def write_events(self):
model = self.parser.parsedModel
for i in range(len(model.listOfEvents)):
self.out_file.write(" if( ")
self.out_file.write(mathml_condition_parser(model.EventCondition[i]))
self.out_file.write("){\n")
list_of_assignment_rules = model.listOfEvents[i].getListOfEventAssignments()
for j in range(len(list_of_assignment_rules)):
self.out_file.write(" ")
event_variable = model.eventVariable[i][j]
if not (event_variable in model.speciesId):
self.out_file.write(event_variable)
else:
string = "y[" + repr(model.speciesId.index(event_variable)) + "]"
self.out_file.write(string)
self.out_file.write("=")
string = model.EventFormula[i][j]
for q in range(len(model.speciesId)):
string = self.rep(string, model.speciesId[q], 'y[' + repr(q) + ']')
for q in range(len(model.parameterId)):
parameter_id = model.parameterId[q]
if not (parameter_id in model.ruleVariable):
flag = False
for r in range(len(model.eventVariable)):
if parameter_id in model.eventVariable[r]:
flag = True
if not flag:
string = self.rep(string, parameter_id, 'tex2D(param_tex,' + repr(q) + ',tid)')
self.out_file.write(string)
self.out_file.write(";\n")
self.out_file.write(" }\n")
self.out_file.write("\n")
def write_assignment_rules(self):
model = self.parser.parsedModel
for i in range(len(model.listOfRules)):
if model.listOfRules[i].isAssignment():
self.out_file.write(" ")
rule_variable = model.ruleVariable[i]
if not (rule_variable in model.speciesId):
self.out_file.write("float ")
self.out_file.write(rule_variable)
else:
string = "y[" + repr(model.speciesId.index(rule_variable)) + "]"
self.out_file.write(string)
self.out_file.write("=")
string = mathml_condition_parser(model.ruleFormula[i])
for q in range(len(model.speciesId)):
string = self.rep(string, model.speciesId[q], 'y[' + repr(q) + ']')
for q in range(len(model.parameterId)):
parameter_id = model.parameterId[q]
if not (parameter_id in model.ruleVariable):
flag = False
for r in range(len(model.eventVariable)):
if parameter_id in model.eventVariable[r]:
flag = True
if not flag:
string = self.rep(string, parameter_id, 'tex2D(param_tex,' + repr(q) + ',tid)')
self.out_file.write(string)
self.out_file.write(";\n")
self.out_file.write("\n")
def write(self):
"""
Write the cuda file with ODE functions using the information taken by the parser
"""
num_species = len(self.parser.parsedModel.species)
p = re.compile('\s')
# Write number of parameters and species
self.out_file.write("#define NSPECIES " + str(num_species) + "\n")
self.out_file.write("#define NPARAM " +
str(len(self.parser.parsedModel.parameterId)) +
"\n")
self.out_file.write("#define NREACT " +
str(self.parser.parsedModel.numReactions) + "\n")
self.out_file.write("\n")
# The user-defined functions used in the model must be written in the file
self.out_file.write("//Code for texture memory\n")
num_events = len(self.parser.parsedModel.listOfEvents)
num_rules = len(self.parser.parsedModel.listOfRules)
num = num_events + num_rules
if num > 0:
self.out_file.write("#define leq(a,b) a<=b\n")
self.out_file.write("#define neq(a,b) a!=b\n")
self.out_file.write("#define geq(a,b) a>=b\n")
self.out_file.write("#define lt(a,b) a<b\n")
self.out_file.write("#define gt(a,b) a>b\n")
self.out_file.write("#define eq(a,b) a==b\n")
self.out_file.write("#define and_(a,b) a&&b\n")
self.out_file.write("#define or_(a,b) a||b\n")
for i in range(len(self.parser.parsedModel.listOfFunctions)):
arg_string = ",".join(
map(lambda s: "float " + s,
self.parser.parsedModel.functionArgument[i]))
self.out_file.write(
"__device__ float %s (%s){\n" %
(self.parser.parsedModel.listOfFunctions[i].getId(),
arg_string))
self.out_file.write(" return %s;\n" %
self.parser.parsedModel.functionBody[i])
self.out_file.write("}\n\n")
self.out_file.write("\n")
self.out_file.write(
"__device__ void step(float *y, float t, unsigned *rngRegs, int tid){\n"
)
num_species = len(self.parser.parsedModel.species)
# write rules and events
for i in range(len(self.parser.parsedModel.listOfRules)):
if self.parser.parsedModel.listOfRules[i].isRate():
rule_variable = self.parser.parsedModel.ruleVariable[i]
if not (rule_variable in self.parser.parsedModel.speciesId):
self.out_file.write(" %s = " % rule_variable)
else:
self.out_file.write(
" y[%s] = " %
self.parser.parsedModel.speciesId.index(rule_variable))
string = self.parser.parsedModel.ruleFormula[i]
string = self.replace_names(string, 'tex2D(param_tex,%s,tid)')
self.out_file.write("%s;\n" % string)
for i in range(len(self.parser.parsedModel.listOfEvents)):
self.out_file.write(" if( ")
self.out_file.write(
mathml_condition_parser(
self.parser.parsedModel.EventCondition[i]))
self.out_file.write("){\n")
