Known subclasses: pingspice.analysis.diode.Diode_Setups, pingspice.analysis.mosfet.MOSFET_Setups, pingspice.analysis.thermal.Package_Setups, pingspice.examples.rlc-04.RLC_Setups, pingspice.examples.rlc-06.RLC_Setups, pingspice.test.analysis.test_param.AParameterFinder.Setups, pingspice.test.analysis.test_solver.RC_Setups, pingspice.test.lib.test_diodes.SillySetups

Include an instance of your subclass of me, with my class attributes defined, as the setups attribute to your ParameterFinder subclass.

If your subclass of me has a setup method, it will be called at the beginning of my init, right after setdefaults and construction of an EvalHelper.

The goal values for the parameter finder are in my dict Xs, one 2-D Numpy array per setup ID. If you don't want to specify all those numbers in your subclass, set the goals attribute to text, or a path of a file containing text, of goals readable by DataParser.

The class attributes analyzers, independents, Xs, are dicts, keyed by setup ID, or callables that take a Specs object as the sole argument and return a dict. Be sure to order your setups with the least CPU-intensive ones numbered lowest because evaluations returning huge SSEs can prevent ones from higher-numbered setups from even bothering to run.

The class attributes weights and transforms are dicts keyed by vector name.

Class Variable	goals	Text or a text file defining the device's performance goals. Parsed by `DataParser`.
Class Variable	specs	Text or a text file defining the device's specifications. Parsed by `SpecsParser`.
Class Variable	EvalHelper	A class with a constructor that takes a `Specs` instance as its sole argument. I automatically construct it right away if you define the class and you can refer to the instance via my eh attribute.
Class Variable	nameLists	A dict of lists of vector names to retrieve after each analysis, each keyed by setup ID.
Class Variable	plotSpecs	(Optional) A setup ID-keyed dict of plot specificiations, if parameter-fit plotting is desired. For each setup: `None` if no plot for that setup, or a list with: [<plot title>, <x-axis vector name>, <y-axis vector name(s)>]
Class Variable	analyzers	A setup ID-keyed dict of lists defining analyzers. The first item in each is a subclass of `sim.Analyzer` or an instance of `psim.Analyzer` followed by its setup args.
Class Variable	independents	A setup ID-keyed dict of lists of names of the vectors that are independent variables.
Class Variable	Xs	A setup ID-keyed dict of lists of goals that were not otherwise defined with goals. Items can be either a 2-D Numpy array of expected goal values or a callable. An array must have one column for each vector name listed in the nameLists list for that setup. A callable must accept the vectors named in the nameLists list as 1-D Numpy arrays and return a real-valued sum-of-squared error based on how closely the values in those vectors conform to some user-defined goal. If there is no Xs entry for a given setup ID (after I populate it from parsing goals, no analysis for that setup will be included.
Class Variable	weights	A dict of weights (real values or single-argument callables) to apply to the value from each vector, keyed by the names found in nameLists. A callable must take two args, the integer ID of the series being evaluated and a 1-D Numpy array of values from the vector, and return a positive real-valued weight.
Class Variable	transforms	A dict, keyed by the name of a transform-produced vector, of callables that each transform one or more vectors from the Ngspice simulation into a vector listed in nameLists but not included in the Ngspice simulation. The result of the transform (which must be a 1-D Numpy array) will be added to the `V` object under the name by which the transform is keyed in this dict.
Class Variable	logValues	A set of vector names that should be evaluated and plotted in log space rather than linear. I will log-transform any goal points in X corresponding to such vector names. It's up to whoever uses me to do the same log transformation on simulation results.
Class Variable	defaultSpecs	A dict of default values for attributes that don't appear in a given subclass's specs.
Instance Variable	netlists	A setup ID-keyed dict of netlist file paths or text. Usually populated by `ParameterFinder`.
Class Method	setdefaults	Examines my class (or subclass), or an object if supplied as a sole argument, for attributes named in my defaultSpecs dict. Those not found are set to the default values.
Method	__init__	`Setups(**kw)`
Method	__contains__	I contain a setup ID if it's for a setup I run.
Method	__len__	My length is the number of setups I run.
Method	__iter__	I iterate over my setup IDs. You can change my IDs attribute during iteration if you want.
Method	__getstate__	For pickling during wire evaluation.
Method	__setstate__	For unpickling during wire evaluation.
Method	setup	Override this if you want custom setup at the beginning of my standard setup, right after `setdefaults` gets called and an `EvalHelper` instance is constructed.
Method	_setupDicts	Called by my constructor with a `Specs` object to replace any dict-producing callable attributes with their results.
Method	_prune	Called by my constructor to get (1) a sorted copy of my class-wide IDs list and (2) a pruned version of nameLists containing just entries for the enabled IDs.
Method	_checkWeights	Called by my constructor to make sure my weights dict is legit.
Method	_finalizeXs	Called by my constructor to finalize the setting up of a dict of evaluation goals, one Numpy 2-D array or callable evaluator per setup ID.

goals =

Text or a text file defining the device's performance goals. Parsed by DataParser.

specs =

Text or a text file defining the device's specifications. Parsed by SpecsParser.

