def execute(self, context):
if self.typ == 1:
handle = handle_read(self.handle_name)
prop = handle[1]
Elman = SvNeuro_Elman()
if handle[0]:
prop['wA']=Elman.init_w(prop['InA'], prop['InB'], prop['trashold'])
prop['wB']=Elman.init_w(prop['InB'], prop['InC'], prop['trashold'])
handle_write(self.handle_name, prop)
return {'FINISHED'}
def execute(self, context):
if self.typ == 1:
handle = handle_read(self.handle_name)
prop = handle[1]
Elman = SvNeuro_Elman()
if handle[0]:
prop['wA']=Elman.init_w(prop['InA'], prop['InB'], prop['trashold'])
prop['wB']=Elman.init_w(prop['InB'], prop['InC'], prop['trashold'])
handle_write(self.handle_name, prop)
return {'FINISHED'}
def draw_buttons(self, context, layout):
#row.prop(self, 'groupname', text='')
col = layout.column(align=True)
row = col.row(align=True)
row.prop_search(self,
'groupname',
bpy.data,
'groups',
text='',
icon='HAND')
row = col.row()
addon = context.user_preferences.addons.get(sverchok.__name__)
if addon.preferences.over_sized_buttons:
row.scale_y = 4.0
op_text = "G E T"
else:
row.scale_y = 1
op_text = "Get selection"
opera = row.operator('node.sverchok_object_insertion', text=op_text)
opera.node_name = self.name
opera.tree_name = self.id_data.name
opera.grup_name = self.groupname
opera.sort = self.sort
col = layout.column(align=True)
row = col.row(align=True)
row.prop(self, 'sort', text='Sort', toggle=True)
row.prop(self, "modifiers", text="Post", toggle=True)
# row = layout.row(align=True)
row.prop(self, "vergroups", text="VeGr", toggle=True)
row = col.row(align=True)
opera = row.operator('node.sverchok_object_in_selector', text='Select')
opera.node_name = self.name
opera.tree_name = self.id_data.name
handle = handle_read(self.name + self.id_data.name)
if self.objects_local:
if handle[0]:
for i, o in enumerate(handle[1]):
if i > 4:
layout.label('. . . more ' + str(len(handle[1]) - 5) +
' items')
break
layout.label(o)
else:
handle_write(self.name + self.id_data.name,
literal_eval(self.objects_local))
else:
layout.label('--None--')
def draw_buttons(self, context, layout):
#row.prop(self, 'groupname', text='')
col = layout.column(align=True)
row = col.row(align=True)
row.prop_search(self, 'groupname', bpy.data, 'groups', text='', icon='HAND')
row = col.row()
addon = context.user_preferences.addons.get(sverchok.__name__)
if addon.preferences.over_sized_buttons:
row.scale_y = 4.0
op_text = "G E T"
else:
row.scale_y = 1
op_text = "Get selection"
opera = row.operator('node.sverchok_object_insertion', text=op_text)
opera.node_name = self.name
opera.tree_name = self.id_data.name
opera.grup_name = self.groupname
opera.sort = self.sort
col = layout.column(align=True)
row = col.row(align=True)
row.prop(self, 'sort', text='Sort', toggle=True)
row.prop(self, "modifiers", text="Post", toggle=True)
# row = layout.row(align=True)
row.prop(self, "vergroups", text="VeGr", toggle=True)
row = col.row(align=True)
opera = row.operator('node.sverchok_object_in_selector', text='Select')
opera.node_name = self.name
opera.tree_name = self.id_data.name
handle = handle_read(self.name + self.id_data.name)
if self.objects_local:
if handle[0]:
for i, o in enumerate(handle[1]):
if i > 4:
layout.label('. . . more ' + str(len(handle[1]) - 5) + ' items')
break
layout.label(o)
else:
handle_write(self.name + self.id_data.name, literal_eval(self.objects_local))
else:
layout.label('--None--')
def enable(self, name_no, name_tr, handle, sorting):
objects = []
if self.grup_name and bpy.data.groups[self.grup_name].objects:
objs = bpy.data.groups[self.grup_name].objects
elif bpy.context.selected_objects:
objs = bpy.context.selected_objects
else:
self.report({'WARNING'}, 'No object selected')
return
for o in objs:
objects.append(o.name)
if sorting:
objects.sort()
handle_write(name_no + name_tr, objects)
if bpy.data.node_groups[name_tr]:
handle = handle_read(name_no + name_tr)
bpy.data.node_groups[name_tr].nodes[name_no].objects_local = str(handle[1])
def enable(self, name_no, name_tr, handle, sorting):
objects = []
if self.grup_name and bpy.data.groups[self.grup_name].objects:
objs = bpy.data.groups[self.grup_name].objects
elif bpy.context.selected_objects:
objs = bpy.context.selected_objects
else:
self.report({'WARNING'},'No object selected')