list_of_assignment_rules = self.parser.parsedModel.listOfEvents[
i].getListOfEventAssignments()
for j in range(len(list_of_assignment_rules)):
event_variable = self.parser.parsedModel.eventVariable[i][j]
if not (event_variable in self.parser.parsedModel.speciesId):
self.out_file.write(" %s = " % event_variable)
else:
self.out_file.write(
" y[%s] = " %
self.parser.parsedModel.speciesId.index(event_variable)
)
string = self.parser.parsedModel.EventFormula[i][j]
string = self.replace_names(string, 'tex2D(param_tex,%s,tid)')
self.out_file.write("%s;\n" % string)
self.out_file.write("}\n")
self.out_file.write("\n")
for i in range(len(self.parser.parsedModel.listOfRules)):
if self.parser.parsedModel.listOfRules[i].isAssignment():
rule_variable = self.parser.parsedModel.ruleVariable[i]
if not (rule_variable in self.parser.parsedModel.speciesId):
self.out_file.write(" float %s = " % rule_variable)
else:
self.out_file.write(
" y[%s] = " %
self.parser.parsedModel.speciesId.index(rule_variable))
string = mathml_condition_parser(
self.parser.parsedModel.ruleFormula[i])
string = self.replace_names(string, 'tex2D(param_tex,%s,tid)')
self.out_file.write("%s;\n" % string)
# Write the derivatives
for i in range(num_species):
species_id = self.parser.parsedModel.species[i]
if species_id.getConstant(
) == False and species_id.getBoundaryCondition() == False:
self.out_file.write(" float d_y%s = DT * (" % i)
if species_id.isSetCompartment():
self.out_file.write("(")
reaction_written = False
for k in range(self.parser.parsedModel.numReactions):
if not self.parser.parsedModel.stoichiometricMatrix[i][
k] == 0.0:
if reaction_written and self.parser.parsedModel.stoichiometricMatrix[
i][k] > 0.0:
self.out_file.write("+")
reaction_written = True
self.out_file.write(
repr(
self.parser.parsedModel.stoichiometricMatrix[i]
[k]))
self.out_file.write("*(")
string = self.parser.parsedModel.kineticLaw[k]
string = self.replace_names(string,
'tex2D(param_tex,%s,tid)')
string = p.sub('', string)
self.out_file.write(string)
self.out_file.write(")")
if species_id.isSetCompartment():
self.out_file.write(")/")
my_species_compartment = species_id.getCompartment()
for j in range(len(
self.parser.parsedModel.listOfParameter)):
if self.parser.parsedModel.listOfParameter[j].getId(
) == my_species_compartment:
parameter_id = self.parser.parsedModel.parameterId[
j]
if not (parameter_id
in self.parser.parsedModel.ruleVariable):
flag = False
for r in range(
len(self.parser.parsedModel.
eventVariable)):
if parameter_id in self.parser.parsedModel.eventVariable[
r]:
flag = True
if not flag:
self.out_file.write(
"tex2D(param_tex,%s,tid));" % j)
break
else:
self.out_file.write("%s);" % parameter_id)
break
else:
self.out_file.write(");")
self.out_file.write("\n")
self.out_file.write("\n")
# check for columns of the stochiometry matrix with more than one entry
random_variables = ["*randNormal(rngRegs,sqrt(DT))"
] * self.parser.parsedModel.numReactions
for k in range(self.parser.parsedModel.numReactions):
num_entries = 0
for i in range(num_species):
if self.parser.parsedModel.stoichiometricMatrix[i][k] != 0.0:
num_entries += 1
# define specific randomVariable
if num_entries > 1:
self.out_file.write(
" float rand%s = randNormal(rngRegs,sqrt(DT));\n" % k)
random_variables[k] = "*rand%s" % k
self.out_file.write("\n")
# write noise terms
for i in range(num_species):
species = self.parser.parsedModel.species[i]
if not (species.getConstant() or species.getBoundaryCondition()):
self.out_file.write(" d_y%s += (" % i)
if species.isSetCompartment():
self.out_file.write("(")
reaction_written = False
for k in range(self.parser.parsedModel.numReactions):
if not self.parser.parsedModel.stoichiometricMatrix[i][
k] == 0.0:
if reaction_written and self.parser.parsedModel.stoichiometricMatrix[
i][k] > 0.0:
self.out_file.write("+")
reaction_written = True
self.out_file.write(
repr(
self.parser.parsedModel.stoichiometricMatrix[i]
[k]))
self.out_file.write("*sqrt(")
string = self.parser.parsedModel.kineticLaw[k]
string = self.replace_names(string,
'tex2D(param_tex,%s,tid)')
string = p.sub('', string)
self.out_file.write(string)
# multiply random variable
self.out_file.write(")")
self.out_file.write(random_variables[k])
if species.isSetCompartment():
self.out_file.write(")/")
my_species_compartment = species.getCompartment()
for j in range(len(
self.parser.parsedModel.listOfParameter)):
if self.parser.parsedModel.listOfParameter[j].getId(
) == my_species_compartment:
parameter_id = self.parser.parsedModel.parameterId[
j]
if not (parameter_id
in self.parser.parsedModel.ruleVariable):
flag = False
for r in range(
len(self.parser.parsedModel.