EvalHelper =

A class with a constructor that takes a Specs instance as its sole argument. I automatically construct it right away if you define the class and you can refer to the instance via my eh attribute.

nameLists =

A dict of lists of vector names to retrieve after each analysis, each keyed by setup ID.

plotSpecs =

(Optional) A setup ID-keyed dict of plot specificiations, if parameter-fit plotting is desired. For each setup: None if no plot for that setup, or a list with:

[<plot title>, <x-axis vector name>, <y-axis vector name(s)>]

analyzers =

A setup ID-keyed dict of lists defining analyzers. The first item in each is a subclass of sim.Analyzer or an instance of psim.Analyzer followed by its setup args.

independents =

A setup ID-keyed dict of lists of names of the vectors that are independent variables.

Xs =

A setup ID-keyed dict of lists of goals that were not otherwise defined with goals. Items can be either a 2-D Numpy array of expected goal values or a callable. An array must have one column for each vector name listed in the nameLists list for that setup. A callable must accept the vectors named in the nameLists list as 1-D Numpy arrays and return a real-valued sum-of-squared error based on how closely the values in those vectors conform to some user-defined goal. If there is no Xs entry for a given setup ID (after I populate it from parsing goals, no analysis for that setup will be included.

weights =

A dict of weights (real values or single-argument callables) to apply to the value from each vector, keyed by the names found in nameLists. A callable must take two args, the integer ID of the series being evaluated and a 1-D Numpy array of values from the vector, and return a positive real-valued weight.

transforms =

A dict, keyed by the name of a transform-produced vector, of callables that each transform one or more vectors from the Ngspice simulation into a vector listed in nameLists but not included in the Ngspice simulation. The result of the transform (which must be a 1-D Numpy array) will be added to the V object under the name by which the transform is keyed in this dict.

logValues =

A set of vector names that should be evaluated and plotted in log space rather than linear. I will log-transform any goal points in X corresponding to such vector names. It's up to whoever uses me to do the same log transformation on simulation results.

defaultSpecs =

A dict of default values for attributes that don't appear in a given subclass's specs.

netlists =

A setup ID-keyed dict of netlist file paths or text. Usually populated by ParameterFinder.

@classmethod
def setdefaults(cls, *args):

Examines my class (or subclass), or an object if supplied as a sole argument, for attributes named in my defaultSpecs dict. Those not found are set to the default values.

def __init__(self, **kw):

Setups(**kw)

You can set an instance's attributes with constructor keywords.

def __contains__(self, ID):

I contain a setup ID if it's for a setup I run.

def __len__(self):

My length is the number of setups I run.

def __iter__(self):

I iterate over my setup IDs. You can change my IDs attribute during iteration if you want.

def __getstate__(self):

For pickling during wire evaluation.

def __setstate__(self, state):

For unpickling during wire evaluation.

def setup(self, s):

overridden in pingspice.analysis.thermal.Package_Setups

Override this if you want custom setup at the beginning of my standard setup, right after setdefaults gets called and an EvalHelper instance is constructed.

def _setupDicts(self, s):

Called by my constructor with a Specs object to replace any dict-producing callable attributes with their results.

The nameLists dict/callable gets done first, with a reference to it included in s so that each setup's vector names are available to the others.

The Xs dict/callable gets done second, with my goals string/file parsed and a reference to it included in s so that its values are available to the others.

def _prune(self):

Called by my constructor to get (1) a sorted copy of my class-wide IDs list and (2) a pruned version of nameLists containing just entries for the enabled IDs.

If my IDs list is empty, (1) contains all IDs of all keys in nameLists, still sorted, and (2) contains entries for all IDs.

Except: Any ID appearing in my skippedIDs list is always omitted.

def _checkWeights(self):

Called by my constructor to make sure my weights dict is legit.

def _finalizeXs(self):

Called by my constructor to finalize the setting up of a dict of evaluation goals, one Numpy 2-D array or callable evaluator per setup ID.

Removes IDs of setups with no goals defined.
Sorts independent variables in ascending order so that sse.ErrorTabulator.closenessToIndepInterpolated may be used.
Log-transforms any log-space goal values (as indicated by presence of their vector name in my logValues set) into linear space.

API Documentation for pingspice, generated by pydoctor at 2021-09-18 08:41:11.

pingspice.analysis.param.Setups(object) class documentation

`pingspice.analysis.param.Setups(object)` class documentation