18:30:33ZMartin ReineckeField.dot->Field.vdotIssue #151: rename Field.dot() to Field.vdot().Issue #151: rename Field.dot() to Field.vdot().https://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/152Field testing2018-01-25T12:48:03ZTheo SteiningerField testinghttps://gitlab.mpcdf.mpg.de/ift/nifty/-/merge_requests/151added square in create_power_operator2017-06-06T11:59:06ZJakob Knollmuelleradded square in create_power_operatorpower_synthesize takes square root, diagonal of power operator has to be this squaredpower_synthesize takes square root, diagonal of power operator has to be this squared