nam issueshttps://gitlab.gwdg.de/groups/nam/-/issues2020-05-05T07:58:05Zhttps://gitlab.gwdg.de/nam/ProxPython/-/issues/2sqrt in evaluateChange - gretchko2020-05-05T07:58:05ZGijsbert Simon Matthijs Jansensqrt in evaluateChange - gretchkoCompared to the old SimpleAlgorithm, it appears that we are missing a square root of the change.
We should verify this.Compared to the old SimpleAlgorithm, it appears that we are missing a square root of the change.
We should verify this.https://gitlab.gwdg.de/nam/ProxPython/-/issues/7Ptychography not working2020-06-17T13:58:41ZRussell LukePtychography not working[nix-shell:~/work/git/ProxPython/demos]$ python3 Ptychography_PHeBIE_Wilke2.py
Traceback (most recent call last):
File "Ptychography_PHeBIE_Wilke2.py", line 21, in <module>
Pty = PtychographyExperiment(debug = False, warmup = True...[nix-shell:~/work/git/ProxPython/demos]$ python3 Ptychography_PHeBIE_Wilke2.py
Traceback (most recent call last):
File "Ptychography_PHeBIE_Wilke2.py", line 21, in <module>
Pty = PtychographyExperiment(debug = False, warmup = True, MAXIT=10, warmup_params=warmupParams)
File "/home/rluke1/work/git/ProxPython/demos/../proxtoolbox/experiments/experiment.py", line 39, in __call__
obj.initialize() # initialize experiment object
File "/home/rluke1/work/git/ProxPython/demos/../proxtoolbox/experiments/ptychography/ptychographyExperiment.py", line 639, in initialize
super(PtychographyExperiment, self).initialize() # call parent's method
File "/home/rluke1/work/git/ProxPython/demos/../proxtoolbox/experiments/experiment.py", line 218, in initialize
self.loadData()
File "/home/rluke1/work/git/ProxPython/demos/../proxtoolbox/experiments/ptychography/ptychographyExperiment.py", line 185, in loadData
P = mat_dict['P']
KeyError: 'P'
[nix-shell:~/work/git/ProxPython/demos]$Release in support of books.gretchkos.gretchkohttps://gitlab.gwdg.de/nam/ProxPython/-/issues/6IterateMonitor (gretchko)2020-06-17T14:03:39ZGijsbert Simon Matthijs JansenIterateMonitor (gretchko)The iterate monitor has a property u_monitor, which is set to the initial (random) guess and is not updated anywhere during operation.The iterate monitor has a property u_monitor, which is set to the initial (random) guess and is not updated anywhere during operation.https://gitlab.gwdg.de/nam/ProxPython/-/issues/4Capitalization - Gretchko vs other branches2020-06-17T15:42:45ZGijsbert Simon Matthijs JansenCapitalization - Gretchko vs other branchesI noticed that the capitalization of the folders proxtoolbox/Algorithms and proxtoolbox/ProxOperators was changed to lower case. In principle, this is good, but it causes issues when switching between branches (on windows at least), wher...I noticed that the capitalization of the folders proxtoolbox/Algorithms and proxtoolbox/ProxOperators was changed to lower case. In principle, this is good, but it causes issues when switching between branches (on windows at least), where I have to manually change the capitalization of these folders. We should aim to stick with one naming across branches, and be consistenthttps://gitlab.gwdg.de/nam/ProxPython/-/issues/3Gretchko - Proxoperator arguments2020-08-03T12:01:37ZGijsbert Simon Matthijs JansenGretchko - Proxoperator argumentsWhy is it necessary to include a prox_index argument for all proxoperators if this argument is in most cases not used? Can't we allow for keyword arguments using **kwargs?Why is it necessary to include a prox_index argument for all proxoperators if this argument is in most cases not used? Can't we allow for keyword arguments using **kwargs?https://gitlab.gwdg.de/nam/ProxPython/-/issues/5easy data loading (gretchko)2020-08-03T12:02:04ZGijsbert Simon Matthijs Janseneasy data loading (gretchko)There should be a method to select which datafile is loaded as a user. (paths cannot be hardcoded)
A path might be passed as an argument to the ExperimentThere should be a method to select which datafile is loaded as a user. (paths cannot be hardcoded)
A path might be passed as an argument to the Experimenthttps://gitlab.gwdg.de/nam/ProxPython/-/issues/9Noise not working in JWST experiment2021-01-31T16:17:48ZRussell LukeNoise not working in JWST experimentI created a demo to run CDRl on JWST with noise. I got the following error:
[nix-shell:~/work/git/ProxPython/demos]$ python3 JWST_CDRl.py
Loading data.
