from typing import Any, Dict, List, Tuple
import cmath as cm
import time
import re
import itertools
import inspect
from copy import deepcopy
import numpy as np
from numpy import ndarray
from pathlib import PosixPath
from defusedxml.ElementTree import parse
[docs]def fromSI(value: str) -> float:
"""converts from SI unit values to metric
Args:
value (str): a value in SI units, e.g. 1.3u
Returns:
float: the value in metric units.
"""
return float(value.replace("u", "e-6"))
[docs]def universal_sparam_filereader(
nports: int, sfilename: str, sfiledir: PosixPath, format_type: str = "auto"
) -> Tuple[ndarray, ndarray]:
"""
Function to automatically detect the sparameter file format and use appropriate method to delimit and format sparam data
This function is a unified version of sparameter reader function defined in https://github.com/BYUCamachoLab/simphony
Args:
nports: Number of ports
sfilename: XML look-up-table filename
sfiledir: Path to the directory containing the XML file
format_type: Format type. For more information: https://support.lumerical.com/hc/en-us/articles/360036618513-S-parameter-file-formats
"""
numports = nports
filename = sfiledir / sfilename
if format_type == "auto":
try:
# print("try A")
result = universal_sparam_filereader(nports, sfilename, sfiledir, "A")
return result
except Exception:
try:
# print("try B")
result = universal_sparam_filereader(nports, sfilename, sfiledir, "B")
return result
except Exception:
# print("try C")
result = universal_sparam_filereader(nports, sfilename, sfiledir, "C")
return result
elif format_type == "A":
"""
dc_halfring_te_1550
Returns the s-parameters across some frequency range for the Sparam fileformat A
input:
["port 1",""]
["port 2",""]
["port 3",""]
["port 4",""]
("port 1","mode 1",1,"port 1",1,"transmission")
(101, 3)
output:
[frequency, s-parameters]
"""
F = []
S = []
with open(filename, "r") as fid:
for i in range(5):
line = fid.readline()
line = fid.readline()
numrows = int(tuple(line[1:-2].split(","))[0])
S = np.zeros((numrows, numports, numports), dtype="complex128")
r = m = n = 0
for line in fid:
if line[0] == "(":
continue
data = line.split()
data = list(map(float, data))
if m == 0 and n == 0:
F.append(data[0])
S[r, m, n] = data[1] * np.exp(1j * data[2])
r += 1
if r == numrows:
r = 0
m += 1
if m == numports:
m = 0
n += 1
if n == numports:
break
return (np.array(F), S)
elif format_type == "B":
"""
ebeam_bdc_te1550, nanotaper, ebeam_y_1550
Returns the s-parameters across some frequency range for the Sparam fileformat A
input:
('port 1','TE',1,'port 1',1,'transmission')
(51,3)
output:
[frequency, s-parameters]
"""
F = []
S = []
with open(filename, "r") as fid:
line = fid.readline()
line = fid.readline()
numrows = int(tuple(line[1:-2].split(","))[0])
S = np.zeros((numrows, numports, numports), dtype="complex128")
r = m = n = 0
for line in fid:
if line[0] == "(":
continue
data = line.split()
data = list(map(float, data))
if m == 0 and n == 0:
F.append(data[0])
S[r, m, n] = data[1] * np.exp(1j * data[2])
r += 1
if r == numrows:
r = 0
m += 1
if m == numports:
m = 0
n += 1
if n == numports:
break
return (np.array(F), S)
elif format_type == "C":
"""
ebeam_gc_te1550
Returns the s-parameters across some frequency range for the Sparam fileformat A
input:
columns with space delimiter
output:
[frequency, s-parameters]
"""
with open(filename) as fid:
# grating coupler compact models have 100 points for each s-matrix index
arrlen = 100
lines = fid.readlines()
F = np.zeros(arrlen)
S = np.zeros((arrlen, 2, 2), "complex128")
for i in range(0, arrlen):
words = lines[i].split()
F[i] = float(words[0])
S[i, 0, 0] = cm.rect(float(words[1]), float(words[2]))
S[i, 0, 1] = cm.rect(float(words[3]), float(words[4]))
S[i, 1, 0] = cm.rect(float(words[5]), float(words[6]))
S[i, 1, 1] = cm.rect(float(words[7]), float(words[8]))
F = F[::-1]
S = S[::-1, :, :]
return (np.array(F), S)
[docs]def LUT_reader(filedir: PosixPath, lutfilename: str, lutdata: List[List[str]]):
"""
Reads look up table data.
