Merge pull request #36 from inwoo-park/simba

NEW: add "loadmodel_netcdf.py" in "src/m/contrib/inwoo"

src/m/contrib/inwoo/loadmodel_netcdf.m

@@ -0,0 +1,38 @@
+%{
+% Explain
+%  load netcdf format ISSM model (~ export_netCDF).
+%
+% Usage
+%  md = loadmodel_netcdf('Antarctica.nc');
+%
+% Inputs
+%
+% Outputs
+%}
+function md = loadmodel_netcdf(fname)
+	warning('WARNING: this script is under developement.');
+
+	% initialize model
+	md = model;
+
+	% get information.
+	info = ncinfo(fname);
+
+	% get groups;
+	GroupNames = {info.Groups.Name};
+
+	for i = 1:length(GroupNames)
+		GroupName = GroupNames{i};
+		posGroup = strcmpi(GroupName, GroupNames); 
+
+		disp(GroupName);
+		for j = 1:length(info.Groups(posGroup).Variables)
+			VarName = info.Groups(posGroup).Variables(j).Name;
+			md.(GroupName).(VarName) = transpose(ncread(fname,[GroupName '/' VarName]));
+		end
+		break;
+	end
+
+	% figure out groups!
+
+end

src/m/contrib/inwoo/loadmodel_netcdf.py

@@ -0,0 +1,165 @@
+#!/usr/bin/env python3
+import warnings
+from model import *
+from pairoptions import pairoptions
+import netCDF4
+import copy
+import pydoc
+from results import solution
+
+__all__ = ['loadmodel_netcdf']
+def str2bool(v:str): # {{{
+    '''
+    Okay, convert string type bool, True or False, to bool
+    '''
+    if not isinstance(v,str):
+        raise Exception('ERROR: Given type of input value (={type(v)}) is not string.')
+
+    if v in ['True','1']:
+        return True
+    else:
+        return False
+    # }}}
+
+def loadmodel_netcdf(filename, verbose:bool=0):
+    '''
+    Explain
+    -------
+    Import model dataset and results from netcdf format. Netcdf format ISSM model result is generated with "export_NetCDF".
+
+    Example
+    -------
+    >>> import loadmodel_netcdf import loadmodel_netcdf
+    >>> md = loadmodel_netcdf('./model.nc') 
+
+    Parameters
+    ----------
+    verbose:bool (default: 0)
+        Show progress with verbosity.
+
+    Returns
+    -------
+    md: model class in ISSM.
+    '''
+    # initialize model class..
+    md = model();
+
+    # check file name type
+    if not filename.endswith('.nc'):
+        raise Exception('ERROR: we cannot load %s. This file should be ".nc"(netCDF) file format.')
+
+    # loadmodel
+    nc = netCDF4.Dataset(filename, 'r')
+
+    # check groups..
+    groupNames = np.array(list(nc.groups.keys()))
+    for group, groupName in zip(nc.groups, groupNames):
+
+        # configure variables in class...
+        try:
+            if (not nc[groupName].variables == {}) | \
+                    (groupName in ['private','debug','miscellaneous','verbose','outputdefinition']):
+                    #(nc[groupName].classtype == 'results.results') | \
+                    #(nc[groupName].classtype == 'private.private') | \
+                    #(nc[groupName].classtype == 'debug.debug') | \
+                    #(nc[groupName].classtype == 'miscellaneous.miscellaneous') | \
+                    #(nc[groupName].classtype == 'verbose.verbose') | \
+                    #(nc[groupName].classtype == 'outputdefinition.outputdefinition'):
+                mclass = nc[groupName].classtype.split('.')[0]
+            elif groupName in ['results']:
+                mclass = 'results'
+            else: # skip if we cannot configure specific group.
+                '''
+                Current status
+                Some class will be skipeed.
+                '''
+                warnings.warn(f'WARNING: Skip group = {groupName}')
+                continue
+        except:
+            print(nc.groups)
+            raise Exception('ERROR: we cannot find "%s" group.'%(group))
+
+        # get class group.. such as mesh. geometry, inversion, initialization, material etc...
+        if hasattr(nc.groups[group],'classgroup'):
+            gclass = nc[group].classgroup
+        elif hasattr(nc.groups[group],'classtype'):
+            gclass = nc.groups[group].classtype
+        else:
+            gclass = None
+
+        if groupName == 'results': # results special.
+            # Loading results with "solution" in ISSM class.
+            for solutionname in nc['results'].groups.keys():
+                print('loadmodel_netcdf: assign md.results.%s'%(solutionname))
+                _solution = solution()
+                for fieldname in nc['results'][solutionname].variables.keys():
+                    #print(solutionname, fieldname)
+                    setattr(_solution, fieldname, np.array(nc['results'][solutionname][fieldname]))
+
+                setattr(md.results, solutionname, _solution)
+        elif groupName == 'private': 
+            # For private
+            md.private.runtimename  = nc[groupName].runtimename
+            md.private.isconsistent = str2bool(nc[groupName].isconsistent)
+            md.private.solution     = nc[groupName].solution
+        elif groupName == 'debug':
+            md.debug.valgrind  = str2bool(nc[groupName].valgrind)
+            md.debug.gprof     = str2bool(nc[groupName].gprof)
+            md.debug.profiling = str2bool(nc[groupName].profiling)
+        elif groupName == 'verbose':
+            tmp = nc[groupName]
+            md.verbose.mprocessor  = str2bool(tmp.mprocessor)
+            md.verbose.module      = str2bool(tmp.module)
+            try:
+                md.verbose.solution    = str2bool(tmp.solution)
+            except:
+                warnings.warn('Skip loading "md.verbose.solution"')
+            md.verbose.solver      = str2bool(tmp.solver)
+            md.verbose.convergence = str2bool(tmp.convergence)
+            try:
+                md.verbose.control     = str2bool(tmp.control)
+            except:
+                warnings.warn('Skip loading "md.verbose.control"')
+            md.verbose.qmu         = str2bool(tmp.qmu)
+            md.verbose.autodiff    = str2bool(tmp.autodiff)
+            md.verbose.smb         = str2bool(tmp.smb)
+        elif groupName == 'miscellaneous':
+            # For "miscellaneous"
+            md.miscellaneous.notes = nc[groupName].notes
+            md.miscellaneous.name  = nc[groupName].varname
+        elif groupName == 'outputdefinition':
+            definitions = md.outputdefinition.definitions
+            if definitions == 'emtpycell':
+                md.outputdefinition.definitions = []
+            else:
+                md.outputdefinition.definitions = definitions
+        elif (not mclass in ['verbose',
+                                'hpc_simba','fourierlove',
+                                ]) & isinstance(gclass,str): # skip specific class..
+            # show process..
+            if verbose:
+                print('    class: {} / classtype: {}'.format(gclass,mclass))
+
+            # get class...
+            mod=getattr(__import__(mclass),mclass)()
+
+            # get class
+            setattr(md, group, mod)
+
+            # assign loaded calss to md.(field)
+            variables = list(nc.groups[group].variables)
+            for variable in variables:
+                value = nc.groups[group][variable]
+                try:
+                    if value.shape == ():
+                        setattr(getattr(md,group),variable,value[:])
+                    else:
+                        setattr(getattr(md,group),variable,np.array(value))
+                except:
+                    value = nc['%s/%s'%(group,variable)]
+                    print(value)
+                    print(value.shape)
+                    raise Exception(f'ERROR: we cannot find {variable}')
+
+    # return outputs.
+    return md