Traceback (most recent call last):
File "JWST_CDRl.py", line 6, in <module>
...I created a demo to run CDRl on JWST with noise. I got the following error:
[nix-shell:~/work/git/ProxPython/demos]$ python3 JWST_CDRl.py
Loading data.
Traceback (most recent call last):
File "JWST_CDRl.py", line 6, in <module>
JWST = JWST_Experiment(algorithm = 'CDRl', lambda_0 = 0.5, lambda_max = 0.5, noise = 'Poisson', MAXIT = 1000)
File "/home/rluke1/work/git/ProxPython/demos/../proxtoolbox/experiments/experiment.py", line 39, in __call__
obj.initialize() # initialize experiment object
File "/home/rluke1/work/git/ProxPython/demos/../proxtoolbox/experiments/experiment.py", line 223, in initialize
self.loadData()
File "/home/rluke1/work/git/ProxPython/demos/../proxtoolbox/experiments/phase/JWST_Experiment.py", line 150, in
loadData
k_j[ii,jj]= np.random.poisson(k_j[ii,jj])*snr #use built in numpy possion instead of
utils.PoissonRan(k[ii,jj,i])*snr
File "mtrand.pyx", line 3556, in numpy.random.mtrand.RandomState.poisson
File "_common.pyx", line 815, in numpy.random._common.disc
File "_common.pyx", line 379, in numpy.random._common.check_constraint
ValueError: lam value too largehttps://gitlab.gwdg.de/nam/ProxPython/-/issues/8divide by zero2021-01-31T16:18:19ZRussell Lukedivide by zeroIn the warmup algorithm for the Ptychography experiment, you get a divide by zero error:
Running warmup algorithm DRl
/home/rluke1/work/git/ProxPython/demos/../proxtoolbox/proxoperators/fourier.py:69: RuntimeWarning: divide by zero enco...In the warmup algorithm for the Ptychography experiment, you get a divide by zero error:
Running warmup algorithm DRl
/home/rluke1/work/git/ProxPython/demos/../proxtoolbox/proxoperators/fourier.py:69: RuntimeWarning: divide by zero encountered in true_divide
tmp = u_hat_sq / self.data_sq[j]
This does not seem to kill the run, but probably should be dealt with.https://gitlab.gwdg.de/nam/ProxPython/-/issues/12Elser TO DO2022-10-27T13:14:29ZRussell LukeElser TO DOElser TO DO:
* fix file I/O (Daniel)
* get the Elser data in the form used for phase problems in ProxToolbox (Jazib)
* benchmarkElser TO DO:
* fix file I/O (Daniel)
* get the Elser data in the form used for phase problems in ProxToolbox (Jazib)
* benchmarkRussell LukeMuhammad Jazib ZafarDaniel SchellhornRussell Lukehttps://gitlab.gwdg.de/nam/ProxPython/-/issues/10How to: install a new algorithm2023-05-05T07:29:31ZRussell LukeHow to: install a new algorithmI have just tried - and given up - installing a new algorithm. There seem to be lots of submodules like __init__.py that need to be modified. Of course, some algorithms will be more difficult than others, but it would be nice to see a...I have just tried - and given up - installing a new algorithm. There seem to be lots of submodules like __init__.py that need to be modified. Of course, some algorithms will be more difficult than others, but it would be nice to see a step-by-step how-to for a ``new algorithm" which takes exactly the same format as something already there.https://gitlab.gwdg.de/nam/ProxPython/-/issues/13Refine automatic data downloader2023-05-05T07:31:19ZDaniel SchellhornRefine automatic data downloaderWith the deletions of only the CDP data, the automatic downloader fails to recognize that the data is missing, since the other phase data is still there. I will update this automation to a more refined one soon, also for all the other da...With the deletions of only the CDP data, the automatic downloader fails to recognize that the data is missing, since the other phase data is still there. I will update this automation to a more refined one soon, also for all the other data types.Daniel SchellhornDaniel Schellhornhttps://gitlab.gwdg.de/nam/ProxPython/-/issues/14not importing h5py or mat732023-05-05T07:32:33ZRussell Lukenot importing h5py or mat73I have updated the pyproject.toml file to load mat73, and then open my editor in the poetry shell, but when I import this into Phasepack_Experiment.py, it seems to think that this package has not been installed. What am I doing wrong?I have updated the pyproject.toml file to load mat73, and then open my editor in the poetry shell, but when I import this into Phasepack_Experiment.py, it seems to think that this package has not been installed. What am I doing wrong?