Args:
filedir: Directory of the XML look-up-table file.
lutfilename: Look-up-table filename.
lutdata: Look-up-table arguments.
"""
xml = parse(filedir / lutfilename)
root = xml.getroot()
for node in root.iter("association"):
sample = [[each.attrib["name"], each.text] for each in node.iter("value")]
if sorted(sample[0:-1]) == sorted(lutdata):
break
sparam_file = sample[-1][1].split(";")
return (sparam_file, xml, node)
[docs]def LUT_processor(
filedir: PosixPath,
lutfilename: str,
lutdata: List[List[str]],
nports: int,
sparam_attr: str,
verbose: bool = False,
) -> Tuple[Tuple[ndarray, ndarray], str]:
"""process look up table data"""
start = time.time()
sparam_file, xml, node = LUT_reader(filedir, lutfilename, lutdata)
# read data
if ".npz" in sparam_file[0] or ".npz" in sparam_file[-1]:
npzfile = [each for each in sparam_file if ".npz" in each][0]
tempdata = np.load(filedir / npzfile)
sdata = (tempdata["f"], tempdata["s"])
npz_file = npzfile
else:
if verbose:
print("numpy datafile not found. reading sparam file instead..")
sdata = universal_sparam_filereader(nports, sparam_file[-1], filedir, "auto")
# create npz file name
npz_file = sparam_file[-1].split(".")[0]
# save as npz file
np.savez(filedir / npz_file, f=sdata[0], s=sdata[1])
# update xml file
sparam_file.append(npz_file + ".npz")
sparam_file = list(set(sparam_file))
for each in node.iter("value"):
if each.attrib["name"] == sparam_attr:
each.text = ";".join(sparam_file)
xml.write(filedir / lutfilename)
if verbose:
print("SParam data extracted in ", time.time() - start)
return (sdata, npz_file)
[docs]def NetlistProcessor(spice_filepath, Network, libraries, c_, circuitData, verbose=True):
"""
Processes a spice netlist to setup and simulate a circuit.
Args:
spice_filepath: Path to the spice netlist file.
Network:
"""
if verbose:
for key, value in circuitData.items():
print(key, str(value))
# define frequency range and resolution
freq = np.linspace(
c_ / circuitData["sim_params"][0],
c_ / circuitData["sim_params"][1],
circuitData["sim_params"][2],
)
# create a circuit
subckt = Network(network_id=circuitData["networkID"], f=freq)
# get library
all_libraries = dict(
[
each
for each in inspect.getmembers(libraries, inspect.ismodule)
if each[0][0] != "_"
]
)
libs_comps = {}
for each_lib in list(set(circuitData["compLibs"])):
# temp_comps = dict(inspect.getmembers(all_libraries[each_lib], inspect.isclass))
libs_comps[each_lib] = all_libraries[each_lib].component_factory
# add circuit components
for i in range(len(circuitData["compModels"])):
# get component model
comp_model = libs_comps[circuitData["compLibs"][i]][
circuitData["compModels"][i]
]
# clean attributes
cls_attrs = deepcopy(comp_model.cls_attrs) # class attributes
comp_attrs = circuitData["compAttrs"][i] # component attributes
# clean up attributes
for each_attrs in cls_attrs.keys():
if each_attrs in comp_attrs.keys():
cls_attrs[each_attrs] = fromSI(comp_attrs[each_attrs])
subckt.add_component(
libs_comps[circuitData["compLibs"][i]][circuitData["compModels"][i]],
params=cls_attrs,
component_id=circuitData["compLabels"][i],
)
# add circuit netlist
subckt.global_netlist = circuitData["circuitNets"]
# add unique net component connections
subckt.current_connections = circuitData["circuitConns"]
return subckt
[docs]class netlistParser:
"A netlist parser to read spi files generated by SiEPIC tools"
def __init__(self, mainfile_path: PosixPath) -> None:
self.circuitComponents = []
self.circuitConnections = []
self.mainfile_path = mainfile_path
[docs] def readfile(self) -> Dict[str, Any]:
filepath = self.mainfile_path
circuitID = ""
inp = ""
out = ""
inp_net = 0
out_net = []
circuitLabels = []
