Learning Forum

[lipee36]

I have a question about co-simulating python with Interconnect. I've ran all the examples successfully, and I was able to reproduce simple scenarios based on them.

My question is regarding the capabilities of the co-simulation itself. What I am looking to do is to create a circuit on interconnect and execute it on a loop, using dynamic inputs generated by my python script. A very dummy example of this functionality:

[Guanhui Wang]

Hi lipee36
Thank you for the post. I believe what you have described is possible using the Python API. Please check out this example here: https://support.lumerical.com/hc/en-us/articles/360034936773-Python-co-simulation-with-INTERCONNECT

[lipee36]

thank you very much for your suggestion. Indeed, I've tried to replicate this example. If I run the example, it returns an error (something like "co_simulation_example.icp could not be read" - although I can open the project and executes it from the interface, or using the .lsf script.
Anyhow, I've tried to adapt it to a very simple example, and yet I have errors:
in lumapi.evalScript(h,'''File "F:\Programs\Lumerical\api\python\lumapi.py", line 251, in evalScriptraise LumApiError("Failed to evaluate code")lumapi.LumApiError: 'Failed to evaluate code'
This is the simple circuit I am trying to use as example:
and this is the adapted script:
import sys, os, random, pdb
import numpy as np
os.add_dll_directory('F:\\Programs\\Lumerical\\api\\python')

sys.path.append("F:\\Programs\\Lumerical\\api\\python")
CWD = os.path.dirname(os.path.abspath(__file__)) ## No need to run this line if using python console
os.chdir("F:\\Programs\\Lumerical\\api\\python")

import lumapi
h = lumapi.open("interconnect", hide=True)

#####################################################
# Define user parameters
#
time_window = 5e-9
Nsamples = 8192
power_in_1 = 100
power_in_2 = 0
frequency_in = 1550
phase_in = 0
# end user parameters
#####################################################

# we must load the file after starting INTERCONNECT
lumapi.evalScript(h,''' load("MZI.icp"); ''');

# intialize simply by pushing data to Lumerical script workspace
lumapi.putDouble(h,"time_window",time_window);
lumapi.putDouble(h,"Nsamples",Nsamples);
lumapi.putDouble(h,"power_in_1",power_in_1);
lumapi.putDouble(h,"power_in_2",power_in_2);
lumapi.putDouble(h,"frequency_in",frequency_in);
lumapi.putDouble(h,"phase_in",phase_in);
lumapi.evalScript(h,'''
# turn off history and redrawing
redrawoff;
historyoff;

# time vectors
t = linspace(0,time_window,Nsamples+1);
t = t(1:end-1);
dt = t(2)-t(1);

# restore design mode in case we are in analysis mode
switchtodesign;

setnamed("::Root Element","time window",time_window);
setnamed("::Root Element","number of samples",Nsamples);

setnamed("CWL_1","power",power_in_1);
setnamed("CWL_1","frequency",frequency_in);

setnamed("CWL_2","power",power_in_2);
setnamed("CWL_2","frequency",frequency_in);

setnamed("MZI_1","phase_shift",phase_in);

# intialize the simulation
runinitialize;

# initialize struct used for passing co-sim data
v = struct;

# End initialization
#####################################################

''')

output = np.empty((314,2))

counter = 0
while [phase_in < 3.15]:
lumapi.evalScript(h,'''

#set parameters
setvalue("CWL_1","power",power_in_1);
setvalue("CWL_1","frequency",frequency_in);
setvalue("CWL_2","power",power_in_2);
setvalue("CWL_2","frequency",frequency_in);

setvalue("MZI_1","phase_shift",phase_in);

# run INTERCONNECT step
runstep;

# read signals from INTERCONNECT simulation
v.index = 1;
v1 = getvalue("COSIM_1","port",v);
v.index = 2;
v2 = getvalue("COSIM_1","port",v);
''');
output[counter][0] = lumapi.getVar(h,"v1")
output[counter][1] = lumapi.getVar(h,"v2")

phase_in = phase_in + 0.01

# run finalize of INTERCONNECT
lumapi.evalScript(h,"runfinalize;")

print(output)
lumapi.close(h)

The error I've mentioned before is due the line I've highlighted bold . Can you see any obvious mistake/problem?
Thanks again!

[liqiande]

õ¢áÕÑ¢´╝îµêæþø«Õë쵡úÕ£¿ÕüÜÞ┐ÖþºìÕÀÑõ¢£´╝îµêæÞâ¢Úù«õ¢áÕÉù´╝ƒ