return
for o in objs:
objects.append(o.name)
if sorting:
objects.sort()
handle_write(name_no+name_tr, objects)
# временное решение с группой. надо решать, как достать имя группы узлов
if bpy.data.node_groups[name_tr]:
handle = handle_read(name_no+name_tr)
#print ('exec',name)
bpy.data.node_groups[name_tr].nodes[name_no].objects_local = str(handle[1])
def draw_buttons(self, context, layout):
row = layout.row()
addon = context.user_preferences.addons.get(sverchok.__name__)
if addon.preferences.over_sized_buttons:
row.scale_y = 4.0
op_text = "G E T"
else:
row.scale_y = 1
op_text = "Get selection"
opera = row.operator("node.sverchok_object_insertion", text=op_text)
opera.node_name = self.name
opera.tree_name = self.id_data.name
opera.grup_name = self.groupname
opera.sort = self.sort
row = layout.row(align=True)
opera = row.operator("node.sverchok_object_in_selector", text="Select")
opera.node_name = self.name
opera.tree_name = self.id_data.name
row = layout.row(align=True)
row.prop(self, "groupname", text="")
row.prop(self, "sort", text="Sort objects")
row = layout.row(align=True)
row.prop(self, "modifiers", text="Post modifiers")
# row = layout.row(align=True)
row.prop(self, "vergroups", text="Vert groups")
handle = handle_read(self.name + self.id_data.name)
if self.objects_local:
if handle[0]:
for i, o in enumerate(handle[1]):
if i > 4:
layout.label(". . . more " + str(len(handle[1]) - 5) + " items")
break
layout.label(o)
else:
handle_write(self.name + self.id_data.name, literal_eval(self.objects_local))
else:
layout.label("--None--")
def process(self):
handle_name = self.name + self.id_data.name
handle = handle_read(handle_name)
props = handle[1]
if not handle[0]:
elman = SvNeuroElman()
props = {
'InA': 2,
'InB': 5,
'InC': 1,
'wA': [],
'wB': [],
'gister': 0.01,
'k_learning': 0.1,
'epsilon': 1.3,
'cycles': 3,
'trashold': 0.01,
'k_lambda': 0.0001,
'Elman': elman,
}
self.elman = props['Elman']
result = []
if self.outputs['result'].is_linked and self.inputs['data'].is_linked:
if self.inputs['etalon'].is_linked:
input_etalon = self.inputs['etalon'].sv_get()
is_learning = True
else:
input_etalon = [[0]]
is_learning = False
if (props['InA'] != self.lA +
1) or props['InB'] != self.lB or props['InC'] != self.lC:
props['InA'] = self.lA + 1
props['InB'] = self.lB
props['InC'] = self.lC
props['wA'] = self.elman.init_w(props['InA'], props['InB'],
props['trashold'])
props['wB'] = self.elman.init_w(props['InB'], props['InC'],
props['trashold'])
props['gister'] = self.gisterezis
props['k_learning'] = self.k_learning
props['epsilon'] = self.epsilon
props['k_lambda'] = self.k_lambda
props['cycles'] = self.cycles
props['trashold'] = self.treshold
input_data = self.inputs['data'].sv_get()
if type(input_etalon[0]) not in [list, tuple]:
input_etalon = [input_etalon]
if type(input_data[0]) not in [list, tuple]:
input_data = [input_data]
for idx, data in enumerate(input_data):
let = len(input_etalon) - 1
eta = input_etalon[min(idx, let)]
data2 = [1.0] + data
if type(eta) not in [list, tuple]:
eta = [eta]
result.append(
self.elman.neuro(data2, eta, self.maximum, is_learning,
props))
else:
result = [[]]
handle_write(handle_name, props)
self.outputs['result'].sv_set(result)
def process(self):
name = self.name + self.id_data.name
handle = handle_read(name)
#reload handle if possible
if self.objects_local and not handle[0]:
handle_write(name, literal_eval(self.objects_local))
handle = handle_read(name)
if handle[0]:
objs = handle[1]
edgs_out = []
vers_out = []
vers_out_grouped = []
pols_out = []
mtrx_out = []
for obj_ in objs: # names of objects
edgs = []
vers = []
vers_grouped = []
pols = []
mtrx = []
obj = bpy.data.objects[obj_] # objects itself
if obj.type == 'EMPTY':
for m in obj.matrix_world:
mtrx.append(m[:])
else:
scene = bpy.context.scene
settings = 'PREVIEW'
# create a temporary mesh
obj_data = obj.to_mesh(scene, self.modifiers, settings)
for m in obj.matrix_world:
mtrx.append(list(m))
for k, v in enumerate(obj_data.vertices):
if self.vergroups and v.groups.values():
vers_grouped.append(k)
vers.append(list(v.co))
edgs = obj_data.edge_keys
for p in obj_data.polygons:
pols.append(list(p.vertices))
# remove the temp mesh
bpy.data.meshes.remove(obj_data)