eventVariable)):
if parameter_id in self.parser.parsedModel.eventVariable[
r]:
flag = True
if not flag:
self.out_file.write(
"tex2D(param_tex,%s,tid))" % j)
break
else:
self.out_file.write(parameter_id + ")")
break
else:
self.out_file.write(")")
self.out_file.write(";\n")
self.out_file.write("\n")
# add terms
for i in range(num_species):
species = self.parser.parsedModel.species[i]
if species.getConstant() == False and species.getBoundaryCondition(
) == False:
self.out_file.write(" y[%s] += d_y%s;\n" % (i, i))
self.out_file.write("}\n")
# same file
p = re.compile('\s')
# The user-defined functions used in the model must be written in the file
self.out_file.write("//Code for shared memory\n")
num_events = len(self.parser.parsedModel.listOfEvents)
num_rules = len(self.parser.parsedModel.listOfRules)
num = num_events + num_rules
if num > 0:
self.out_file.write("#define leq(a,b) a<=b\n")
self.out_file.write("#define neq(a,b) a!=b\n")
self.out_file.write("#define geq(a,b) a>=b\n")
self.out_file.write("#define lt(a,b) a<b\n")
self.out_file.write("#define gt(a,b) a>b\n")
self.out_file.write("#define eq(a,b) a==b\n")
self.out_file.write("#define and_(a,b) a&&b\n")
self.out_file.write("#define or_(a,b) a||b\n")
for i in range(len(self.parser.parsedModel.listOfFunctions)):
function = self.parser.parsedModel.listOfFunctions[i]
arg_string = ",".join(
map(lambda s: "float " + s,
self.parser.parsedModel.functionArgument[i]))
self.out_file.write("__device__ float %s (%s){\n" %
(function.getId(), arg_string))
self.out_file.write(" return %s;\n" %
self.parser.parsedModel.functionBody[i])
self.out_file.write("}\n\n")
self.out_file.write("\n")
self.out_file.write(
"__device__ void step(float *parameter, float *y, float t, unsigned *rngRegs){\n"
)
num_species = len(self.parser.parsedModel.species)
# write rules and events
for i in range(len(self.parser.parsedModel.listOfRules)):
if self.parser.parsedModel.listOfRules[i].isRate():
rule_variable = self.parser.parsedModel.ruleVariable[i]
if not (rule_variable in self.parser.parsedModel.speciesId):
self.out_file.write(
" %s = " % self.parser.parsedModel.ruleVariable[i])
else:
self.out_file.write(" y[%s] = " % repr(
self.parser.parsedModel.speciesId.index(rule_variable))
)
string = self.parser.parsedModel.ruleFormula[i]
string = self.replace_names(string, 'parameter[%s]')
self.out_file.write("%s;\n" % string)
for i in range(len(self.parser.parsedModel.listOfEvents)):
self.out_file.write(" if( ")
self.out_file.write(
mathml_condition_parser(
self.parser.parsedModel.EventCondition[i]))
self.out_file.write("){\n")
list_of_assignment_rules = self.parser.parsedModel.listOfEvents[
i].getListOfEventAssignments()
for j in range(len(list_of_assignment_rules)):
event_variable = self.parser.parsedModel.eventVariable[i][j]
if not (event_variable in self.parser.parsedModel.speciesId):
self.out_file.write(" %s =" % event_variable)
else:
self.out_file.write(
" y[%s] =" %
self.parser.parsedModel.speciesId.index(event_variable)
)
string = self.parser.parsedModel.EventFormula[i][j]
string = self.replace_names(string, 'parameter[%s]')
self.out_file.write("%s;\n" % string)
self.out_file.write("}\n")
self.out_file.write("\n")
for i in range(len(self.parser.parsedModel.listOfRules)):
if self.parser.parsedModel.listOfRules[i].isAssignment():
rule_variable = self.parser.parsedModel.ruleVariable[i]
if not (rule_variable in self.parser.parsedModel.speciesId):
self.out_file.write(" float %s = " % rule_variable)
else:
self.out_file.write(
" y[%s] = " %
self.parser.parsedModel.speciesId.index(rule_variable))
string = mathml_condition_parser(
self.parser.parsedModel.ruleFormula[i])
string = self.replace_names(string, 'parameter[%s]')
self.out_file.write("%s;\n" % string)
# Write the derivatives
for i in range(num_species):
species = self.parser.parsedModel.species[i]
