Source code for finesse.model

"""A sub-module containing the configuration container class :class:`.Model` which is
used for building and manipulating interferometer systems."""

from __future__ import annotations

from collections import OrderedDict, defaultdict
from copy import copy, deepcopy
from functools import reduce
import inspect
import math
from numbers import Number
import weakref
import logging
from contextlib import contextmanager
from functools import wraps
from typing import Any
import warnings
from pathlib import Path

import numpy as np
import networkx as nx
from deprecated import deprecated

from .config import config_instance
from .components import (
    Port,
    NodeDirection,
    Connector,
    Space,
    Surface,
    FrequencyGenerator,
    Cavity,
    Wire,
    Gauss,
)
from .components.general import InteractionType, Variable
from .components.readout import _Readout, ReadoutDetectorOutput
from .components.dof import DegreeOfFreedom
from .components.general import CouplingType
from .components.node import NodeType, OpticalNode, Node, SignalNode
from .components.trace_dependency import TraceDependency
from . import detectors
from .element import ModelElement
from .env import warn, has_pygraphviz
from .exceptions import (
    FinesseException,
    NodeException,
    ComponentNotConnected,
    BeamTraceException,
    ModelMissingAttributeError,
    ModelClassAttributeError,
    ContextualValueError,
)
from .freeze import Freezable
from .frequency import Fsig
from .gaussian import BeamParam, transform_beam_param
from .locks import Lock
from .paths import OpticalPath
from .tree import TreeNode
from .solutions import BaseSolution, BeamTraceSolution, SeriesSolution
from .tracing.forest import TraceForest
from .tracing import tools as tracetools
from .utilities import valid_name, pairwise, ngettext, is_iterable
from .utilities.components import refractive_index
from .utilities.homs import make_modes, insert_modes, remove_modes
from .utilities.tables import NumberTable
from finesse.utilities.network_filter import NetworkType
from .warnings import ModelParameterSettingWarning, CavityUnstableWarning
from .simulations import BaseSimulation
from .simulations.base import ModelSettings
from .parameter import Parameter
from .plotting.graph import plot_format
from finesse.utilities.collections import OrderedSet


[docs]def locked_when_built(func): @wraps(func) def wrapper(self, *args, **kwargs): if self.is_built: raise Exception( f"Model has been built for a simulation, cannot use {func} here" ) return func(self, *args, **kwargs) return wrapper
LOGGER = logging.getLogger(__name__)
[docs]class IOMatrix:
[docs] def __init__(self, model): self._model = model self._terms = {}
[docs] def clear(self): self._terms.clear()
def _check_key(a, b): raise NotImplementedError() def __getitem__(self, key): if type(key) is tuple: if key[0] in self._terms: return self._terms[key[0]][key[1]] else: raise KeyError(f"{key[0]} not in Matrix") else: if key in self._terms: return self._terms[key] else: raise KeyError(f"{key} not in Matrix") def __setitem__(self, key, value): if len(key) != 2: raise Exception("Expected pair of string names as key") a, b = key if key[0] not in self._terms: self._terms[key[0]] = {} self._terms[key[0]][key[1]] = value def __bool__(self): return bool(self._terms)
[docs]class OutputMatrix(IOMatrix): def _check_key(self, a, b): if not isinstance(a, _Readout): raise Exception(f"First index should be a Readout element name not `{a}`") if not isinstance(b, DegreeOfFreedom): raise Exception( f"Second index should be a DegreeOfFreedom element name not `{a}`" )
[docs]class InputMatrix(IOMatrix): def _check_key(self, a, b): if not isinstance(a, DegreeOfFreedom): raise Exception( f"First index should be a DegreeOfFreedom element name not `{a}`" ) if not isinstance(b, _Readout): raise Exception(f"Second index should be a Readout element name not `{a}`")
[docs]class Event(list): """Event subscription. A list of callable objects. Calling an instance of this will cause a call to each item in the list in ascending order by index. Example Usage: >>> def f(x): ... print 'f(%s)' % x >>> def g(x): ... print 'g(%s)' % x >>> e = Event() >>> e() >>> e.append(f) >>> e(123) f(123) >>> e.remove(f) >>> e() >>> e += (f, g) >>> e(10) f(10) g(10) >>> del e[0] >>> e(2) g(2) Notes ----- Code from https://stackoverflow.com/questions/1092531/event-system-in-python """ def __call__(self, *args, **kwargs): for f in self: f(*args, **kwargs) def __repr__(self): return "<Event(%s)>" % list.__repr__(self)
[docs]def make_optical_network_view(model): """From a given model return a view of the full network that just contains the optical nodes and edges.""" return nx.subgraph_view( model.network, filter_node=lambda n: model.network.nodes[n]["type"] == NodeType.OPTICAL, filter_edge=lambda i, o: model.network[i][o]["coupling_type"] == CouplingType.OPTICAL_TO_OPTICAL, )
[docs]class Model(Freezable): """Optical configuration class for handling models of interferometers. This class stores the interferometer configuration as a directed graph and contains methods to interface with this data structure. """
[docs] def __init__(self): # The model graph. self.__network = nx.DiGraph() # The optical graph. We need to store optical network separately for tracing # correct paths through optical-optical couplings only. self.__opt_net_view = make_optical_network_view(self) # The syntax used to build the model, if parsed from KatScript. self.syntax_graph = None self.__is_built = False # Some simple events that occur within the model that elements can register # themselves with. self._on_pre_build = Event() # Called just before a simulation is built. self._on_unbuild = Event() # Called after unbuild() has been completed. # Components and detectors. self.__cavities = OrderedDict() self.__components = OrderedDict() self.__detectors = OrderedDict() self.__locks = OrderedDict() self.__dofs = OrderedDict() self.__lock_model_change = False self.__not_removable = OrderedSet() # Frequency storage. self.__frequencies = OrderedSet() self.__frequency_change_callbacks = [] self.__freq_map = None self._frequency_generators = [] # internal dict to store mode() call args for unparsing self.__mode_setting = {} self.alternate_name_map = {} # Storage class for all the model settings that affects the output of a # simulation. self._settings = ModelSettings() # Higher order mode related attributes. self._settings.homs = np.zeros(1, dtype=(np.intc, 2)) # [(0, 0)] self._settings.is_modal = False self._settings.phase_config.zero_k00 = False self._settings.phase_config.zero_tem00_gouy = True # whether to use the Finesse2 phase conventions on transmission self._settings.phase_config.v2_transmission_phase = False # Constants. config_consts = config_instance()["constants"] self._settings.set_lambda0(config_consts.getfloat("lambda0")) # Ratio of epsilon_0/c, used in converting between power and optical fields. # Typically use just renormalise and use 1, putting optical fields in units of # sqrt{W}. self._settings.EPSILON0_C = 1 self._settings.UNIT_VACUUM = 1 self._settings.x_scale = config_consts.getfloat("x_scale") # Beam tracing attributes. self.__gausses = {} self.__last_trace = None self.__trace_order = [] # Order of TraceDependency objects for beam tracing # Get argument names and values from beam_trace method and set the default # simulation tracing config dict with these self.__default_sim_trace_config = { arg: value.default for arg, value in inspect.signature(self.beam_trace).parameters.items() } self.__default_sim_trace_config["retrace"] = True self.__default_sim_trace_config["unstable_handling"] = "auto" self.__sim_trace_config = self.__default_sim_trace_config.copy() self.__trace_forest = TraceForest(self, self.sim_trace_config["symmetric"]) self._rebuild_trace_forest = True # The root action. self.__analysis = None self.__inmtx_dc = InputMatrix(self) self.yaxis = None self.force_refill = False # List of all model elements. self.__elements = OrderedDict() self.__model_parameters = OrderedDict() # Models always have a signal element. self.add(Fsig("fsig", None), unremovable=True) self._freeze()
def __setattr__(self, name, value): # Need this method to stop overwriting model elements by accident, or # on purpose obj = self.__dict__.get(name) if obj is None: super().__setattr__(name, value) elif isinstance(obj, Parameter): obj.value = value elif isinstance(obj, ModelElement): raise AttributeError(f"Cannot overwrite element `{repr(obj)}` in model") else: super().__setattr__(name, value)
[docs] def info( self, modes=True, components=True, detectors=True, cavities=True, locks=True, ): """Get string containing information about this model. Parameters ---------- modes, components, detectors, cavities, locks : bool, optional Show model component information. Returns ------- str The model information. """ from .utilities import format_section, format_bullet_list pieces = [] if modes: pieces.append( format_section( ngettext(len(self.homs), "{n} optical mode", "{n} optical modes"), format_bullet_list(self.homs), ) ) if components: pieces.append( format_section( ngettext(len(self.components), "{n} component", "{n} components"), format_bullet_list(self.components), ) ) if detectors: pieces.append( format_section( ngettext(len(self.detectors), "{n} detector", "{n} detectors"), format_bullet_list(self.detectors), ) ) if cavities: pieces.append( format_section( ngettext(len(self.cavities), "{n} cavity", "{n} cavities"), format_bullet_list(self.cavities), ) ) if locks: pieces.append( format_section( ngettext(len(self.locks), "{n} locking loop", "{n} locking loops"), format_bullet_list(self.locks), ) ) text = str(self) + "\n\n" + "\n".join(pieces) return text
[docs] def deepcopy(self): return deepcopy(self)
def __deepcopy__(self, memo): new = object.__new__(type(self)) memo[id(self)] = new new.after_deepcopy = [] # fields to exclude from copying straight away # -> see below for cavities # -> TraceForest instances are non-copyable by design so # new model must rebuild its trace_forest by itself exclude = [ "_Model__cavities", "_Model__trace_forest", ] sdict = {} for k, v in self.__dict__.items(): if k in exclude: sdict[k] = None else: sdict[k] = v # need to ensure cavities are copied after everything else otherwise # the copy operations on the source and target nodes may result in # key errors on the memo as the new nodes haven't been created yet cdict = {"_Model__cavities": self.__dict__["_Model__cavities"]} new.__dict__.update(deepcopy(sdict, memo)) new.__dict__.update(deepcopy(cdict, memo)) # update all the weakrefs we have in the network for n in new.network.nodes: new.network.nodes[n]["weakref"] = weakref.ref( memo[id(self.network.nodes[n]["weakref"]())] ) new.network.nodes[n]["owner"] = weakref.ref( memo[id(self.network.nodes[n]["owner"]())] ) for e in new.network.edges: new.network.edges[e]["in_ref"] = weakref.ref( memo[id(self.network.edges[e]["in_ref"]())] ) new.network.edges[e]["out_ref"] = weakref.ref( memo[id(self.network.edges[e]["out_ref"]())] ) new.network.edges[e]["owner"] = weakref.ref( memo[id(self.network.edges[e]["owner"]())] ) new.__opt_net_view = make_optical_network_view(new) new.__trace_forest = TraceForest( new, self.sim_trace_config["symmetric"], self.trace_forest.forest.trees ) new._rebuild_trace_forest = True new.__last_trace = None if self.is_built: # If we deepcopy from a built state then make sure we clean up the model # state to get rid of any old data stored new.unbuild() # Final run of any later updates for _ in new.after_deepcopy: _() del new.after_deepcopy # cleanup as this shouldn't be kept return new
[docs] def add_parameter( self, name, value, *, description=None, dtype=float, units="", is_geometric=False, changeable_during_simulation=True, ): """Adds a new parameter to the Model. This can be used to add a custom parameter to the model that are not defined by elements themselves. Parameters ---------- name : str Name of the parameter value : numeric | symbolic The initial value of the parameter, this can be a simple numeric value or some FINESSE symbolic desription : str, optional A descritive text of what this parameter is dtype : str, optional The datatype, typically float, int, bool units : str, optional Units of the parameter for display purposes is_geometric : bool, optional Whether this parameter is being used to recompute what the ABCD state of the model is. This should only be true if this is directly used to set what the ABCD matrices are. If you are just using a reference of this for another geometric parameter this does not have to be True. changeable_during_simulation : bool, optional Whether this parameter is allowed to be changed during the simulation add_to_model_namespace : bool, optional Whether to add this to the main model namespace or not Examples -------- Here we add some new parameters: >>> import finesse >>> model = finesse.Model() >>> A = model.add_parameter('A', 1, units='W') >>> B = model.add_parameter('B', A.ref +1) These act like normal element parameters, so references can be made to make symbolic equations and links between elements. These will be unparsed as variables in KatScript: >>> print(model.unparse()) variable A value=1.0 units='W' variable B value=(A+1) """ if name in self.__model_parameters: raise FinesseException(f"A parameter with the name `{name}` already exists") if hasattr(self, name): raise FinesseException( f"The `{name}` is already used by `{getattr(self, name)}` in this model, choose a different one" ) from finesse.parameter import Parameter, ParameterInfo pinfo = ParameterInfo( name, description, dtype, dtype, units, is_geometric, changeable_during_simulation, ) p = self.__model_parameters[name] = Parameter(pinfo, self) p.value = value # Store in the global model namespace for access self._unfreeze() setattr(self, name, p) self._freeze() return p
@property def parameters(self): "Returns any parameters added to this model, see also :attr:`.Model.all_parameters`." return tuple(self.__model_parameters.values()) @property def all_parameters(self): """Returns a generator of all the parameters in this model and the elements added to this model.""" from itertools import chain return chain( (p for el in self.elements.values() for p in el.parameters), self.__model_parameters.values(), )
[docs] def sort_elements(self, key): """Sort the display order of the elements in the model. This order is used for determining the order of plot traces and other listings. Element sorting is useful for example when parsing KatScript into a model, where adding of elements to the model may not be performed in the same order as the corresponding definitions in the script. To ensure consistency to the user, this method can be used to sort the parsed elements back into their original script order. Notes ----- The sort performed by this method is stable. Parameters ---------- key : callable Specifies a function that takes a single argument - a tuple containing the element name and object - and returns a comparison key. """ LOGGER.debug("sorting model elements (redefining __elements dict)") self.__elements = dict(sorted(self.__elements.items(), key=key))
[docs] def element_order(self, element): """Get the order in which `element` was added to the model.""" return list(self.elements).index(element.name)