edgs_out.append(edgs)
vers_out.append(vers)
vers_out_grouped.append(vers_grouped)
pols_out.append(pols)
mtrx_out.append(mtrx)
if vers_out[0]:
if self.outputs['Vertices'].is_linked:
SvSetSocketAnyType(self, 'Vertices', vers_out)
if self.outputs['Edges'].is_linked:
SvSetSocketAnyType(self, 'Edges', edgs_out)
if self.outputs['Polygons'].is_linked:
SvSetSocketAnyType(self, 'Polygons', pols_out)
if self.vergroups and self.outputs['Vers_grouped'].is_linked:
SvSetSocketAnyType(self, 'Vers_grouped', vers_out_grouped)
if self.outputs['Matrixes'].is_linked:
SvSetSocketAnyType(self, 'Matrixes', mtrx_out)
def copy(self, node):
if self.objects_local:
name = self.name + self.id_data.name
handle_write(name, literal_eval(self.objects_local))
def process(self):
handle_name = self.name + self.id_data.name
handle = handle_read(handle_name)
props = handle[1]
if not handle[0]:
props = {'InA':2,
'InB':5,
'InC':1,
'wA':[],
'wB':[],
'gister':0.01,
'k_learning':0.1,
'epsilon':1.3,
'cycles':3,
'trashold':0.01,
'k_lambda':0.0001}
props['wA'] = self.Elman.init_w(props['InA'], props['InB'], props['trashold'])
props['wB'] = self.Elman.init_w(props['InB'], props['InC'], props['trashold'])
self.Elman.gister = abs(self.gisterezis)
self.Elman.k_learning = self.k_learning
result = []
if self.outputs['result'].is_linked and self.inputs['data'].is_linked:
if self.inputs['etalon'].is_linked:
etalon = self.inputs['etalon'].sv_get()[0]
flag = True
else:
flag = False
etalon = [[0]]
if (props['InA']!=self.lA+1) or props['InB']!=self.lB or props['InC']!=self.lC:
props['InA'] = self.lA+1
props['InB'] = self.lB
props['InC'] = self.lC
props['wA'] = self.Elman.init_w(props['InA'], props['InB'])
props['wB'] = self.Elman.init_w(props['InB'], props['InC'])
props['gister'] = self.gisterezis
props['k_learning'] = self.k_learning
props['epsilon'] = self.epsilon
props['k_lambda'] = self.k_lambda
props['cycles'] = self.cycles
props['trashold'] = self.treshold
data_ = self.inputs['data'].sv_get()[0]
if type(etalon[0]) not in [list, tuple]: etalon = [etalon]
if type(data_[0]) not in [list, tuple]: data_ = [data_]
for idx, data in enumerate(data_):
let = len(etalon)-1
eta = etalon[min(idx,let)]
data2 = [1.0]+data
if type(eta) not in [list, tuple]: eta = [eta]
result.append([self.Elman.neuro(data2, eta, self.maximum, flag, props)])
else:
result = [[[]]]
handle_write(handle_name, props)
self.outputs['result'].sv_set(result)
def process(self):
handle_name = self.name + self.id_data.name
handle = handle_read(handle_name)
props = handle[1]
if not handle[0]:
props = {'InA':2,
'InB':5,
'InC':1,
'wA':[],
'wB':[],
'gister':0.01,
'k_learning':0.1,
'epsilon':1.3,
'cycles':3,
'trashold':0.01,
'k_lambda':0.0001}
props['wA'] = self.Elman.init_w(props['InA'], props['InB'], props['trashold'])
props['wB'] = self.Elman.init_w(props['InB'], props['InC'], props['trashold'])
self.Elman.gister = abs(self.gisterezis)
self.Elman.k_learning = self.k_learning
result = []
if 'result' in self.outputs and len(self.outputs['result'].links) > 0 \
and 'data' in self.inputs and len(self.inputs['data'].links) > 0:
if 'etalon' in self.inputs and len(self.inputs['etalon'].links) > 0:
etalon = SvGetSocketAnyType(self, self.inputs['etalon'])[0]
flag = True
else:
flag = False
etalon = [[0]]
if (props['InA']!=self.lA+1) or props['InB']!=self.lB or \
props['InC']!=self.lC:
props['InA'] = self.lA+1
props['InB'] = self.lB
props['InC'] = self.lC
props['wA'] = self.Elman.init_w(props['InA'], props['InB'])
props['wB'] = self.Elman.init_w(props['InB'], props['InC'])
props['gister'] = self.gisterezis
props['k_learning'] = self.k_learning
props['epsilon'] = self.epsilon
props['k_lambda'] = self.k_lambda
props['cycles'] = self.cycles
props['trashold'] = self.treshold
data_ = SvGetSocketAnyType(self, self.inputs['data'])[0]
if type(etalon[0]) not in [list, tuple]: etalon = [etalon]
if type(data_[0]) not in [list, tuple]: data_ = [data_]
for idx, data in enumerate(data_):
let = len(etalon)-1
eta = etalon[min(idx,let)]
data2 = [1.0]+data
if type(eta) not in [list, tuple]: eta = [eta]
result.append([self.Elman.neuro(data2, eta, self.maximum, flag, props)])
else:
result = [[[]]]
handle_write(handle_name, props)
SvSetSocketAnyType(self, 'result', result)