if species.getConstant() == False and species.getBoundaryCondition(
) == False:
self.out_file.write(" float d_y%s = DT * (" % i)
if species.isSetCompartment():
self.out_file.write("(")
reaction_written = False
for k in range(self.parser.parsedModel.numReactions):
if not self.parser.parsedModel.stoichiometricMatrix[i][
k] == 0.0:
if reaction_written and self.parser.parsedModel.stoichiometricMatrix[
i][k] > 0.0:
self.out_file.write("+")
reaction_written = True
self.out_file.write(
repr(
self.parser.parsedModel.stoichiometricMatrix[i]
[k]))
self.out_file.write("*(")
string = self.parser.parsedModel.kineticLaw[k]
string = self.replace_names(string,
'tex2D(param_tex,%s,tid)')
string = p.sub('', string)
self.out_file.write(string)
self.out_file.write(")")
if species.isSetCompartment():
self.out_file.write(")/")
my_species_compartment = species.getCompartment()
for j in range(len(
self.parser.parsedModel.listOfParameter)):
parameter_id = self.parser.parsedModel.parameterId[j]
if self.parser.parsedModel.listOfParameter[j].getId(
) == my_species_compartment:
if not (parameter_id
in self.parser.parsedModel.ruleVariable):
flag = False
for r in range(
len(self.parser.parsedModel.
eventVariable)):
if parameter_id in self.parser.parsedModel.eventVariable[
r]:
flag = True
if not flag:
self.out_file.write("parameter[%s]);" % j)
break
else:
self.out_file.write("%s);" % parameter_id)
break
else:
self.out_file.write(");")
self.out_file.write("\n")
self.out_file.write("\n")
# check for columns of the stochiometry matrix with more than one entry
random_variables = ["*randNormal(rngRegs,sqrt(DT))"
] * self.parser.parsedModel.numReactions
for k in range(self.parser.parsedModel.numReactions):
num_entries = 0
for i in range(num_species):
if self.parser.parsedModel.stoichiometricMatrix[i][k] != 0.0:
num_entries += 1
# define specific randomVariable
if num_entries > 1:
self.out_file.write(
" float rand%s = randNormal(rngRegs,sqrt(DT));\n" % k)
random_variables[k] = "*rand%s" % k
self.out_file.write("\n")
# write noise terms
for i in range(num_species):
species = self.parser.parsedModel.species[i]
if species.getConstant() == False and species.getBoundaryCondition(
) == False:
self.out_file.write(" d_y%s+= (" % i)
if species.isSetCompartment():
self.out_file.write("(")
reaction_written = False
for k in range(self.parser.parsedModel.numReactions):
if not self.parser.parsedModel.stoichiometricMatrix[i][
k] == 0.0:
if reaction_written and self.parser.parsedModel.stoichiometricMatrix[
i][k] > 0.0:
self.out_file.write("+")
reaction_written = True
self.out_file.write(
repr(
self.parser.parsedModel.stoichiometricMatrix[i]
[k]))
self.out_file.write("*sqrt(")
string = self.parser.parsedModel.kineticLaw[k]
string = self.replace_names(string, 'parameter[%s]')
string = p.sub('', string)
self.out_file.write(string)
# multiply random variable
self.out_file.write(")")
self.out_file.write(random_variables[k])
if species.isSetCompartment():
self.out_file.write(")/")
my_species_compartment = species.getCompartment()
for j in range(len(
self.parser.parsedModel.listOfParameter)):
parameter_id = self.parser.parsedModel.parameterId[j]
if self.parser.parsedModel.listOfParameter[j].getId(
) == my_species_compartment:
if not (parameter_id
in self.parser.parsedModel.ruleVariable):
flag = False
for r in range(
len(self.parser.parsedModel.
eventVariable)):
if parameter_id in self.parser.parsedModel.eventVariable[
r]:
flag = True
if not flag:
self.out_file.write("parameter[%s])" % j)
break
else:
self.out_file.write("%s)" % parameter_id)
break
else:
self.out_file.write(")")
self.out_file.write(";\n")
self.out_file.write("\n")
# add terms
for i in range(num_species):
species = self.parser.parsedModel.species[i]
if species.getConstant() == False and species.getBoundaryCondition(
) == False:
self.out_file.write(" y[%s] += d_y%s;\n" % (i, i))
self.out_file.write("}\n")