circuitModels = []
circuitConns = []
circuitNets = []
componentLibs = []
componentAttrs = []
component_locations = []
temp_file = open(filepath, "r")
temp_lines = temp_file.readlines()
free_node_idx = -1
freq_data = []
seek_component = 0
seek_ona = 0
orthogonal_ID = 0
# extract circuit connectivity
for each_line in temp_lines:
each_line = re.sub(" +", " ", each_line.strip()) # remove empty lines
if each_line.startswith("*"):
continue
else:
each_line = "".join(
[
"".join(filter(None, each_section.split(" ")))
if ('"' in each_section)
else each_section
for each_section in re.split(
r"""("[^"]*"|'[^']*')""", each_line
)
]
)
temp_data = each_line.split(" ")
if len(temp_data) > 1: # if line is not an empty one
MC_location = []
if temp_data[0] == ".subckt":
circuitID = temp_data[1]
inp = temp_data[2]
out = [temp_data[x] for x in range(3, len(temp_data))]
seek_component = 1
elif temp_data[0] == ".param":
continue
elif temp_data[0] == ".ends":
seek_component = 0
elif temp_data[0] == ".ona":
seek_ona = 1
elif seek_ona == 1:
# ONA related data
if len(temp_data) < 3:
temp_data = [0] + temp_data[-1].split("=")
if temp_data[1] == "orthogonal_identifier":
orthogonal_ID = int(temp_data[-1])
elif temp_data[1] == "start":
freq_data.append(float(temp_data[-1]))
elif temp_data[1] == "stop":
freq_data.append(float(temp_data[-1]))
elif temp_data[1] == "number_of_points":
freq_data.append(int(temp_data[-1]))
elif seek_component == 1:
# otherwise its component data
circuitLabels.append(temp_data[0])
temp_ports = []
found_ports = 0
found_library = 0
for i in range(1, len(temp_data)):
# if its an optical port
if (
"N$" in temp_data[i]
and "N$None".lower() != temp_data[i].lower()
):
temp_ports.append(int(temp_data[i].replace("N$", "")))
found_ports = 1
elif "N$None".lower() == temp_data[i].lower():
temp_ports.append(free_node_idx)
free_node_idx -= 1
found_ports = 1
elif inp == temp_data[i]:
temp_ports.append(free_node_idx)
inp_net = free_node_idx
free_node_idx -= 1
found_ports = 1
elif out[0] == temp_data[i]:
temp_ports.append(free_node_idx)
out_net.append(free_node_idx)
free_node_idx -= 1
if len(out) > 1:
out.pop(0)
if len(out) == 0:
found_ports = 1
elif found_ports == 1 and "N$" not in temp_data[i]:
circuitModels.append(temp_data[i])
temp_cls_atrr = (
{}
) # deepcopy(lib[temp_data[i]].cls_attrs)
found_ports = -1
elif "lay" in temp_data[i] or "sch" in temp_data[i]:
if "lay" in temp_data[i]:
MC_location.append(
fromSI(temp_data[i].split("=")[-1]) * 1e6
)
# ignore layout and schematic position data for now.
# adapt opics models to accept this data
# they are component parameters
elif "library" in temp_data[i]:
# cprint(temp_data[i])
temp_lib = (
temp_data[i].replace('"', "").split("=")[1].split()
)
componentLibs.append(
temp_lib[-1].split("/")[-1].lower()
)
found_library = 1
elif "=" in temp_data[i] and found_library == 1:
# if its a components' attribute
temp_attr = temp_data[i].split("=")
# print(temp_attr[0])
# if(temp_attr[0] in temp_cls_atrr):
temp_cls_atrr[temp_attr[0]] = temp_attr[1].strip('"')
componentAttrs.append(temp_cls_atrr)
circuitNets.append(temp_ports)
if bool(MC_location):
component_locations.append(MC_location)
circuitConns = list(set(list(itertools.chain(*circuitNets))))
# remove IOs from component connections' list
circuitConns = [each for each in circuitConns if each >= 0]
# return all data
return {
"circuitNets": circuitNets,
"circuitConns": circuitConns,
"compLibs": componentLibs,
"compModels": circuitModels,
"compLabels": circuitLabels,
"compAttrs": componentAttrs,
"compLocs": component_locations,
"networkID": circuitID,
"inp_net": inp_net,
"out_net": out_net,
"sim_params": freq_data,
"OID": orthogonal_ID,
}