@property def input_matrix_dc(self): """The DC input matrix is used to relate degrees of freedoms and readouts within a model. This information is used to generate DC locks to put the model at an operating point defined by the error signals used. :`getter`: Returns an InputMatrix object. """ return self.__inmtx_dc @property def analysis(self): """The root action to apply to the model when :meth:`Model.run` is called. :`getter`: Returns the root analysis attached to this model. :`setter`: Sets the model's root analysis. """ return self.__analysis @analysis.setter @locked_when_built def analysis(self, action): self.__analysis = action @property def readouts(self): """Returns all readouts in the model.""" return self.get_elements_of_type(_Readout) @property def elements(self): """Dictionary of all the model elements with the keys as their names.""" return self.__elements.copy() @property def network(self): """The directed graph object containing the optical configuration as a :class:`networkx.DiGraph` instance. The `network` stores :class:`.Node` instances as nodes and :class:`.Space` instances as edges, where the former has access to its associated component via :attr:`.Node.component`. See the NetworkX documentation for further details and a reference to the data structures and algorithms within this module. :`getter`: Returns the directed graph object containing the configuration (read-only). """ return self.__network @property def optical_network(self): """A read-only view of the directed graph object stored by :attr:`Model.network` but only containing nodes of type :class:`.OpticalNode` and only with edges that have couplings to optical nodes. :`getter`: Returns the optical-only directed graph-view (read-only). """ assert self.__opt_net_view._graph is self.network return self.__opt_net_view def filter_node(node): return self.network.nodes[node]["type"] == NodeType.OPTICAL return nx.subgraph_view(self.network, filter_node=filter_node)
[docs] def to_component_network(self, add_edge_info: bool = False): """Generate an undirected graph containing components as the nodes of the graph and connections (spaces, wires) between component nodes as the edges of the graph. Returns ------- :class:`networkx.Graph` The component network. """ c_net = nx.Graph() # Add components as nodes. for component in self.components: c_net.add_node(component.name) # Add spaces, wires as edges. for u, v, data in self.network.edges(data=True): # Resolve the owning object. owner = data["owner"]() if not isinstance(owner, (Space, Wire)): continue # Resolve the connected objects. in_ref = data["in_ref"]() out_ref = data["out_ref"]() u = in_ref.component.name v = out_ref.component.name if u not in c_net.nodes or v not in c_net.nodes: # Only add edges between components (not e.g. connections between components and # spaces, like phase couplings). continue if add_edge_info: # add info on connections between nodes for verbose component tree out_name = f"{out_ref.component.name}.{out_ref.port.name}" in_name = f"{in_ref.component.name}.{in_ref.port.name}" c_net.nodes[u][v] = f"{out_name}{in_name}" c_net.nodes[v][u] = f"{in_name}{out_name}" # Since the network is a graph, not a digraph, we only need to add one direction. if (u, v) not in c_net.edges: c_net.add_edge(u, v, connection=data["owner"]) return c_net
def __nodes_of(self, *node_types) -> list[Node]: return [ attr["weakref"]() for _, attr in self.__network.nodes(data=True) if attr["weakref"]().type in node_types ] @property def optical_nodes(self) -> list[Node]: """The optical nodes stored in the model. :`getter`: Returns a list of all the optical nodes in the model (read-only). """ return self.__nodes_of(NodeType.OPTICAL) @property def signal_nodes(self) -> list[Node]: """The signal nodes stored in the model. :`getter`: Returns a list of all the signal nodes in the model (read-only). """ return self.__nodes_of(NodeType.ELECTRICAL, NodeType.MECHANICAL) @property def gausses(self): """A dictionary of optical node to :class:`.Gauss` instance mappings. :`getter`: Returns the dictionary of user defined beam parameter nodes (read-only). """ return self.__gausses @property def lambda0(self): """The default wavelength to use for the model. :`getter`: Returns wavelength in meters :`setter`: Sets the wavelength in meters """ return self._settings.lambda0 @lambda0.setter @locked_when_built def lambda0(self, value): self._settings.set_lambda0(float(value)) # Update the wavelengths of Gauss object beam parameters for gauss in self.__gausses.values(): gauss.qx.wavelength = self._settings.lambda0 gauss.qy.wavelength = self._settings.lambda0 @property def f0(self): """The default frequency to use for the model. This is determinde by the value of `lambda0`. :`getter`: Returns frequency in Hertz """ return self._settings.f0 @property def k0(self): """The default wavenumber used for the model. This is determinde by the value of `lambda0`. :`getter`: Returns frequency in Hertz """ return self._settings.k0 @property def is_modal(self): """Flag indicating whether the model is modal or plane-wave. :`getter`: `True` if the modal is modal, `False` if it is plane-wave. """ return self._settings.is_modal @property def hom_labels(self): """Labels for the HOMs present in this model.""" return tuple(f"{n},{m}" for n, m in self.homs) @property def homs(self): """An array of higher-order modes (HOMs) included in the model. :`getter`: Returns a copy of the array of the HOMs in the model. :`setter`: Sets the HOMs to be included in the model. See :meth:`Model.modes` for the options available. """ return self._settings.homs @homs.setter @locked_when_built def homs(self, value): self.modes(value) @property def modes_setting(self): # Used by unparser. return self.__mode_setting @property def mode_index_map(self): """An ordered dictionary where the key type is the modes in the model and the mapped type is the index of the mode. :`getter`: Returns the map of modes to indices (read-only). """ return {(n, m): i for i, (n, m) in enumerate(self._settings.homs)}
[docs] @locked_when_built def include_modes(self, modes): """Inserts the mode indices in `modes` into the :attr:`.Model.homs` array at the correct (sorted) position(s). Parameters ---------- modes : sequence, str A single mode index pair or an iterable of mode indices. Each element must unpack to two integer convertible values. """ self._settings.homs = insert_modes(self._settings.homs, modes) if not self._settings.is_modal: self._settings.is_modal = True LOGGER.info(f"Turning on HOMs --> switching model: {self!r} to modal.") self.__mode_setting["include"] = modes
[docs] @locked_when_built def remove_modes(self, modes): """Removes the mode indices in `modes` from the :attr:`.Model.homs` array. Parameters ---------- modes : sequence, str A single mode index pair or an iterable of mode indices. Each element must unpack to two integer convertible values. """ self._settings.homs = remove_modes(self._settings.homs, modes) self.__mode_setting["remove"] = modes
[docs] @locked_when_built def switch_off_homs(self): """Turns off HOMs, switching the model to a plane wave basis.""" LOGGER.info("Turning off HOMs --> switching model to plane wave.") self._settings.homs = np.zeros(1, dtype=(np.intc, 2)) self._settings.is_modal = False
[docs] @locked_when_built def modes(self, modes=None, maxtem=None, include=None, remove=None): """Select the HOM indices to include in the model. See :ref:`selecting_modes` for examples on using this method. Parameters ---------- modes : sequence, str, optional; default: None Identifier for the mode indices to generate. This can be: - An iterable of mode indices, where each element in the iterable must unpack to two integer convertible values. - A string identifying the type of modes to include, must be one of "off", "even", "odd", "x" or "y". maxtem : int, optional; default: None Optional maximum mode order. include : sequence, str, optional A single mode index pair, or an iterable of mode indices, to include. Each element must unpack to two integer convertible values. remove : sequence, str, optional A single mode index pair, or an iterable of mode indices, to remove. Each element must unpack to two integer convertible values. See Also -------- Model.include_modes : Insert mode indices into :attr:`Model.homs` at the correct (sorted) positions. Model.remove_modes : Remove mode index pairs from the model. Examples -------- See :ref:`selecting_modes`. """ if ( (modes is None) and (maxtem is None) and (include is None) and (remove is None) ): if self._settings.is_modal: return self.homs else: return None if not self._settings.is_modal: self._settings.is_modal = True LOGGER.info("Turning on HOMs --> switching model to modal.") # NOTE (sjr) Commenting this out for now as it can cause confusion when # parsing a file - e.g. adding a cavity, gauss object before # maxtem is set will trigger this statement and erroneously lead # the user to believe that only the 00 mode is present # if modes is None and maxtem == 0 and include is None and remove is None: # warn( # "Modal model enabled with only HG00. Call Model.modes to add modes." # ) def do_make_modes(*args, **kwargs): try: return make_modes(*args, **kwargs) except ContextualValueError as e: # make_modes uses different parameters to this method so we need to # rewrite the parameters. if "select" in e.params: # Note: ensure the signature order is retained! params = {"modes": e.params["select"], **e.params} params.pop("select") e.params = params raise e if modes is None: # maxtem if maxtem is None: self.switch_off_homs() else: self._settings.homs = do_make_modes(maxtem=maxtem) elif isinstance(modes, str): # identifier if modes.casefold() == "off": self.switch_off_homs() else: if maxtem is None: raise ValueError( f"Argument maxtem must be specified for modes argument of " f"{modes}" ) self._settings.homs = do_make_modes(modes, maxtem) else: # iterable of mode indices self._settings.homs = do_make_modes(modes) if include is not None: self.include_modes(include) if remove is not None: self.remove_modes(remove) # remember the modes setting for use in the unparser self.__mode_setting["modes"] = modes self.__mode_setting["maxtem"] = maxtem self.__mode_setting["include"] = include self.__mode_setting["remove"] = remove
[docs] @locked_when_built def add_all_ad(self, node, f=0): """Adds amplitude detectors at the specified `node` and frequency `f` for all Higher Order Modes in the model. Parameters ---------- node : :class:`.OpticalNode` Node to add the detectors at. f : scalar, :class:`.Parameter` or :class:`.ParameterRef` Frequency of the field to detect. Returns ------- dets : list A list of all the amplitude detector instances added to the model. """ node_name = node.tag if node_name is None or not valid_name(node_name): node_name = node.full_name.replace(".", "_") ads = [] for n, m in self._settings.homs: ad = detectors.AmplitudeDetector(f"ad_{n}_{m}_{node_name}", node, f, n, m) self.add(ad) ads.append(ad) return ads
[docs] @locked_when_built def add_fd_to_every_node(self, freq=0): """Adds a FieldDetector at every optical node in model at a given frequency The name of each FieldDetector is automatically generated using the following pattern: ``E_<component>_<port>_<node>`` Parameters ---------- f : scalar, :class:`.Parameter` or :class:`.ParameterRef` Frequency of the field to detect. Returns ------- dets : list A list of all the field detector instances added to the model. """ nodes = self.optical_nodes fds = [] for node in nodes: fd_name = f"E_{node.full_name.replace('.','_')}" fd = detectors.FieldDetector(fd_name, node, freq) self.add(fd) fds.append(fd) return fds
@property def phase_level(self): """An integer corresponding to the phase level given to the phase command of a Finesse 2 kat script. :`getter`: Returns the phase level. :`setter`: Sets the phase level - turns on/off specific flags for the scaling of coupling coefficient and Gouy phases. """ warn( "`Model.phase_level = n` deprecrated, use `Model.phase_config` instead", DeprecationWarning, ) lvl = 0 if self._settings.phase_config.zero_k00: lvl += 1 if self._settings.phase_config.zero_tem00_gouy: lvl += 2 return lvl @phase_level.setter @locked_when_built def phase_level(self, value): # No public alternative... warn( "`Model.phase_level = n` deprecrated, use `Model.phase_config` instead", DeprecationWarning, ) if value == 0: self._settings.phase_config.zero_k00 = False self._settings.phase_config.zero_tem00_gouy = False elif value == 1: self._settings.phase_config.zero_k00 = True self._settings.phase_config.zero_tem00_gouy = False elif value == 2: self._settings.phase_config.zero_k00 = False self._settings.phase_config.zero_tem00_gouy = True elif value == 3: self._settings.phase_config.zero_k00 = True self._settings.phase_config.zero_tem00_gouy = True else: raise ValueError(f"Value is not valid {value}")
[docs] @locked_when_built def phase_config(self, zero_k00=False, zero_tem00_gouy=True): """Coupling coefficient and Gouy phase scaling: - phase_level 3 == zero_k00=True, zero_tem00_gouy=True - phase_level 2 == zero_k00=False, zero_tem00_gouy=True - phase_level 1 == zero_k00=True, zero_tem00_gouy=False - phase_level 0 == zero_k00=False, zero_tem00_gouy=False See also :ref:`phase_configurations` This can be used to change the computation of light field phases in the Hermite-Gauss mode. In general, in the presence of higher order modes, the "macroscopic length" of a space is no longer an integer number of wavelengths for the TEM00 mode because of the Gouy phase. Furthermore, the coupling coefficients :math:`k_{nmnm}` contribute to the phase when there is a mode mismatch. For correct analysis these effects have to be taken into account. On the other hand, these extra phase offsets make it very difficult to set a resonance condition or operating point intuitively. In most cases another phase offset can be added to all modes so that the phase of the TEM00 becomes zero. This method allows setting these phase offsets for the propagation through free space, for the coupling coefficients, or both. Regardless of the setting, the phases for all higher modes are changed accordingly such that the relative phases remain correct. Parameters ---------- zero_k00 : bool, optional Scale phase for k0000 (TEM00 to TEM00) coupling coefficients to 0. Defaults to True. zero_tem00_gouy : bool, optional Ensure that all Gouy phases for TEM00 are 0. Defaults to True. """ self._settings.phase_config.zero_k00 = zero_k00 self._settings.phase_config.zero_tem00_gouy = zero_tem00_gouy
@property def Nhoms(self): """Number of higher-order modes (HOMs) included in the model. :`getter`: Returns the number of HOMs in the model (read-only). """ return self._settings.num_HOMs @property def frequencies(self): """The frequencies stored in the model as a :py:class:`list` instance. :`getter`: Returns a list of the model frequencies (read-only). """ return tuple(self.__frequencies) @property def source_frequencies(self): """The source frequencies stored in the model as a :py:class:`dict` instance. The frequencies are the keys, and the values are list of components that depend on the given frequency. :`getter`: Returns a dict of the model source frequencies (read-only). """ return tuple(self.__source_frequencies.keys()) @property def components(self): """The components stored in the model as a tuple object. :`getter`: Returns a tuple of the components in the model (read-only). """ return tuple(self.__components.keys()) @property def detectors(self): """The detectors stored in the model as a tuple object. :`getter`: Returns a tuple of the detectors in the model (read-only). """ return tuple(self.__detectors.keys()) @property def cavities(self): """The cavities stored in the model as a tuple object. :`getter`: Returns a tuple of the cavities in the model (read-only). """ return tuple(self.__cavities.keys()) @property def locks(self): return tuple(self.__locks.keys()) @property def dofs(self): return tuple(self.__dofs.keys()) @property def is_built(self): """Flag indicating whether the model has been built. When this evaluates to `True`, the structure of the underlying matrix should not be changed. :`getter`: `True` if the model has been built, `False` otherwise. """ return self.__is_built @property def trace_order(self): """A list of beam tracing dependencies, ordered by their tracing priority. Dependency (i.e. :class:`.Cavity` and :class:`.Gauss`) objects are ordered in this list according to the priority in which they will be traced during the beam tracing routine. This ordering is strictly defined as follows: Dependencies will be sorted in order of *descending* :attr:`.TraceDependency.priority` value. Any dependencies which have equal :attr:`.TraceDependency.priority` value are sorted *alphabetically* according to their names. Please be aware that this means if no priority values have been given to any :class:`.TraceDependency` instance in the model, as is the default when creating these objects, then this trace order list is simply sorted alphabetically by the dependency names. .. note:: Regardless of their positions in this list, the *internal* traces of :class:`.Cavity` objects will always be performed first. Internal cavity traces are defined as the traces which propagate the cavity eigenmode through all the nodes of the cavity path. Importantly, however, the order in which :class:`.Cavity` objects appear in this trace order list *will* also determine the order in which their internal traces are performed. This is relevant only for when there are overlapping cavities in the model - recycling cavities in dual-recycled Michelson interferometer configurations are a typical case of this. As always see :meth:`.Model.beam_trace` and :ref:`tracing_manual` for more details on the inner workings of the beam tracing routines. Temporary overriding of this order for a given :meth:`.Model.beam_trace` call can be performed by specifying the ``order`` argument for this method call. To override this ordering for a simulation, one should use the ``"order"`` keyword argument of :meth:`.Model.sim_trace_config_manager` to temporarily use any arbitrary dependency order within a context. :`getter`: Returns a list giving the order in which dependencies will be traced. Read-only. """ return self.__trace_order.copy() @property def trace_order_names(self): """A convenience property to retrieve a list of the names of each :class:`.TraceDependency` instance in :attr:`.Model.trace_order`. :`getter`: Returns a list of the names of the dependencies in the order they will be traced. Read-only. """ return [d.name for d in self.trace_order] def __insert_trace_dependency(self, dep): self.__trace_order.append(dep) self._resort_trace_dependencies() def _resort_trace_dependencies(self): # Sort in order of descending priority and ensure dependencies # with equal priority are sorted alphabetically by name self.__trace_order.sort(key=lambda x: (-x.priority, x.name)) @property def trace_forest(self): """The :class:`.TraceForest` instance held by the model. This is a representation of the beam tracing paths from each dependency which takes on a form corresponding to the last call to :meth:`.Model.beam_trace`. See the documentation for :class:`.TraceForest` itself for details on what exactly this object is, and the various methods and properties it exposes. .. hint:: Most of the time users will not need to touch this property as it is generally just used internally. Beam tracing functionality should instead be used via the carefully designed interfaces, i.e: :meth:`.Model.beam_trace` for full model beam traces, :meth:`.Model.propagate_beam` for propagating an arbitrary beam through a path etc. See :mod:`.tracing.tools` for details on various beam tracing tools. Despite the above, it *can* sometimes be useful to query this property to get a visual representation of how the beam tracing paths look in your model. To do this one can simply print the return of this property, i.e.:: print(model.trace_forest) to get a forest-like structure of all the beam tracing trees which represent the current state (as of the last :meth:`.Model.beam_trace` call) of the model. :`getter`: The :class:`.TraceForest` object associated with this model. Read-only. """ return self.__trace_forest
[docs] @contextmanager def sim_trace_config_manager(self, **kwargs): """Change the :attr:`.Model.sim_trace_config` within a context. This provides a convenient pattern through which one can temporarily set the simulation beam tracing behaviour in a ``with`` block. The method :meth:`.Model.reset_sim_trace_config` is called on exit. Parameters ---------- kwargs : keyword arguments See :attr:`.Model.sim_trace_config`. Examples -------- Temporarily change the tracing order:: with model.sim_trace_config_manager(order=["cavXARM", "gaussBS", "cavYARM"]): out = model.run("noxaxis()") Disable certain dependencies in a context:: with model.sim_trace_config_manager(disable="cavIMC"): out = model.run("noxaxis()") Switch off re-tracing and enable only two specific trace dependencies:: with model.sim_trace_config_manager( retrace=False, enable_only=["cavXARM", "cavYARM"] ): out = model.run("noxaxis()") Use asymmetric tracing and mask all data points where any unstable cavity is encountered:: with model.sim_trace_config_manager(symmetric=False, unstable_handling="mask"): out = model.run("noxaxis()") """ diff = set(kwargs) - set(self.sim_trace_config) if diff: raise KeyError( "The following arguments are not sim_trace_config " f"options: {', '.join(diff)}" ) self.sim_trace_config.update(kwargs) try: yield finally: self.reset_sim_trace_config()
@property def sim_trace_config(self): """Dictionary corresponding to beam tracing configuration options for simulations. The (string) keys of this dict are: * The arguments of :meth:`.Model.beam_trace`, see the linked docs for descriptions of each of these. These config values are passed to the initial beam trace call when building a modal simulation, thereby determining the structure of both the :attr:`.Model.trace_forest` used for computing the initial beam parameters, as well as the trace forest of changing beam paths as stored by the simulation itself. * "retrace" --- flag determining whether beam tracing is re-executed, during a simulation, whenever some dependent parameter changes. This is ``True`` by default, meaning that any paths in the model with changing geometric parameters will automatically be retraced during the simulation. Setting this to ``False`` means that the initial beam parameters (from the beam trace executed at the start of the simulation) are used for all data points, regardless of whether any geometric parameter is changing or not. * "unstable_handling" --- the strategy to use when encountering unstable optical cavities during a simulation (as a potential result of scanning geometric parameters). The accepted values for this config option are: * "auto" --- (default) contingency :class:`.TraceForest` instances are created when entering unstable cavity regions; or, if there are no stable :class:`.TraceDependency` objects, detector outputs are masked appropriately in these regions. * "mask" --- detector outputs masked appropriately whenever an unstable cavity is encountered; i.e. nothing else (scatter matrices, gouy phases, refills etc.) is computed for such data points. * "abort" --- immediately aborts the simulation, if an unstable cavity is encountered, by raising a :class:`.BeamTraceException`. .. hint:: Most of the time it is better to use :meth:`.Model.sim_trace_config_manager` to temporarily set simulation beam tracing configuration options, rather than modifying the entries here directly (which then requires manual re-setting as outlined below). :`getter`: Beam tracing configuration options for simulations. Examples -------- One can use this property to change the behaviour of beam tracing for a simulation. For example, this:: model.sim_trace_config["disable"] = "cav1" would switch off tracing from the trace-dependency named `"cav1"` during a simulation. It can also be used to temporarily override the trace order used, without modifying :attr:`.TraceDependency.priority` values and, thus, without modifying the actual :attr:`.Model.trace_order`. For example:: model.sim_trace_config["order"] = ["gL0", "cav2", "cav1"] would set the trace ordering for the next simulation using this model to the order given. To reset the `sim_trace_config` dict entries to the default values, call :meth:`.Model.reset_sim_trace_config`. """ return self.__sim_trace_config @property def sim_initial_trace_args(self): """Filtered dictionary of :attr:`.Model.sim_trace_config` corresponding to only those options which match the arguments of :meth:`.Model.beam_trace`. The arguments of :meth:`.Model.beam_trace`, see the linked docs for descriptions of each of these. These config values are passed to the initial beam trace call when building a modal simulation, thereby determining the structure of both the :attr:`.Model.trace_forest` used for computing the initial beam parameters, as well as the trace forest of changing beam paths as stored by the simulation itself. .. note:: The return value is a new filtered dict, not a sub-view of :attr:`.Model.sim_trace_config`, thus modifying this dict does not affect the entries in :attr:`.Model.sim_trace_config`. """ return { k: self.sim_trace_config[k] for k in inspect.signature(self.beam_trace).parameters.keys() }
[docs] def reset_sim_trace_config(self): """Resets the simulation beam tracing configuration dict, given by :attr:`.Model.sim_trace_config`, to the default values.""" self.__sim_trace_config = self.__default_sim_trace_config.copy()
@property def is_traced(self): """Flag indicating whether the model has been traced. .. warning:: This flag only indicates whether a beam trace has been performed on the model, it *does not* mean that the last stored beam trace (i.e. :attr:`.Model.last_trace`) corresponds to the latest state of the model. :`getter`: `True` if the mode has been traced at least once, `False` otherwise. """ return self.last_trace is not None @property def last_trace(self): """An instance of :class:`.BeamTraceSolution` containing the output of the most recently stored beam trace performed on the model. :`getter`: Returns a copy of the most recently stored beam trace output. Read- only. """ return copy(self.__last_trace)
[docs] @locked_when_built def tag_node(self, node, tag): """Tag a node with a unique name. Access to this node can then be performed with:: node = model.<tag_name> Parameters ---------- node : :class:`.Node` An instance of a node already present in the model. tag : str Unique tag name of the node. """ self.__node_exists_check(node) if hasattr(self, tag): raise Exception(f"Tagged name: {tag} already exists in the model.") node._set_tag(tag) self._unfreeze() setattr(self, tag, node) self._freeze()
[docs] def get_parameters( self, *, include=None, exclude=None, are_changing=None, are_symbolic=None ): """Get all or a filtered list of parameters from all elements in the model. Parameters ---------- include : [iterable|str], optional Parameters that *should* be included. If a single string is given it can be a Unix file style wildcard (See ``fnmatch``). A value of None means everything is included. If an iterable is provided it must be a list of names or Parameter objects. exclude : [iterable|str], optional Parameters that *should not* be included. If a single string is given it can be a Unix file style wildcard (See ``fnmatch``). A value of None means nothing is excluded. If an iterable is provided it must be a list of names or Parameter objects. are_changing : boolean, optional Filter if Parameter has a value that will be marked as changing during a simulation. Note this might be a user chaniging variable or a symbolic value whose arguments are changing. are_symbolic : boolean, optional Filter if Parameter has a symbolic value. If set to None, not filtering is done. Returns ------- parameters : list List of filtered Parameters Examples -------- >>> model = finesse.Model() >>> model.parse(''' ... l l1 ... l l2 ... lens L1 f=100 ... link(l1, L1, l2) ... pd P1 L1.p1.o ... pd P2 L1.p2.o ... ''') >>> print(model.get_parameters(include='*f*')) >>> print(model.get_parameters(include='*f*', exclude='l2*')) [<fsig.f=None @ 0x7f7fc449a580>, <l1.f=0.0 @ 0x7f7fc449a340>, <l2.f=0.0 @ 0x7f7fc449a880>, <L1.f=100.0 @ 0x7f7fc449aac0>] [<fsig.f=None @ 0x7f7fc449a580>, <l1.f=0.0 @ 0x7f7fc449a340>, <L1.f=100.0 @ 0x7f7fc449aac0>] """ from fnmatch import fnmatch name_predicates = [] bool_predicates = [] if exclude is not None: if isinstance(exclude, str): name_predicates.append(lambda x: not fnmatch(x, exclude)) else: name_predicates.append(lambda x: (x not in exclude)) if include is not None: if isinstance(include, str): name_predicates.append(lambda x: fnmatch(x, include)) else: name_predicates.append(lambda x: (x in include)) if are_changing is not None: bool_predicates.append(lambda x: x.is_changing == are_changing) if are_symbolic is not None: bool_predicates.append(lambda x: x.is_symbolic == are_symbolic) return [ # get all parameters p for p in self.all_parameters if all(pred(p.full_name) for pred in name_predicates) and all(pred(p) for pred in bool_predicates) ]
[docs] @contextmanager def temporary_parameters(self, include=None, exclude=None): """Context manager that lets user change any ModelParameter then return it to the original value once completed. When the Model is in this temporary state it cannot have any structural changes, such as adding or removing components. There is also the ability to include or exclude certain parameters from reverting back to their previous values if needed. Parameters ---------- include : iterable or str, optional Parameters that *should* be reverted once the context has exitted. If a single string is given it can be a Unix file style wildcard (See ``fnmatch``). A value of None means everything is included. If an iterable is provided it must be a list of names or Parameter objects. exclude : iterable or str, optional Parameters that *should not* be reverted once the context has exitted. If a single string is given it can be a Unix file style wildcard (See ``fnmatch``). A value of None means nothing is excluded. If an iterable is provided it must be a list of names or Parameter objects. Examples -------- .. code-block:: import finesse model = finesse.Model() model.parse(''' l l1 P=1 m m1 R=0.99 T=0.01 Rc=-1934 m m2 R=1 T=0 Rc=2245 m m3 R=1 T=0 Rc=10000 ''' ) with model.temporary_parameters(): model.m1.Rc = 100 print(model.m1.Rc) print(model.m1.Rc) # Only reset m2 parameters with model.temporary_parameters(include="m2.*"): ... # Only reset m2.phi and m1.phi parameters with model.temporary_parameters(include=("m2.phi", "m1.phi")): ... # Reset everything apart from all phi parameters with model.temporary_parameters(exclude="*.phi"): ... # Reset everything apart from all phi parameters with model.temporary_parameters(exclude="m[1-3].phi"): ... """ params = self.get_parameters(include=include, exclude=exclude) try: self.__lock_model_change = True # stop any structural changes to model # Get the current values of all parameters initial_values = {p: copy(p.value) for p in params} initial_settings = deepcopy(self._settings) yield self except Exception as ex: raise ex finally: # Reset the values on exit with warnings.catch_warnings(): warnings.filterwarnings("ignore", category=ModelParameterSettingWarning) for p in params: if p.value != initial_values[p]: p.value = initial_values[p] self._settings = initial_settings self.__lock_model_change = False
[docs] @locked_when_built def remove(self, obj): """Removes an object from the model. .. note:: If a string is passed, it will be looked up via self.elements. Parameters ---------- obj : :class:`.Frequency` or sub-class of :class:`.ModelElement` The object to remove from the model. Raises ------ Exception If the matrix has already been built or there is no component with the given name in the model. """ if self.__lock_model_change: raise FinesseException("Model is locked and cannot be added or removed to.") if isinstance(obj, str): obj = self.elements[obj] if obj in self.__not_removable: raise FinesseException(f"{obj} cannot be removed from the model.") if isinstance(obj, ModelElement): # Check that any of this elements parameters are not being referenced by # someone else in the model symbolic_params = ( p for p in self.all_parameters if p.is_symbolic and p.owner is not obj ) for sp in symbolic_params: sp_id_parameters = tuple(id(_) for _ in sp.value.parameters()) for p in obj.parameters: if id(p) in sp_id_parameters: raise FinesseException( f"Cannot remove element {repr(obj)}.\nParameter {repr(p)} is being used by {repr(sp)}." ) if obj in self.__components: nodes = obj.nodes # stores (port1, port2) keys where port1 should now connect to port2 pairs = {} for name, node in nodes.items(): if node.type != NodeType.OPTICAL or not node.is_input: continue predecessors = list(self.network.predecessors(name)) if not predecessors: continue pre = predecessors[0] # immediate successors of the node - always exist im_succ = list(self.network.successors(name)) # find only the successor node which is at a different port match = list( filter( lambda n: self.network.nodes[n]["weakref"]().port != node.port, im_succ, ) )[0] # successors of the immediate successors successors = list(self.network.successors(match)) if not successors: continue succ = successors[0] p1 = self.network.nodes[pre]["weakref"]().port p2 = self.network.nodes[succ]["weakref"]().port # don't add the port pair again if we already did the # previous propagation from the opposite port if (p2, p1) not in pairs: p1_space = p1.space p2_space = p2.space pairs[(p1, p2)] = p1_space.L.value + p2_space.L.value self.network.remove_edge(pre, name) self.network.remove_edge( node.opposite.full_name, self.network.nodes[pre]["weakref"]().opposite.full_name, ) self.network.remove_edge(match, succ) self.network.remove_edge( self.network.nodes[succ]["weakref"]().opposite.full_name, self.network.nodes[match]["weakref"]().opposite.full_name, ) for n in p1.nodes: try: self.network.remove_edge( p1_space.phase_sig.i.full_name, n.full_name ) except nx.exception.NetworkXError: pass for n in p2.nodes: try: self.network.remove_edge( p2_space.phase_sig.i.full_name, n.full_name ) except nx.exception.NetworkXError: pass self.network.remove_node(name) del self.__elements[p1_space.name] del self.__elements[p2_space.name] for (p, s), L in pairs.items(): self.connect(p, s, L=L) del self.__components[obj] self._rebuild_trace_forest = True elif isinstance(obj, (Space, Wire)): # Remove any connections from this element self.network.remove_edges_from( ( (n[0].full_name, n[1].full_name) for n in obj.all_internal_connections.values() ) ) # Remove any signal nodes too self.network.remove_nodes_from((n.full_name for n in obj.signal_nodes)) if isinstance(obj, Space): if isinstance(obj.p1.i.component, Surface): obj.p1.i.component._resymbolise_ABCDs() if isinstance(obj.p2.o.component, Surface): obj.p2.o.component._resymbolise_ABCDs() elif obj in self.detectors: del self.__detectors[obj] elif obj in self.__cavities or obj in self.__gausses.values(): args = ("order", "disable", "enable_only") for arg in args: thing = self.__sim_trace_config[arg] if thing is not None: if obj in thing: thing.remove(obj) elif obj.name in thing: thing.remove(obj.name) if obj in self.__trace_order: self.__trace_order.remove(obj) if obj in self.__cavities: del self.__cavities[obj] elif obj in self.__gausses.values(): del self.__gausses[obj.node] self._rebuild_trace_forest = True elif obj in self.__locks: del self.__locks[obj] elif obj in self.__dofs: del self.__dofs[obj] elif isinstance(obj, ReadoutDetectorOutput): self.network.remove_nodes_from( (n.full_name for n in obj.readout.signal_nodes) ) self._rebuild_trace_forest = True else: raise Exception( "Element with name {} not in this model".format(obj.name) ) if hasattr(obj, "_on_remove"): obj._on_remove() obj._set_model(None) del self.__elements[obj.name] # Remove element from any namespaces for namespace in obj._namespace: parent = self for _ in namespace.split("."): if len(_) > 0: if hasattr(parent, _): parent = getattr(parent, _) delattr(parent, obj.name) else: raise TypeError( "Object {} not recongised as a Finesse ModelElement.".format(obj) )
[docs] def get(self, attr): """Get an attribute of the model using a string path representation like `l1.p1.o.q`. If a :class:`ModelElement`, :class:`Node`, or a :class:`Port` is given it will try to return the equivalent object in this model. Parameters ---------- attr : [str | ModelElement | Node | Port] An object to get from the model. Could be a generic string, or Examples -------- Parse a simple model and extract the laser power model parameter: >>> import finesse >>> kat = finesse.Model() >>> kat.parse(''' ... l l1 P=2 ... s s1 l1.p1 m1.p1 ... m m1 R=0.5 T=0.5 ... pd Pr m1.p1 ... pd Pt m1.p2 ... ''') >>> kat.get("l1.P") <l1.P=2.0 @ 0x11aa56588> See Also -------- :method:`Model.get_element` """ if isinstance(attr, str): return self._reduce_model_path(attr.strip().split(".")) elif isinstance(attr, ModelElement): return self.get_element(attr) elif hasattr(attr, "full_name"): return self.get(attr.full_name) elif hasattr(attr, "name"): return self.get(attr.name) else: raise FinesseException( f"model.get() should be provided with a str, ModelElement, Node, or Port, or some object that has a name or full_name attribute. Not {repr(attr)} ({type(attr)})" )
[docs] @deprecated(reason="Use `Model.get`.") def reduce_get_attr(self, attr): return self.get(attr)
[docs] def set(self, attr, value): """Set an attribute of the model using a string path representation like `l1.p1.o.q`.""" attrs = attr.split(".") # Set the final attribute. setattr( self._reduce_model_path(attrs[:-1]), attrs[-1], value, )
[docs] @deprecated(reason="Use `Model.set`.") def reduce_set_attr(self, attr, value): return self.set(attr, value)
def _reduce_model_path( self, attrs: list[str], check_namespaces: bool = True, ) -> Any: """Recursively follow the attribute names in `attrs` by calling `getattr`, starting at this model. Only meant to find object attributes, like `s1.p1.o`, not class attributes like `parse`, `__dict__` etcetera. Parameters ---------- attrs : list[str] List of attribute names describing a model path, e.g. `l1.P` as `['l1', 'P'] check_namespaces: bool Whether to check namespaces such as `.spaces` or `.wires`. Defaults to True Returns ------- Any The resolved attribute Raises ------ ModelClassAttributeError When the model path resolves to a class attribute that is not a property ModelMissingAttributeError When the model path fails to resolve """ targets = [self] if check_namespaces: # we check the namespaces first, since the model is the most likely target # and a missing attribute on the model produces the most helpful suggestions # https://gitlab.com/ifosim/finesse/finesse3/-/issues/622 targets = self._get_namespaces() + targets orig_attrs = deepcopy(attrs) while len(targets): resolved_attrs = [] target = targets.pop(0) attrs = deepcopy(orig_attrs) while len(attrs): attr_name = attrs.pop(0) try: attribute = getattr(target, attr_name) try: class_attr = getattr(type(target), attr_name) if not isinstance(class_attr, property): raise ModelClassAttributeError( target, resolved_attrs, attr_name ) except AttributeError: pass if len(attrs): target = attribute else: return attribute except AttributeError: if not len(targets): raise ModelMissingAttributeError( target, resolved_attrs, attr_name ) resolved_attrs.append(attr_name) def _get_namespace_paths(self) -> OrderedSet[str]: # namespaces are not really documented, but every ModelElement has a # '_namespaces' attribute. This is (by convention) a tuple of strings # starting with a '.', like '(".", ".spaces"). This tells the Model class # to save the ModelElement under these namespaces, creating "Freezable" # objects as parents where necessary. namespace_paths: OrderedSet[str] = OrderedSet() for element in self.__elements.values(): assert isinstance(element, ModelElement) namespace_paths |= OrderedSet(element._namespace) if "." in namespace_paths: namespace_paths.remove(".") return namespace_paths def _get_namespaces(self) -> list[Freezable]: namespaces = [] for namespace in sorted(self._get_namespace_paths()): if namespace.startswith("."): namespace = namespace[1:] namespaces.append(reduce(getattr, namespace.split("."), self)) return namespaces
[docs] @locked_when_built def add(self, obj, *, unremovable=False): """Adds an element (or sequence of elements) to the model - these can be :class:`.ModelElement` sub-class instances. When the object is added, an attribute defined by `obj.name` is set within the model allowing access to the object just added via `model.obj_name` where `obj_name = obj.name`. Parameters ---------- obj : Sub-class of :class:`.ModelElement` (or sequence of) The object(s) to add to the model. unremovable : bool, optional When True, this object will not be able to be removed from this model Returns ------- element : ModelElement The object that was added Raises ------ Exception If the matrix has already been built, the component has already been added to the model or `obj` is not of a valid type. """ if self.__lock_model_change: raise FinesseException("Model is locked and cannot be added or removed to.") if is_iterable(obj): return [self.add(o) for o in obj] try: if obj._model is not None: raise FinesseException( f"Element {obj.name} already thinks it is attached to a different model" ) except ComponentNotConnected: pass if obj.name in self.__elements: raise FinesseException( f"An element with the name {obj.name} is already present" f" in the model ({self.__elements[obj.name]})" ) if obj._unique_element: if len(self.get_elements_of_type(type(obj))) > 0: raise FinesseException( f"An element with type {type(obj)} has already been added to the model" ) assert isinstance(obj, ModelElement) obj._set_model(self) add_to_elements = True if isinstance(obj, Space): for node in obj.nodes.values(): self.__add_node_to_graph(node, obj) for key in obj._registered_connections: From, To = obj._registered_connections[key] From_name, To_name = obj._registered_connections[key] From = obj.nodes[From_name] To = obj.nodes[To_name] self.__add_connection_to_graph(key, From, To, obj) # Need to re-calculate symbolic ABCD matrices for connected # surfaces now that refractive index symbols are different if isinstance(obj.p1.i.component, Surface): obj.p1.i.component._resymbolise_ABCDs() if isinstance(obj.p2.o.component, Surface): obj.p2.o.component._resymbolise_ABCDs() elif isinstance(obj, Wire): # Wires only have one connection each self.__add_connection_to_graph("WIRE", obj.nodeA, obj.nodeB, obj) elif isinstance(obj, detectors.Detector): # Tell component to associate itself with this model if obj.name in self.__detectors: raise FinesseException( f"Detector with name {obj.name} already added to this model" ) if obj.node is not None and obj.node._model != self: raise FinesseException( f"The node of detector {obj.name}, {obj.node}, is not part of {self!r}" ) elif isinstance(obj, detectors.CavityPropertyDetector): obj._set_cavity() elif isinstance(obj, detectors.Gouy): obj._lookup_spaces() self.__detectors[obj] = len(self.__detectors) # If the detector needs a modal basis defined then # switch on modes if not modal already if obj.needs_trace: if not self.is_modal: self.modes(maxtem=0) elif isinstance(obj, Cavity): if obj.name in self.__cavities: raise Exception(f"cavity {repr(obj.name)} already added to this model") self.__cavities[obj] = len(self.__cavities) # compute all the cavity properties (including path determination) obj.initialise() self.__insert_trace_dependency(obj) if not obj.is_stable: warn( f"cavity {repr(obj.name)} added to the model is unstable", CavityUnstableWarning, ) # Turn on HOMs if model is still plane-wave if not self.is_modal: self.modes(maxtem=0) elif isinstance(obj, Connector): if obj.name in self.__components: raise Exception(f"Element {repr(obj.name)} already added to this model") for _ in obj.nodes.values(): self.__add_node_to_graph(_, obj) for key in obj._registered_connections: From_name, To_name = obj._registered_connections[key] From = obj.nodes[From_name] To = obj.nodes[To_name] self.__add_connection_to_graph(key, From, To, obj) self.__components[obj] = len(self.__components) if isinstance(obj, FrequencyGenerator): self._frequency_generators.append(obj) if isinstance(obj, DegreeOfFreedom): self.__dofs[obj] = len(self.__dofs) elif isinstance(obj, Lock): self.__locks[obj] = len(self.__locks) elif isinstance(obj, Gauss): obj.qx.wavelength = self.lambda0 obj.qy.wavelength = self.lambda0 space = obj.node.space if space is not None: nr = space.nr.value else: nr = 1.0 obj.qx.nr = nr obj.qy.nr = nr current_gauss = self.__gausses.get(obj.node) if current_gauss is not None: raise ValueError( f"A Gauss object with name {current_gauss.name} already exists " f"at the node {obj.node.full_name}" ) self.__gausses[obj.node] = obj self.__insert_trace_dependency(obj) # Turn on HOMs if model is still plane-wave if not self.is_modal: self.modes(maxtem=0) elif isinstance(obj, Variable): # Variables are not normal model elements in this regard as they # simply add a new parameter to the model but the elemnt isn't actually # saved anymore obj._add_to_model_namespace = False add_to_elements = False return self.add_parameter( obj.name, obj.value, description=obj.description, units=obj.units, is_geometric=obj.is_geometric, changeable_during_simulation=obj.changeable_during_simulation, dtype=float, ) elif isinstance(obj, ModelElement): # Model elements come in a variety of forms, which are dealt with above # if the object isn't any of these the bare minimum contract is that we # tell it to associate itself with this model and call _on_add to let it # do some initialisation if necessary pass else: raise Exception("Could not add object {}".format(str(obj))) if obj._add_to_model_namespace: # If the elment requests asks it will be added to the model namespace if hasattr(self, obj.name): raise Exception( f"Not a valid {obj.__class__.__name__} name. An attribute " f"called `{obj.name}` already exists in the Model" ) # Can add objects to multiple namespaces for collecting # together like components as well as easy access assert isinstance(obj._namespace, tuple) for namespace in obj._namespace: parent = self for _ in namespace.split("."): if len(_) > 0: if hasattr(parent, _): curr = getattr(parent, _) else: curr = Freezable() parent._unfreeze() setattr(parent, _, curr) parent._freeze() parent = curr parent._unfreeze() setattr(parent, obj.name, obj) parent._freeze() if hasattr(obj, "_on_add"): obj._on_add(self) if hasattr(obj, "_on_frequency_change"): self.__frequency_change_callbacks.append(obj._on_frequency_change) if add_to_elements: self.__elements[obj.name] = obj if unremovable: self.__not_removable.add(obj) # All model elements should be freezable so that users don't get confused # when they misspell certain attributes and set to them, such as mirror.r obj._freeze() # Notify that the TraceForest needs rebuilding on next beam_trace call # when adding an object which can change the forest structure self._rebuild_trace_forest = isinstance(obj, (Connector, TraceDependency)) return obj
@locked_when_built def __add_node_to_graph(self, node, owner): if not self.network.has_node(node.full_name): ref = node.full_name self.network.add_node( ref, weakref=weakref.ref(node), owner=weakref.ref(owner) ) self.network.nodes[ref]["type"] = node.type if node.type == NodeType.OPTICAL: self.network.nodes[ref]["optical"] = True elif node.type == NodeType.MECHANICAL: self.network.nodes[ref]["mechanical"] = True elif node.type == NodeType.ELECTRICAL: self.network.nodes[ref]["electrical"] = True else: raise Exception("Type unhandled") elif not owner._borrows_nodes and node.type == NodeType.OPTICAL: raise Exception("Node {} already added".format(node)) @locked_when_built def __add_connection_to_graph(self, name, From, To, owner): for _ in [From, To]: if not self.network.has_node(_.full_name): raise Exception("Node {} hasn't been added".format(_)) self.network.add_edge( From.full_name, To.full_name, name=name, in_ref=weakref.ref(From), out_ref=weakref.ref(To), owner=weakref.ref(owner), length=1, coupling_type=owner.coupling_type(From, To), internal=From is To, # if edge is to the same element )
[docs] @locked_when_built def parse(self, text, spec=None): """Parses kat script and adds the resulting objects to the model. Parameters ---------- text : :py:class:`str` The kat script to parse. spec : :class:`.KatSpec`, optional The language specification to use. Defaults to the shared :class:`.KatSpec` instance. See Also -------- parse_file : Parse script file. parse_legacy : Parse Finesse 2 kat script. parse_legacy_file : Parse Finesse 2 kat script file. """ from .script import parse parse(text, model=self, spec=spec)
[docs] @locked_when_built def parse_file(self, path, spec=None): """Parses kat script from a file and adds the resulting objects to the model. Parameters ---------- path : str, :class:`pathlib.Path`, or file-like The path or file object to read kat script from. If an open file object is passed, it will be read from and left open. If a path is passed, it will be opened, read from, then closed. spec : :class:`.KatSpec`, optional The language specification to use. Defaults to the shared :class:`.KatSpec` instance. See Also -------- parse : Parse script. parse_legacy : Parse Finesse 2 kat script. parse_legacy_file : Parse Finesse 2 kat script file. """ from .script import parse_file parse_file(path, model=self, spec=spec)
[docs] @locked_when_built def parse_legacy(self, text): """Parses legacy (Finesse 2) kat script and adds the resulting objects to the model. Parameters ---------- text : :py:class:`str` The kat script to parse. See Also -------- parse_legacy_file : Parse Finesse 2 kat script file. parse : Parse Finesse 3 kat script. parse_file : Parse Finesse 3 kat script file. """ from .script import parse_legacy parse_legacy(text, model=self)
[docs] @locked_when_built def parse_legacy_file(self, path): """Parses legacy (Finesse 2) kat script from a file and adds the resulting objects to the model. Parameters ---------- path : str, :class:`pathlib.Path`, or file-like The path or file object to read kat script from. If an open file object is passed, it will be read from and left open. If a path is passed, it will be opened, read from, then closed. See Also -------- parse_legacy : Parse Finesse 2 kat script. parse : Parse Finesse 3 kat script. parse_file : Parse Finesse 3 kat script file. """ from .script import parse_legacy_file parse_legacy_file(path, model=self)
[docs] def unparse(self, inplace=True): """Serialise the model to kat script. Returns ------- str The generated kat script. """ from .script import unparse return unparse(self, ref_graph=self.syntax_graph if inplace else None)
[docs] def unparse_file(self, path, inplace=True): """Serialise the model to kat script in a file. Parameters ---------- path : str, :class:`pathlib.Path`, or file-like The path or file object to write kat script to. If an open file object is passed, it will be written to and left open. If a path is passed, it will be opened, written to, then closed. """ from .script import unparse_file unparse_file(path, self, ref_graph=self.syntax_graph if inplace else None)
[docs] @locked_when_built def merge( self, other, from_comp, from_port, to_comp, to_port, name=None, L=0, nr=1 ): """Merges the model `other` with this model using a connection at the specified ports. .. note:: Upon completion of this method call the `Model` instance `other` will be invalidated. All components and nodes within `other` will be associated with **only** this model. Parameters ---------- other : :class:`.Model` A model configuration to merge into this model instance. from_comp : Sub-class of :class:`.Connector` The component to start a connection from. from_port : int Port of `from_comp` to initiate the connection from. to_comp : Sub-class of :class:`.Connector` The component to bridge the connection to. to_port : int Port of `to_comp` to bridge the connection to. name : str Name of connecting :class:`.Space` instance. L : float Length of the connecting space. nr : float Index of refraction of the connecting space. """ self._unfreeze() raise NotImplementedError() self._settings.homs_view += other.homs # combine components/detectors for node in list(other.network.nodes): try: self.add(node.component) except Exception: continue # combine spaces for edge_tuple in list(other.network.edges): for edge in edge_tuple: try: self.add(edge.space) except Exception: continue # combine cavities for cav in other.cavities: self.add(cav) # check self.__opt_net_view correctly points to correct new graph if this # is ever reused self.__network = nx.compose(self.__network, other.network) self.connect(from_comp, from_port, to_comp, to_port, name=name, L=L, nr=nr) self._freeze()
[docs] @locked_when_built def add_frequency(self, freq): """Add a specific optical carrier frequency to the model description. Parameters ---------- :class:`float` or :class:`.Frequency` The frequency to add. """ if freq in self.__frequencies: warn(f"frequency {repr(freq.name)} already added to model") else: self.__frequencies.add(freq)
[docs] def get_frequency_object(self, frequency_value): if frequency_value in self.__freq_map: return self.__freq_map[frequency_value][0] else: return None
[docs] @locked_when_built def chain(self, *args, start=None, port=None): """Utility function for connecting multiple connectable objects in a sequential list together. Between each item the connection details can be specified, such as length or refractive index. This function also adds the elements to the model and returns those as a tuple to for the user to store if required. Examples -------- Make a quick 1m cavity and store the added components into variables:: l1, m1, m2 = ifo.chain(Laser('l1'), Mirror('m1'), 1, Mirror('m2')) Or be more specific about connection parameters by providing a dictionary. This dictionary is passed to the :meth:`Model.connect` method as kwargs so see there for which options you can specify. For optical connections we can set lengths and refractive index of the space using a dictionary:: ifo.chain( Laser('l1'), Mirror('AR'), {'L':1e-2, 'nr':1.45}, Mirror('HR') ) In the above case a auto-generated space name will be made. If you want to explicitly set a name use `{'L':1e-2, 'nr':1.45, 'name':"my_space"},` The starting point of the chain can be specfied for more complicated setups like a Michelson:: ifo = Model() ifo.chain(Laser('lsr'), Beamsplitter('bs')) # connecting YARM to BS ifo.chain( 1, Mirror('itmy'), 1, Mirror('etmy'), start=ifo.bs, port=2, ) # connecting XARM to BS ifo.chain( 1, Mirror('itmx'), 1, Mirror('etmx'), start=ifo.bs, port=3, ) Parameters ---------- start: component, optional This is the component to start the chain from. If None, then a completely new chain of components is generated. port: int, optional (required if `start` defined) The port number at the `start` component provided to start the chain from. This must be a free unconnected port at the `start` component or an exception will be thrown. Returns ------- tuple A tuple containing the objects added. The `start` component is never returned. """ connectors = [] connections = [] was_prev_connector = False if start is not None: if start not in self.components: raise Exception("Component %s is not in this model" % start) if port is None: raise Exception( "Port keyword argument must also be" "provided if specifying a start for" "the chain." ) connectors.append(start) was_prev_connector = True for i, item in enumerate(args): if isinstance(item, Space) or isinstance(item, Wire): connections.append({"connector": item}) was_prev_connector = False elif isinstance(item, Connector): if was_prev_connector: if i == 0 and start is not None: connections.append({"port": port}) else: connections.append({}) connectors.append(item) was_prev_connector = True self.add(item) elif isinstance(item, Number): connections.append({"L": item}) if i == 0 and start is not None: connections[-1]["port"] = port was_prev_connector = False else: connections.append(item) was_prev_connector = False pairs = list(pairwise(connectors)) # There should always be an equal number of connections # for each pair of components, otherwise we are missing something... assert len(pairs) == len(connections) for (a, b), conn in zip(pairs, connections): if "port" in conn: port = conn["port"] del conn["port"] else: if len(a.optical_nodes) == 2: port = a.p1 elif len(a.optical_nodes) == 4: port = a.p2 elif len(a.optical_nodes) > 4: break else: raise Exception("Unhandled: " + str(a, len(a.nodes))) if type(port) is int: port = a.ports[port - 1] self.connect(port, b.p1, **conn) # Don't return the starting point the user specified if start is None: return connectors else: return connectors[1:]
[docs] def get_open_ports(self): """Return all optical ports that are not connected to a space.""" open_ports = {} for node in self.optical_nodes: if node.space is None: # Not connected to anything = open port open_ports[node.port] = None return tuple(open_ports.keys())
[docs] def get_elements_connected_to(self, element): """Returns a set of elements that `element` is connected to. Parameters ---------- element : str or :class:.`Element` Element to query connections of """ element = self.get_element(element) edges = [] for node in element.nodes: for edge in self.network.in_edges(node, data=True): if edge[2]["owner"]() is not element: edges.append(edge) for edge in self.network.out_edges(node, data=True): if edge[2]["owner"]() is not element: edges.append(edge) # spaces and wires connecting this element to something connected_to = OrderedSet() for _ in edges: connected_to.add(_[2]["owner"]()) return connected_to
[docs] @locked_when_built def connect( self, A, B, L=0, nr=1, gain=1, *, delay=None, name=None, verbose=False, connector=None, ): """Connects two ports in a model together. The ports should be of the same type, e.g. both optical ports. This method will also accept components from the user, in such cases it will loop through the ports and use the first one in `.ports` that is currently unconnected. As `connect` will try to be somewhat smart in guessing what the user is trying to connect, use `verbose=True` to print what is actually getting connected. Parameters ---------- A : :class:`.Connector` or :class:`.Port` Component to connect B : :class:`.Connector` or :class:`.Port` Other component to connect L : float, optional Length of newly created :class:`.Space` or :class:`.Wire` instance. If connecting electronics, L will be treated as a delay in seconds nr : float, optional Index of refraction of newly created :class:`.Space`. gain : float, symbol, optional Gain of a wire connection for simply scaling between two signals delay : float, optional Delay time for electrical connections. name : str, optional Name of newly created :class:`.Space` or :class:`.Wire` instance. verbose : bool, optional When True, the actual connections being made will be printed. Raises ------ Exception If matrix has already been built, either of `compA` or `compB` are not present in the model, either of `portA` or `portB` are already connected or either of `portA` or `portB` are not valid options at the specified component(s). """ ports = [None, None] if isinstance(A, str): A = self.get(A) if isinstance(B, str): B = self.get(B) for item in [A, B]: # Check if these have been added to a model, if not add to this # if item._model is None: try: _ = item._model except ComponentNotConnected: if hasattr(item, "component"): # In case some port or node has been provided instead # get the actual component to add self.add(item.component) else: self.add(item) def signal_ports(obj): return tuple( p for p in obj.ports if p.type in (NodeType.MECHANICAL, NodeType.ELECTRICAL) ) def get_input_port(ports): if ports[0].nodes[0].direction == NodeDirection.INPUT: rtn = ports[0] else: rtn = ports[1] if verbose: print(f"Selecting input port {rtn!r} from {ports}") return rtn def get_output_port(ports): if ports[0].nodes[0].direction == NodeDirection.INPUT: rtn = ports[1] else: rtn = ports[0] if verbose: print(f"Selecting output port {rtn!r} from {ports}") return rtn def is_electronic_component(obj): return all(p.type == NodeType.ELECTRICAL for p in obj.ports) def get_next_optical_port(obj): rtn = None for p in obj.ports: if not p.is_connected and p.type == NodeType.OPTICAL: rtn = p break if rtn is None: raise FinesseException( f"No unconnected optical ports left at {obj.name} to automatically choose from. Please specify an electrical or mechanical one if that was your aim." ) if verbose: print(f"Selecting port {rtn!r} for {obj!r}") return p # Aim here is to try and be smart and accept some more user friendly # inputs rather than always specifying the port name. for i, obj in enumerate((A, B)): if isinstance(obj, Port): # User gave us a port so just use that ports[i] = obj elif isinstance(obj, SignalNode): ports[i] = obj elif isinstance(obj, ModelElement): Np = len(obj.ports) if Np == 1: # just select the first port as that's our only option ports[i] = obj.ports[0] elif Np > 1: # multiple ports! So which one to choose? if i == 0: # If we are the first argument... if is_electronic_component(obj): if Np == 2: # if the first component is a two port electronics # grab the output port ports[i] = get_output_port(obj.ports) else: raise Exception( f"Must specify a port when connecting {A!r} as it has more than 1 port." ) else: ports[i] = get_next_optical_port(obj) else: # From the A object try and guess what to look for # on the second B object. ptype = ports[0].type if ptype == NodeType.OPTICAL: # connecting from an optical port, so find the frst optical # port that isn't connected to anything else ports[i] = get_next_optical_port(obj) else: # must be elec or mech signal node # connecting from an elec port, now find an input port sig_ports = signal_ports(obj) Nsp = len(sig_ports) if Nsp == 1: # just select the only option we have ports[i] = sig_ports[0] elif Nsp == 2: ports[i] = get_input_port(sig_ports) else: raise FinesseException( f"Too many signal ports to choose from at {obj!r} to automatically choose from. Specify one of {sig_ports}" ) else: raise Exception("Model element {} has no ports".format(obj.name)) elif isinstance(obj, OpticalNode): raise FinesseException( f"Optical node {obj!r} cannot be explicitly connected to. Use its port instead `{obj.port.full_name}`." ) else: raise FinesseException(f"Input {obj!r} cannot be connected") if connector is None: try: if all(p.type == NodeType.OPTICAL for p in ports): if delay is not None: raise Exception("Can't set delay for an optical space") connector = Space(name, *ports, L=L, nr=nr) elif all(isinstance(p, SignalNode) for p in ports): # User specified signal nodes so just wire them up directly connector = Wire( name, *ports, delay=(delay if delay is None else L), gain=gain ) elif ( ports[0].type == NodeType.ELECTRICAL or ports[0].type == NodeType.MECHANICAL ): if nr != 1: raise Exception( "Can't set refractive index for an electronic connection" ) if delay is None: delay = L connector = Wire(name, *ports, delay=delay, gain=gain) else: raise FinesseException( f"Do not know how to connect {ports[0]!r} to {ports[1]!r}" ) except Exception as ex: raise FinesseException( f"Whilst trying to connect {ports[0]!r} to {ports[1]!r}, the following exception occurred:\n{ex}" ) else: connector.connect(*ports) self.add(connector)
[docs] @locked_when_built def disconnect(self, A, B): """Disconnects two elements `A` and `B`."""
[docs] def get_network( self, network_type: str | NetworkType = NetworkType.FULL, add_edge_info=False ) -> nx.Graph: """Get specified network. Parameters ---------- network_type : str, optional The network type to export: full (nodes, ports and components), "components" (just components), or "optical" (the optical subnetwork). Returns ------- :class:`networkx.DiGraph` The network. Raises ------ ValueError If the specified network_type is unknown. """ if isinstance(network_type, str): network_type = NetworkType(network_type.casefold()) return network_type.get_network(self, add_edge_info)
[docs] def display_signal_blockdiagram(self, *nodes, **kwargs): """Displays a block diagram of the signal paths in this model. It will only contain electrical and mechanical connections made. Optical couplings are not shown. Parameters ---------- remove_mechanical_to_mechanical : bool, optional If true, mechanical to mechanical node edges are removed. On by default as the result block diagram is complicated. *nodes : :class:`Node` Nodes to show the path of """ from finesse.utilities.blockdiag import display_loops_blockdiag display_loops_blockdiag(self, *nodes, **kwargs)
[docs] def plot_graph( self, layout: str = "neato", graphviz=True, network_type: str | NetworkType = NetworkType.COMPONENT, root: str | ModelElement | Node | None = None, show_detectors: bool = False, radius: int | None = None, directed: bool = False, path: Path | None = None, show: bool = True, format: plot_format = "svg", **kwargs, ): """Plot the node network. See also :ref:`model_visualization` Parameters ---------- layout : str, optional The networkx plotting layout engine, see NetworkX manual for more details. Choose from: circular, fruchterman_reingold, kamada_kawai, multipartite, planar, random, shell, spectral, spiral, spring, neato, dot, fdp, sfdp, or circo graphviz : bool, optional Whether to use the graphviz library for better node layouts (needs optional dependency), by default True network_type : str | NetworkType, optional Which network to plot, can be one of 'optical', 'component', 'full', by default NetworkType.COMPONENT root : str | ModelElement | Node | None, optional In combination with ``radius``, the root node of the graph to use for distance based filtering, by default ``None``. When ``network_type`` is ``components``, ``None`` will default to the first :class:`.finesse.components.laser.Laser` found in the model. show_detectors : bool, optional Whether to add detectors to the graph, by default False radius : int >= 1 | None, optional Must be used in combination with ``root``, only include nodes of the network that are ``radius`` or less edges away from the root node, by default None, meaning that no nodes are filtered out. directed : bool, optional Whether to use directed distance-based filtering, by default False. If set to True, will only include outgoing edges from the root node. See also the ``undirected`` argument of :func:`networkx.generators.ego.ego_graph` path : Path | None, optional Save the resulting image to the given path. Defaults to None, which saves in a temporary file that is displayed if 'show' is set to True. show : bool, optional Whether to show the resulting image. In Jupyter environments, shows the plot inline, otherwise opens a webbrowser for svgs and PIL for pngs. Defaults to True. Notes ----- Uses :func:`networkx.generators.ego.ego_graph` for the distance-based filtering. """ from .plotting import plot_graph if ( NetworkType(network_type) is not NetworkType.COMPONENT and root is None and radius is not None ): raise ValueError( "Need to specify a root if `network_type` is not " f"'{NetworkType.COMPONENT}' and `radius` is not None" ) if graphviz and not has_pygraphviz(): warnings.warn( "Graphviz plot requested, but graphviz is not available." "Reverting to networkx.", stacklevel=2, ) graphviz = False if not has_pygraphviz() and layout in ["neato", "dot", "fdp", "sfdp", "circo"]: warnings.warn( f"Requested layout {layout} not available in networkx, reverting to" " spring layout.", stacklevel=2, ) layout = "spring" network_filter = NetworkType(network_type).filter_class( model=self, root=root, radius=radius, undirected=not directed, add_detectors=show_detectors, ) return plot_graph( network_filter.run(), layout=layout, graphviz=graphviz, path=path, show=show, format=format, **kwargs, )
def __toggle_param_locks(self, state): """Togglin.""" LOGGER.info(f"Toggle parameter lock to {state}") for k in self.__elements: el = self.__elements[k] for p in el._params: # if this is marked as tunable then don't lock it if p.is_tunable and state is True: p._locked = False LOGGER.info(f"{repr(p)} is not being locked") else: p._locked = state
[docs] def get_changing_edges_elements(self): """ Returns ------- tuple(set of (node1-name, node2-name) edges, dict(weakref(element):list)) Returns a list of the network edges that will be changing, further information on the edge can be retreived directly from the model network. Also returned is a dictionary of elements and which parameter is changing """ changing_elements = defaultdict(OrderedSet) changing_edges = OrderedSet() for _ in self.network.edges(): owner = self.network.edges[_]["owner"] if owner() not in changing_elements: for p in owner().parameters: if p.is_changing: changing_elements[owner].add(p) changing_edges.add(_) return changing_edges, changing_elements
[docs] def unbuild(self): """If a model has been built then this function undoes the process so the model can be changed and rebuilt if required.""" if not self.is_built: raise Exception("Model has not been built") # To unfreeze a graph you have to rebuild it self.__network = nx.DiGraph(self.__network) # Need to make a new view too otherwise we miss new additions to the graph self.__opt_net_view = make_optical_network_view(self) self.__toggle_param_locks(False) self.__is_built = False self._on_unbuild() # fire event
[docs] @locked_when_built def run( self, analysis=None, return_state=False, progress_bar=False, simulation_type=None, simulation_options=None, ): """Runs the current analysis set for this model. If no analysis has been set in the model and the ``analysis`` argument is None, then this will run a ``Noxaxis()`` on the current model. If a separate analysis has been provided with the `analysis` argument then this will be run instead of what has been set to `model.analysis`. Parameters ---------- analysis : [str, Action], optional KatScript code for an analysis or an analysis object to run. return_state : bool, optional Whether to return the state of each model generated by this analysis. progress_bar : bool, optional Whether to show progress bars or not Returns ------- sol : Solution Object Solution to the analysis being performed states : objects, only when return_state == True States generated by the analysis Examples -------- Run a model with the analysis specified in the original KatScript: >>> import finesse >>> model = finesse.Model() >>> model.parse(''' ... l l1 ... pd P l1.p1.o ... ''') >>> model.run("for(l1.P, [0, 1, 2, 3], print(l1.P))") Or you can run a separate analysis: >>> model.run("noxaxis()") """ from .analysis.actions import Noxaxis, Action curr_analysis = self.analysis if analysis is None: # nothing has been set so use what is in the model or noxaxis if nothing _analysis = self.analysis or Noxaxis() elif isinstance(analysis, str): # Katscript analysis requested self.parse(analysis) _analysis = self.analysis elif isinstance(analysis, Action): _analysis = analysis else: raise ValueError(f"Cannot handle analysis input `{analysis}`") rtn = _analysis._run( self, return_state=return_state, progress_bar=progress_bar, simulation_type=simulation_type, simulation_options=simulation_options, ) self.analysis = curr_analysis if not return_state: s = rtn else: s, states = rtn # Return first actual solution for something rather than base # unless there's multiple children while (type(s) is BaseSolution or type(s) is SeriesSolution) and len( s.children ) == 1: _s = s.children[0] _s.parent = None s.children.clear() s = _s if not return_state: return s else: return s, states
[docs] def get_element(self, value): """Returns an element in this model that matches the requested value. This input can be either a string name or an element that might be owned by another model. In the latter case the `name` attribute will be extracted from the value and a search done in this model for an equivalent element. Parameters ---------- value : [str|`Element`] Name or equivalently named `Element` object to extract from this model. Raises ------ `KeyError` when element cannot be found or `FinesseException` when the input `value` cannot be used. Examples -------- >>> import finesse >>> model = finesse.Model() >>> model.parse(''' ... l l1 P=1 ... l l2 P=2 ... ''') >>> model.get_element('l1'), model.get_element('l2') (<'l1' @ 0x122cf9fa0 (Laser)>, <'l2' @ 0x122cf9f40 (Laser)>) Creating a copy of the model and getting elements will return elements with the same name but different `id`. >>> model2 = model.deepcopy() >>> model2.get_element('l1'), model2.get_element('l2') (<'l1' @ 0x122d35c10 (Laser)>, <'l2' @ 0x122d35df0 (Laser)>) Now if we try and select `l1` using an element from a different model we should get the correct element with name `l1` from that model >>> model.get_element(model2.l1) <'l1' @ 0x122cf9fa0 (Laser)> >>> model2.get_element(model.l1) <'l1' @ 0x122d35c10 (Laser)> """ if isinstance(value, str): return self.elements[value] elif hasattr(value, "name"): return self.elements[value.name] else: raise FinesseException(f"Cannot get an element using the input {value}")
[docs] def get_elements_of_type( self, *element_type: type | str ) -> tuple[ModelElement, ...]: """Extracts elements of a specific type from this model. Parameters ---------- *element_type : type or sequence of types The element type(s) to retrieve. Returns ------- tuple The filtered results. Examples -------- >>> IFO.get_elements_of_type(finesse.components.Mirror) (<'m2' @ 0x7ff81a50b6a0 (Mirror)>, <'m1' @ 0x7ff81a50be48 (Mirror)>) >>> tuple(IFO.get_elements_of_type("Mirror")) (<'m2' @ 0x7ff81a50b6a0 (Mirror)>, <'m1' @ 0x7ff81a50be48 (Mirror)>) >>> IFO.get_elements_of_type(finesse.components.Mirror, finesse.components.Beamsplitter)) (<'m2' @ 0x7ff81a50b6a0 (Mirror)>, <'m1' @ 0x7ff81a50be48 (Mirror)>, <'bs1' @ 0x7ff81a50bf33 (Beamsplitter)>) """ # circular import from finesse.script import KATSPEC element_types = [] for el_type in element_type: if isinstance(el_type, str): element_types.append(KATSPEC.get_element_class(el_type)) elif isinstance(el_type, type): element_types.append(el_type) return tuple( el for el in self.elements.values() if isinstance(el, tuple(element_types)) )
[docs] def get_active_signal_nodes(self): """Returns the electrical and mechanical nodes that are active in a model, i.e. ones that need to be solved for because: * they have both input and output edges * have output edges and is a signal input * is used by a detector * has an input optical edge and some output edge This could be more sophisticated and perhaps use the graph in a more correct way. For example, this will not prune some long line of electrical components connected to a mechanical node that drives some optical field. Or in other words, it will not determine if an input edge has an active node on the other end. Returns ------- tuple of nodes """ active = OrderedSet() for node in self.signal_nodes: # can be some signal gen input # or this node might be used as some output is_active = ( node.has_signal_injection or len(node.used_in_detector_output) > 0 ) if not is_active: full_name = node.full_name # For a node to be potentially active it needs to have # both input and output edges. If not then there will # not be any information flowinging through it, as we're # not injecting, etc. into it either is_active |= ( self.network.in_degree(full_name) > 0 and self.network.out_degree(full_name) > 0 ) if is_active: active.add(node) # add nodes that wires will need if hasattr(self, "wires"): for w in self.wires: active.add(w.nodeA) active.add(w.nodeB) return tuple(active)
def __pre_build_checks(self): # TODO (sjr) Probably need different checks dependent upon what type # of simulation is about to built and executed eventually if not self.is_modal: dets_req_trace = list(filter(lambda d: d.needs_trace, self.detectors)) if dets_req_trace: warn( f"Switching on modes (at maxtem = 0) as the model contains the " f"following detectors which require a modal basis: " f"{[d.name for d in dets_req_trace]}" ) self.modes(maxtem=0) expect_unstable_handle = ("auto", "mask", "abort") unstable_handle = self.sim_trace_config.get("unstable_handling") if unstable_handle not in expect_unstable_handle: raise ValueError( "Expected 'unstable_handling' entry of sim_trace_config to be one of " f"{', '.join(expect_unstable_handle)}; but got {unstable_handle}." ) if self._settings.phase_config.zero_k00: if not np.all(self.homs[0] == [0, 0]): raise RuntimeError( "Model phase configuration is set up to zero HG00 -> HG00 " "coupling coefficients but the HG00 mode is not included!" ) @locked_when_built def _build(self, frequencies=None, simulation_type=None, simulation_options=None): """ NOTE: Consider using :function:`finesse.model.Model.built()` instead of this method. This is the first step required to run a simulation of the model. At this stage the layout and connections of the model must be finalised so that the underlying sparse matrix can be allocated and laid out. If this changes the model must be rebuilt. What this method returns are Simulation objects. These are the matrix representation of the model which are used to generate numerical outputs. The Simulation and Model objects are linked together, changing the Model parameters will result in the Simulation objects using the new values when they are solved. A custom list of carrier frequencies can be specfied to run the simulations with. Audio frequencies will be generated from this list separately by the code. `None` will result in the default algorithm computing the required frequency bins, :func:`.frequency.generate_frequency_list`, which can be called by the user and added to if additional frequency bins are required. Parameters ---------- frequencies : list of :class:`.Symbol` Custom list of carrier frequencies can be specfied to run the simulations with. simulation_options : dict Options for the simulation to run Returns ------- sim : object A single simulation object is returned that should be used to perform simulations with this model. Raises ------ Exception If the model is in a built state already. It must be destroyed first before it can be built again. """ if frequencies: raise NotImplementedError("Custom frequencies not implemented yet") self.__pre_build_checks() self._on_pre_build() nx.freeze(self.network) self.__is_built = True self.__toggle_param_locks(True) if simulation_type is None: from finesse.simulations.sparse.simulation import SparseMatrixSimulation from finesse.simulations.sparse.KLU import KLUSolver simulation_type = SparseMatrixSimulation defaults = { "carrier_solver": KLUSolver, "signal_solver": KLUSolver, } if simulation_options is not None: defaults.update(simulation_options) simulation_options = defaults elif not issubclass(simulation_type, BaseSimulation): raise FinesseException( "A simulation should be based on the BaseSimulation class." ) return simulation_type( self, "sim", simulation_options, )
[docs] @locked_when_built @contextmanager def built(self, simulation_type=None, simulation_options=None): """Context manager for making a simulation to work with. Once the context manager has been closed the simulation object and all its memory will be freed up. Parameters ---------- simulation_options : dict Options for type of simulation to run and its settings Yields ------ BaseSimulation Simulation object to interact with after it has been built. Examples -------- >>> import finesse >>> model = finesse.script.parse(''' ... ... some KatScript ... ... ''') >>> >>> with model.built() as sim: ... # interact with simulation object ... ... """ sim = self._build( simulation_type=simulation_type, simulation_options=simulation_options ) try: sim.__enter__() assert sim is not None # Yield and let the caller do something with the model yield sim finally: sim.__exit__(None, None, None) self.unbuild()
def __node_exists_check(self, node): if not self.network.has_node(node.full_name): raise NodeException("Node {} is not in the model".format(node), node)
[docs] def path(self, from_node, to_node, via_node=None, symbolic=False): """Retrieves an ordered container of the path trace between the specified nodes. The return type is an :class:`.OpticalPath` instance which stores an underlying list of the path data (see the documentation for :class:`.OpticalPath` for details). Parameters ---------- from_node : :class:`.Node` Node to trace from. to_node : :class:`.Node` Node to trace to. via_node : :class:`.Node` (or sequence of) Node(s) to traverse via in the path. symbolic : bool, optional Whether to make a symbolic calculation of path lengths Returns ------- out : :class:`.OpticalPath` A container of the nodes and components between `from_node` and `to_node` in order. Raises ------ e1 : :class:`.NodeException` If either of `from_node`, `to_node` are not contained within the model. e2 : :obj:`networkx.NetworkXNoPath` If no path can be found between `from_node` and `to_node`. """ from_node, to_node, via_node = self.__parse_path_nodes( from_node, to_node, via_node ) try: self.__node_exists_check(from_node) self.__node_exists_check(to_node) except NodeException: raise if (from_node is to_node) and (via_node is None): return OpticalPath([]) from_node = from_node.full_name to_node = to_node.full_name if via_node is None: try: path = nx.shortest_path(self.optical_network, from_node, to_node) except nx.exception.NetworkXNoPath: raise path_between = [] for node in path: node_ref = self.network.nodes[node]["weakref"]() path_between.append((node_ref, node_ref.component)) fullpath = [] for node, comp in path_between: if node.is_input: fullpath.append((node, comp)) else: fullpath.append((node, node.space)) return OpticalPath(fullpath, symbolic=symbolic) if not is_iterable(via_node): via_node = [via_node] via_node = [via.full_name for via in via_node] try: paths = [nx.shortest_path(self.optical_network, from_node, via_node[0])] # find intermediate paths for i, node in enumerate(via_node[:-1]): paths.append( nx.shortest_path(self.optical_network, node, via_node[i + 1]) ) paths.append(nx.shortest_path(self.optical_network, via_node[-1], to_node)) except nx.exception.NetworkXNoPath: raise path_between = [ ( self.network.nodes[node]["weakref"](), self.network.nodes[node]["weakref"]().component, ) for node in paths[0] ] for path in paths[1:]: path_between += [ ( self.network.nodes[node]["weakref"](), self.network.nodes[node]["weakref"]().component, ) for node in path ][ 1: ] # don't want to repeat via node fullpath = [] for node, comp in path_between: if node.is_input: fullpath.append((node, comp)) else: fullpath.append((node, node.space)) return OpticalPath(fullpath)
[docs] def sub_model(self, from_node, to_node): """Obtains a subgraph of the complete configuration graph between the two specified nodes. Parameters ---------- from_node : :class:`.Node` Node to trace from. to_node : :class:`.Node` Node to trace to. Returns ------- G : ``networkx.graphviews.SubDiGraph`` A SubGraph view of the subgraph between `from_node` and `to_node`. """ try: self.__node_exists_check(from_node) self.__node_exists_check(to_node) except NodeException: raise nodes_between_set = { node for path in nx.all_simple_paths( self.__network, source=from_node, target=to_node ) for node in path } return self.__network.subgraph(nodes_between_set)
[docs] def component_tree( self, root: ModelElement | str | None = None, network_type: str | NetworkType = NetworkType.COMPONENT, show_detectors: bool = False, show_ports: bool = False, radius: int | None = None, directed: bool = False, ) -> TreeNode: """Retrieves a tree containing the network representing the components connected to the specified root. See also :ref:`model_visualization`. Parameters ---------- root : str | ModelElement | Node | None, optional Root element/node to start drawing the tree from. Is also used in combination with ``radius``, for distance based filtering, by default ``None``. When ``network_type`` is ``components``, ``None`` will default to the first :class:`.finesse.components.laser.Laser` found in the model. network_type : str | NetworkType, optional Which network to plot, can be one of 'optical', 'component', 'full', by default NetworkType.COMPONENT show_detectors : bool, optional Whether to add detectors to the graph, by default False show_ports : bool, optional Whether to show by which ports components are connected, by default False radius : int >= 1 | None, optional Must be used in combination with ``root``, only include nodes of the network that are ``radius`` or less edges away from the root node, by default None, meaning that no nodes are filtered out. directed : bool, optional Whether to use directed distance-based filtering, by default False. If set to True, will only include outgoing edges from the root node. See also the ``undirected`` argument of :func:`networkx.generators.ego.ego_graph` Returns ------- :class:`.TreeNode` The root tree node, with connected components as branches. Raises ------ ValueError If the specified root is not a model element. Notes ----- Uses :func:`networkx.generators.ego.ego_graph` for the distance-based filtering. """ if NetworkType(network_type) is not NetworkType.COMPONENT and root is None: raise ValueError( "Need to specify a root if `network_type` is not " f"'{NetworkType.COMPONENT}'" ) network_filter = NetworkType(network_type).filter_class( model=self, root=root, radius=radius, undirected=not directed, add_edge_info=show_ports, add_detectors=show_detectors, ) return TreeNode.from_network(network_filter.run(), network_filter.root_str)
# Method name is a little mis-leading here but anyway... def _update_symbolic_abcds(self): """Updates the *numeric* ABCD matrices of all components which have geometric parameters which contain symbolic references to other parameters.""" # TODO (sjr) Replace with a loop over just pre-determined geometric parameters # for efficiency, probably not a big deal currently though (I hope) done = set() for el in self.__elements.values(): if el in done: continue for p in el.parameters: # TODO (jwp): determine if is_symbolic is being updated correctly if p.is_geometric and p.is_symbolic: el._re_eval_abcds() done.add(el)
[docs] def detect_mismatches(self, ignore_AR=True, **kwargs): """Detect the mode mismatches in the model. If you want to display these mismatches in a nicely formatted table then use :meth:`Model.mismatches_table`. Parameters ---------- ignore_AR : bool, optional When True, with surfaces with R=0 the reflection mismatch is ignored kwargs : Keyword arguments Arguments to pass to :meth:`Model.beam_trace`. Returns ------- mismatches : dict A dictionary of `(n1, n2): mms` where `n1` is the From node and `n2` is the To node. The value `mms` is another dict consisting of ``"x"`` and ``"y"`` keys mapping to the mismatch values for the tangential and sagittal planes, respectively. """ # Perform the beam trace first so that trace_forest # gets updated if needs be trace = self.beam_trace(store=False, **kwargs) mismatches = {} couplings = self.trace_forest.find_potential_mismatch_couplings() # No mode mismatch couplings present if not couplings: return mismatches for node1, node2 in sorted(couplings, key=lambda npair: npair[0].full_name): component = node1.component if ( ignore_AR and component.interaction_type(node1, node2) == InteractionType.REFLECTION ): if hasattr(component, "R") and component.R.value == 0: continue qx1, qy1 = trace[node1] qx2, qy2 = trace[node2] nr1 = refractive_index(node1) nr2 = refractive_index(node2) # Both nodes attached to same component by definition comp = node1.component Mx = comp.ABCD(node1, node2, direction="x") My = comp.ABCD(node1, node2, direction="y") qx1_matched = transform_beam_param(Mx, qx1, nr1, nr2) qy1_matched = transform_beam_param(My, qy1, nr1, nr2) mismatches[(node1, node2)] = {} if qx1_matched != qx2: mismatches[(node1, node2)]["x"] = BeamParam.mismatch(qx1_matched, qx2) if qy1_matched != qy2: mismatches[(node1, node2)]["y"] = BeamParam.mismatch(qy1_matched, qy2) # Remove the entry if the mismatch values in both planes were zero if not mismatches[(node1, node2)]: del mismatches[(node1, node2)] return mismatches
[docs] def mismatches_table(self, ignore_AR=True, numfmt="{:.4f}", **kwargs): """Prints the mismatches computed by :meth:`Model.detect_mismatches` to an easily readable table. Parameters ---------- ignore_AR : bool, optional When True, with surfaces with R=0 the reflection mismatch is ignored numfmt : str or func or array, optional Either a function to format numbers or a formatting string. The function must return a string. Can also be an array with one option per row, column or cell. Defaults to "{:.4f}". kwargs : Keyword arguments Arguments to pass to :meth:`Model.beam_trace`. """ headerrow = ["Coupling", "Mismatch (x)", "Mismatch (y)"] rownames = [] table = [] mismatches = self.detect_mismatches(ignore_AR=ignore_AR, **kwargs) for (n1, n2), mm_values in mismatches.items(): mmx = mm_values.get("x", 0) mmy = mm_values.get("y", 0) table.append([mmx, mmy]) rownames.append(f"{n1.full_name} -> {n2.full_name}") return NumberTable(table, headerrow, rownames, numfmt=numfmt)
[docs] def compute_space_gouys(self, deg=True, **kwargs): """Calculate the Gouy phases accumulated over each space. If you want to display these phases in a nicely formatted table then use :meth:`Model.space_gouys_table`. Parameters ---------- deg : bool, optional; default = True Whether to convert each phase to degrees. kwargs : Keyword arguments Arguments to pass to :meth:`Model.beam_trace`. Returns ------- gouys : dict A dictionary of `space: gouy` where `space` is a :class:`.Space` object and `gouy` is another dict consisting of ``"x"`` and ``"y"`` keys mapping to the Gouy phase values for the tangential and sagittal planes, respectively. """ trace = self.beam_trace(store=False, **kwargs) scale_func = math.degrees if deg else lambda x: x gouys = {} for space in self.spaces: qx_p1o, qy_p1o = trace[space.p1.i] qx_p2i, qy_p2i = trace[space.p2.o] gouys[space] = { "x": scale_func(abs(qx_p2i.psi - qx_p1o.psi)), "y": scale_func(abs(qy_p2i.psi - qy_p1o.psi)), } return gouys
[docs] def space_gouys_table(self, deg=True, numfmt="{:.4f}", **kwargs): """Prints the space Gouy phases, as computed by :meth:`Model.compute_space_gouys`, to an easily readable table. Parameters ---------- deg : bool, optional; default = True Whether to convert each phase to degrees. fmtber_format : str or func or array, optional Either a function to format numbers or a formatting string. The function must return a string. Can also be an array with one option per row, column or cell. Defaults to "{:.4f}". kwargs : Keyword arguments Arguments to pass to :meth:`Model.beam_trace`. """ units = "deg" if deg else "rad" headers = ["Space", f"Gouy (x) [{units}]", f"Gouy (y) [{units}]"] table = [] rownames = [] gouys = self.compute_space_gouys(deg=deg, **kwargs) for space, gouy_d in gouys.items(): gouy_x = gouy_d.get("x", 0) gouy_y = gouy_d.get("y", 0) table.append([gouy_x, gouy_y]) rownames.append(f"{space.name}") return NumberTable(table, headers, rownames, numfmt=numfmt)
[docs] @locked_when_built def create_mismatch(self, node, w0_mm=0, z_mm=0): """Sets the beam parameters such that a mismatch of the specified percentage magnitude (in terms of :math:`w_0` and :math:`z`) exists at the given `node`. Parameters ---------- node : :class:`.OpticalNode` The node to to create the mismatch at. w0_mm : float or sequence, optional The percentage magnitude of the mismatch in the waist size. This can also be a two-element sequence specifying the waist size mismatches for an astigmatic beam. Defaults to zero percent for both planes. z_mm : float or sequence, optional The percentage magntiude of the mismatch in the distance to the waist. This can also be a two-element sequence specifying the distance to waist mismatches for an astigmatic beam. Defaults to zero percent for both planes. Returns ------- gauss : :class:`.Gauss` The Gauss object created or modified via this mismatch. """ if not self.__network.has_node(node.full_name): raise NodeException("Specified node does not exist within the model", node) trace = self.beam_trace(store=False) qx, qy = trace[node] qx_w0, qy_w0 = qx.w0, qy.w0 qx_z, qy_z = qx.z, qy.z if not is_iterable(w0_mm): qx_w0 *= 1 + (w0_mm / 100) qy_w0 *= 1 + (w0_mm / 100) else: qx_w0 *= 1 + (w0_mm[0] / 100) qy_w0 *= 1 + (w0_mm[1] / 100) if not is_iterable(z_mm): qx_z *= 1 + (z_mm / 100) qy_z *= 1 + (z_mm / 100) else: qx_z *= 1 + (z_mm[0] / 100) qy_z *= 1 + (z_mm[1] / 100) nr = refractive_index(node) qx_bp = BeamParam(w0=qx_w0, z=qx_z, nr=nr) qy_bp = BeamParam(w0=qy_w0, z=qy_z, nr=nr) if node in self.__gausses: self.update_gauss(node, qx_bp, qy_bp) else: name = f"AUTO_MM_{node.component.name}" self.add(Gauss(name, node, qx=qx_bp, qy=qy_bp)) return self.__gausses.get(node)
[docs] @locked_when_built def add_matched_gauss(self, node, name=None, priority=0, matched_to=None): """Adds a :class:`.Gauss` object mode matched to the model at the specified `node`. If `matched_to` is given then the Gauss object will be matched to this, otherwise it will be mode matched to whichever dependency the `node` relies upon currently. Parameters ---------- node :class:`.OpticalNode` The node instance to add the Gauss at. name : str, optional Optional name of the new Gauss object. If not specified then the name will be automatically given as "AUTO_MM_nodename" where `nodename` is the node's :attr:`full name <.Node.full_name>`. priority : int, optional; default: 0 The priority value of the Gauss object. See :attr:`.TraceDependency.priority` for details on the argument. matched_to : :class:`.TraceDependency`, optional The trace dependency to solely match to. """ if not self.__network.has_node(node.full_name): raise NodeException("Specified node does not exist within the model", node) if node in self.__gausses: raise ValueError(f"A Gauss object already exists at node {node.full_name}") if matched_to is not None and not isinstance( matched_to, (TraceDependency, str) ): raise ValueError("Expected matched_to arg to be of type TraceDependency.") trace = self.beam_trace(enable_only=matched_to, store=False) qx, qy = trace[node] if name is None: name = f"AUTO_MM_{node.component.name}" self.add(Gauss(name, node, qx=qx, qy=qy, priority=priority)) return self.__gausses.get(node)
[docs] def update_gauss(self, node, qx=None, qy=None): """Update the value of a manual beam parameter (i.e. :class:`.Gauss` object) at the specified `node`. Parameters ---------- node : :class:`.OpticalNode` The node instance to update the gauss at. qx : :class:`.BeamParam` or complex, optional Beam parameter in tangential plane. qy : :class:`.BeamParam` or complex, optional Beam parameter in sagittal plane. """ if not isinstance(node, OpticalNode): raise TypeError( f"Expected argument 'node' to be of type OpticalNode " f"but got value of type {type(node)}" ) gauss = self.__gausses.get(node) if gauss is None: raise ValueError( f"Node {node.full_name} has no Gauss object associated with it." ) if qx is None and qy is None: warn("Model.update_gauss called with no change to qx, qy.") if qx is not None: self.__gausses[node].qx = qx if qy is not None: self.__gausses[node].qy = qy
[docs] def cavity_mismatch(self, cav1=None, cav2=None): """See :func:`finesse.tracing.tools.compute_cavity_mismatches`""" return tracetools.compute_cavity_mismatches(self, cav1, cav2)
[docs] def cavity_mismatches_table(self, direction=None, percent=False, numfmt="{:.2e}"): """Prints the mismatches between each cavity in an easily readable table format. If either of each cavity in a coupling is unstable then the mismatch values between these will be displayed as ``nan``. Parameters ---------- direction : str, optional; default: None The plane to compute mismatches in, "x" for tangential, "y" for sagittal. If not given then tables for both planes will be printed. percent : bool, optional; default: False Whether mismatch values are displayed in terms of percentage. Defaults to False such that the values are given as fractional mismatches. numfmt : str or func or array, optional Either a function to format numbers or a formatting string. The function must return a string. Can also be an array with one option per row, column or cell. Defaults to "{:.4f}". """ if not self.cavities: warn("No cavities present in the model.") return mmx, mmy = self.cavity_mismatch() headers = [cav.name for cav in self.cavities] names_x = [] names_y = [] table_x = [] table_y = [] scale = 100 if percent else 1 for cav1 in self.cavities: names_x.append(cav1.name) names_y.append(cav1.name) row_x = [] row_y = [] for cav2 in self.cavities: row_x.append(scale * mmx[(cav1.name, cav2.name)]) row_y.append(scale * mmy[(cav1.name, cav2.name)]) table_x.append(row_x) table_y.append(row_y) if direction is None: direction = "x", "y" result = [] if "x" in direction: headers_x = headers.copy() headers_x.insert(0, "(Tangential plane)") result.append(NumberTable(table_x, headers_x, names_x, numfmt=numfmt)) if "y" in direction: headers_y = headers.copy() headers_y.insert(0, " (Sagittal plane) ") result.append(NumberTable(table_y, headers_y, names_y, numfmt=numfmt)) return result
def __optical_node_from_str(self, name, dir_if_port="o", reject_ports=False): args = name.split(".") is_port = False if len(args) == 2: # name is a port string if reject_ports: raise ValueError( f"Via port {name} is ambiguous, cannot safely assume node " "direction. Please specify the full node string instead." ) name += "." + dir_if_port is_port = True else: if len(args) != 3: raise ValueError(f"Unexpected port or node name format: {name}") # return self.network.nodes[name]["weakref"]() # Issue 388: get full name node from model instead which avoids the issue # of borrowed nodes full names not being stored in the model graph if is_port: return getattr(self.get(name).port, dir_if_port) else: return self.get(name) def __parse_path_nodes(self, from_node, to_node, via_node): if from_node is not None and not isinstance(from_node, OpticalNode): if isinstance(from_node, Port): # Issue 388: bit hacky but to get around the borrowed node stuff # as in the port here might be a space port which has the opposite # i/o nodes to the port of an element which actually owns the nodes # we get the original port and the output of that... from_node = from_node.o.port.o elif isinstance(from_node, str): from_node = self.__optical_node_from_str(from_node, dir_if_port="o") else: raise TypeError("Unexpected type for from_node.") if to_node is not None and not isinstance(to_node, OpticalNode): if isinstance(to_node, Port): to_node = to_node.i.port.i # Issue 388 hack elif isinstance(to_node, str): to_node = self.__optical_node_from_str(to_node, dir_if_port="i") else: raise TypeError("Unexpected type for to_node.") if via_node is not None and not isinstance(via_node, OpticalNode): if isinstance(via_node, Port): raise ValueError( f"Via port {via_node} is ambiguous, cannot safely assume node " "direction. Please specify the optical node instead." ) elif isinstance(via_node, str): via_node = self.__optical_node_from_str(via_node, reject_ports=True) else: raise TypeError("Unexpected type for via_node.") return from_node, to_node, via_node
[docs] def ABCD( self, from_node=None, to_node=None, via_node=None, path=None, direction="x", symbolic=False, solution_name=None, ): """See :func:`finesse.tracing.tools.compute_abcd` .. note:: The only difference to the above function is that any of `from_node`, `to_node`, `via_node` can be specified as strings. """ fn, tn, vn = self.__parse_path_nodes(from_node, to_node, via_node) return tracetools.compute_abcd( fn, tn, vn, path, direction, symbolic, solution_name )
[docs] def acc_gouy( self, from_node=None, to_node=None, via_node=None, path=None, q_in=None, direction="x", symbolic=False, degrees=True, **kwargs, ): """See :func:`finesse.tracing.tools.acc_gouy` .. note:: The only difference to the above function is that any of `from_node`, `to_node`, `via_node` can be specified as strings. """ fn, tn, vn = self.__parse_path_nodes(from_node, to_node, via_node) return tracetools.acc_gouy( fn, tn, vn, path, q_in, direction, symbolic, degrees, **kwargs, )
[docs] def propagate_beam( self, from_node=None, to_node=None, via_node=None, path=None, q_in=None, direction="x", symbolic=False, simplify=False, solution_name=None, **kwargs, ): """See :func:`finesse.tracing.tools.propagate_beam` .. note:: The only difference to the above function is that any of `from_node`, `to_node`, `via_node` can be specified as strings. """ fn, tn, vn = self.__parse_path_nodes(from_node, to_node, via_node) return tracetools.propagate_beam( fn, tn, vn, path, q_in, direction, symbolic, simplify, solution_name, **kwargs, )
[docs] def propagate_beam_astig( self, from_node=None, to_node=None, via_node=None, path=None, qx_in=None, qy_in=None, symbolic=False, solution_name=None, **kwargs, ): """See :func:`finesse.tracing.tools.propagate_beam_astig` .. note:: The only difference to the above function is that any of `from_node`, `to_node`, `via_node` can be specified as strings. """ fn, tn, vn = self.__parse_path_nodes(from_node, to_node, via_node) return tracetools.propagate_beam_astig( fn, tn, vn, path, qx_in, qy_in, symbolic, solution_name, **kwargs, )
[docs] def beam_trace( self, order=None, disable=None, enable_only=None, symmetric=True, store=True, solution_name="beam_trace", ): """Performs a full beam trace on the model, calculating the beam parameters at each optical node. Beam tracing requires at least one stable :class:`.TraceDependency` object - defined as a :class:`.Gauss` or :class:`.Cavity` object - in the model. Note that :class:`.Cavity` objects are not determined automatically, they must be explicitly added to the model. The order in which the beam tracing is performed is as follows: * All internal cavity traces are carried out initially, i.e. any nodes that are part of a path of a :class:`.Cavity` instance in the model will be traced using the cavity eigenmode as the basis. See the note below for what happens in the case of overlapping cavities. * If `order` is **not** specified, then the dependency ordering given by :attr:`.Model.trace_order` is used; i.e. beam traces are performed from each dependency in this list where any overlapping trees from multiple dependencies will always use the first dependency's tree. * If `order` **is** specified, then beam traces from each dependency given in this argument will be carried out in the given order, in the same way as above. This effectively allows temporary overriding of the trace order for specific beam trace calls. Certain dependencies can be switched off via the `disable` or `enable_only` arguments. Full details on the beam tracing algorithm are given in :ref:`tracing_manual`. .. note:: .. rubric:: Overlapping cavities In complicated configurations, such as dual-recycled Michelson interferometers, it is often the case that there will be *overlapping* cavities; i.e. cavities which share common optical nodes in their paths. This naturally leads to the question - in which basis are these common nodes being set? The algorithm used by this method will prioritise the internal trace of the cavity as follows: * If `order` is **not** given then the cavity appearing first in :attr:`.Model.trace_order` will be used for setting the beam parameters of all nodes in this cavity path - *including* any nodes which are shared with other cavity instances. * Otherwise, the cavity appearing first in `order` will be used in the same way. Parameters ---------- order : sequence, optional; default: None A priority list of dependencies to trace in order. These dependencies can be either :class:`.Cavity` / :class:`.Gauss` objects or the names of these objects. If this argument is not specified then beam tracing will be performed using the order defined in :attr:`.Model.trace_order`. This argument allows temporary overriding of the model trace order for a given beam trace call, without needing to change the :attr:`.TraceDependency.priority` values of any dependencies in the model (which would affect all future beam traces in which `order` is not specified). Note that, if specifying this argument, a sub-set of the trace dependencies in the model can be given --- any dependencies not in the given `order` will retain their original ordering. disable : sequence, optional: default: None A single dependency or list of dependencies to disable. These dependencies can be either :class:`.Cavity` / :class:`.Gauss` objects or the names of these objects. Note that this argument is ignored if `enable_only` is specified. enable_only : sequence, optional: default: None A single dependency or list of dependencies to explicity enable, all other dependencies will be switched off for the trace call. These dependencies can be either :class:`.Cavity` / :class:`.Gauss` objects or the names of these objects. symmetric : bool, optional; default: true Flag to determine whether the beam parameters at :attr:`.OpticalNode.opposite` nodes of each node encountered during the beam trace get set to the :meth:`.BeamParam.reverse` of the "forward" propagated beam parameter. store : bool, optional; default: True Flag to determine whether to store the results of a beam trace in in the :attr:`Model.last_trace` property. Note that if this is set to False then accessing beam parameters via :class:`.OpticalNode` instances directly will give the *last stored* beam parameter at that node (i.e. not those given by this trace call). Raises ------ ex : :class:`.BeamTraceException` If there are no :class:`.Gauss` objects or stable :class:`.Cavity` instances present in the model. ex_v: :class:`.ValueError` If `order` was specified and it contains an invalid item. ex_tr : :class:`.TotalReflectionError` If a :class:`.Beamsplitter` object is present in the model with an angle of incidence and associated refractive indices giving total internal reflection. Returns ------- out : :class:`.BeamTraceSolution` An object representing the results of the tracing routine. """ LOGGER.info("Beam trace triggered on %s", self) # No override given if order is None: order = self.trace_order # Override to model trace_order given, so verify and make it else: order = self.__make_trace_order_override(order) # No cavities nor gausses present if not order: raise BeamTraceException(_invalid_trace_reason()) Norder = len(order) Ndeps = len(self.cavities) + len(self.gausses) if Norder != Ndeps: raise BeamTraceException( f"Bug encountered! Length of trace order dependency list ({Norder}) " f"not equal to number of tracing dependencies ({Ndeps})." ) self._update_symbolic_abcds() # Remove specific dependencies from order if disable is not None and enable_only is None: if not is_iterable(disable): disable = [disable] for item in disable: if isinstance(item, str): item_obj = self.__elements.get(item) else: item_obj = item if isinstance(item_obj, TraceDependency): try: order.remove(item_obj) except ValueError: LOGGER.info( f"Ignoring duplicate item {repr(item)} in disable sequence" ) else: LOGGER.info( f"ignoring unrecognised item {repr(item)} in disable sequence" ) # Filter order to just contain dependencies in enable_only if enable_only is not None: if disable is not None: warn("Ignoring argument 'disable' as 'enable_only' has been given.") if not is_iterable(enable_only): enable_only = [enable_only] else: enable_only = list(enable_only) for i, item in enumerate(enable_only): if isinstance(item, str): item_obj = self.__elements.get(item) else: item_obj = item if not isinstance(item_obj, TraceDependency): warn( f"ignoring unrecognised item {repr(item_obj)} in enable_only " f"sequence" ) enable_only[i] = item_obj order = list(filter(lambda x: x in enable_only, order)) all_cavities = list(filter(lambda x: isinstance(x, Cavity), order)) stable_cavities = list(filter(lambda x: x.is_stable, all_cavities)) gausses = list(filter(lambda x: isinstance(x, Gauss), order)) if not stable_cavities and not gausses: raise BeamTraceException(_invalid_trace_reason(all_cavities)) # Ensure only gausses and *stable* cavities get passed to trace forest planting order = list(filter(lambda x: x in stable_cavities or x in gausses, order)) unstable_cavities = OrderedSet(all_cavities).difference(stable_cavities) if unstable_cavities: warn( f"The cavities {[uc.name for uc in unstable_cavities]} are unstable " f"and will not be used for beam tracing.", CavityUnstableWarning, ) # Save the planet! Here we only clear and re-plant the forest if: # - a component has been added / removed from the model since last # call (taken into account in Model.add, Model.remove already) # - the symmetric flag has changed since the last plant # - or the tracing priority has changed self._rebuild_trace_forest |= ( self.trace_forest.symmetric != symmetric or self.trace_forest.dependencies != order ) if self._rebuild_trace_forest: LOGGER.info( "Re-planting the model trace forest with dependency order: %s", [d.name for d in order], ) self.trace_forest.symmetric = symmetric self.trace_forest.plant(order) LOGGER.info("Planted the full model trace forest: %s", self.trace_forest) self._rebuild_trace_forest = False trace_results = self.trace_forest.trace_beam() trace = BeamTraceSolution(solution_name, trace_results) LOGGER.debug("\n%s", trace) if store: self.__last_trace = trace return trace
def __make_trace_order_override(self, override): """Construct an override list for the current trace order.""" if not is_iterable(override): override = [override] order = [] # For all the items in value, append these in # order to the new dependency order list for item in override: if isinstance(item, str): dep = self.__elements.get(item) if dep is None: raise ValueError( f"Error in tracing order override:\n" f" No item of name {item} exists in the model." ) else: dep = item if isinstance(dep, Cavity): if dep not in self.__cavities: raise ValueError( f"Error in tracing order override:\n" f" Cavity {dep.name} is not part of the model." ) elif isinstance(dep, Gauss): if dep not in self.__gausses.values(): raise ValueError( f"Error in tracing order override:\n" f" Gauss {dep.name} is not part of the model." ) else: raise ValueError( f"Error in tracing order override:\n" f" Item {dep} is neither a Cavity nor a Gauss object." ) if dep in order: LOGGER.info( f"ignoring duplicate dependency {repr(dep.name)} in tracing order " f"override" ) else: order.append(dep) # Take all the remaining dependencies that were not in override and # append these to the end of the dependency order list # -> this means that the original order of those dependencies is # retained, the only ordering that has changed is those that # were included in the override for item in self.trace_order: if item not in order: order.append(item) return order
def _invalid_trace_reason(cavities=None): msg = """Unable to perform a beam trace! - No manually set beam parameters have been specified. """ def gfactor_str(cavity): gx, gy = cavity.g if cavity.gx == cavity.gy: return f"g = {gx}" return f"gx = {gx}, gy = {gy}" if cavities is None: cavities = [] crit = ", ".join( [f"{cav.name}: {gfactor_str(cav)}" for cav in cavities if cav.is_critical] ) unst = ", ".join( [ f"{cav.name}: {gfactor_str(cav)}" for cav in cavities if not cav.is_stable and not cav.is_critical ] ) if cavities: msg += f""" - All cavities included in the model are unstable / critical. - Critical cavities -- {crit} - Unstable cavities -- {unst} """ else: msg += """ - No cavities are present in the model. Note that cavity objects must be explicitly added, they are NOT auto-generated from your configuration. """ return msg