Open dimkab opened 4 years ago
I believe this issue is due to the fact that most of the IR.m files in the IRtools folder got updated relatively recently, but they depend on some IR.m files in the Extras folder that were not updated. Not sure if you found your own temp fixes by now, but what I found to work on some similar issues was to add the following snippets to a few files:
IRoption_fields.m
at line 71, but really could go anywhere in the allfields
cell:
'warmrestart';
'SparsityTrans';
'wname';
'wlevels';
'discrflatTol';
IRset.m
at line 287 (again could be anywhere in the switch):
case {'warmrestart'} % off, on
[validvalue, errmsg] = onOffType(field,value);
case {'SparsityTrans'} % none, dwt
[validvalue, errmsg] = sparsityTransType(field,value);
case {'wname'} % db1 default, probably others...
[validvalue, errmsg] = wnameType(field,value);
case {'wlevels'} % probably a positive value?
[validvalue, errmsg] = nonNegscalar(field,value);
case {'discrflatTol'} % probably a positive value?
and then also the newly defined functions for sparsity and wname that I threw at the end of the file:
%------------------------------------------------------------------------
function [valid, errmsg] = sparsityTransType(field,value)
% One of these strings: none, dwt
valid = (ischar(value) && any(strcmpi(value,{'none';'dwt'})));
if ~valid
errmsg = sprintf('Invalid value for OPTIONS parameter %s: must be ''dwt'' or ''none''.',field);
else
errmsg = '';
end
%------------------------------------------------------------------------
function [valid, errmsg] = wnameType(field,value)
% One of these strings: none, dwt
valid = (ischar(value) && any(strcmpi(value,{'none';'db1'})));
if ~valid
errmsg = sprintf('Invalid value for OPTIONS parameter %s: must be ''db1''.',field);
else
errmsg = '';
end
I'm sure there's other things missing, but this at least has me up and running so far.
Experiencing a similar issue with a number of the algorithms which I hope I can fix using the above suggestions.
I believe this issue is due to the fact that most of the IR.m files in the IRtools folder got updated relatively recently, but they depend on some IR.m files in the Extras folder that were not updated. Not sure if you found your own temp fixes by now, but what I found to work on some similar issues was to add the following snippets to a few files:
IRoption_fields.m
at line 71, but really could go anywhere in theallfields
cell:'warmrestart'; 'SparsityTrans'; 'wname'; 'wlevels'; 'discrflatTol';
IRset.m
at line 287 (again could be anywhere in the switch):case {'warmrestart'} % off, on [validvalue, errmsg] = onOffType(field,value); case {'SparsityTrans'} % none, dwt [validvalue, errmsg] = sparsityTransType(field,value); case {'wname'} % db1 default, probably others... [validvalue, errmsg] = wnameType(field,value); case {'wlevels'} % probably a positive value? [validvalue, errmsg] = nonNegscalar(field,value); case {'discrflatTol'} % probably a positive value?
and then also the newly defined functions for sparsity and wname that I threw at the end of the file:
%------------------------------------------------------------------------ function [valid, errmsg] = sparsityTransType(field,value) % One of these strings: none, dwt valid = (ischar(value) && any(strcmpi(value,{'none';'dwt'}))); if ~valid errmsg = sprintf('Invalid value for OPTIONS parameter %s: must be ''dwt'' or ''none''.',field); else errmsg = ''; end %------------------------------------------------------------------------ function [valid, errmsg] = wnameType(field,value) % One of these strings: none, dwt valid = (ischar(value) && any(strcmpi(value,{'none';'db1'}))); if ~valid errmsg = sprintf('Invalid value for OPTIONS parameter %s: must be ''db1''.',field); else errmsg = ''; end
I'm sure there's other things missing, but this at least has me up and running so far.
Do you have a case for discflatTol? the ell1 solver doesn't run without it
I don't have access to my code right now, but I'd guess it's the same as the wlevels case. I.e. Checking for a non neg scalar. Sorry, I probably just didn't copy the next line like I should have!
I don't have access to my code right now, but I'd guess it's the same as the wlevels case. I.e. Checking for a non neg scalar. Sorry, I probably just didn't copy the next line like I should have!
That allowed the code to proceed but then it crashes at 'hybridvariant' more to figure out. Thanks very much for the help!!
The error message when running IRSparsity: