engforge.eng.geometry.ShapelySection

class ShapelySection(*, name='shapely section', shape, coarse=False, mesh_extent_decimation=100, min_mesh_angle=20, min_mesh_size=1e-05, goal_elements=1000, material=None, failure_mode='von_mises', prediction=False, prediction_goal_error=0.025, max_records=10000)

Bases: Profile2D

a 2D profile that takes a shapely section to calculate section properties, use a sectionproperties section with hidden variable _geo to bypass shape calculation

Method generated by attrs for class ShapelySection.

Methods

add_fields	Overwrite this to modify logging fields
add_prediction_record	adds a record to the prediction records, and calcultes the average and variance of the data :type record: :param record: a dict of the record :type extra_add: :param extra_add: if true, the record is added to the prediction records even if the data is inbounds :returns: a boolean indicating if the record was out of bounds of current data (therefor should be added)
basis_expand	run combinations of parameters and permutations of weights against the basis values to populate the stress records, by default using estimation logic to speed up the process
calculate_bounds
calculate_mesh_size
calculate_stress
change_all_log_lvl
check_and_retrain	Checks if more data than threshold to train or if error is sufficiently low to ignore retraining, or if more data already exists than window size (no training)
check_out_of_domain	checks if the record is in bounds of the current data
check_ref_slot_type	recursively checks class slots for the key, and returns the slot type
check_symmetric	checks if the section is symmetric about the x and y axis, by finding the intersection of
cls_compile	compiles this class, override this to compile functionality for this class
collect_all_attributes	collects all the attributes for a system
collect_inst_attributes	collects all the attributes for a system
compile_classes	compiles all subclass functionality
copy_config_at_state	copy the system at the current state recrusively to a certain level, by default copying everything :type levels_deep: int :param levels_deep: how many levels deep to copy, -1 is all :type level: :param level: the current level, defaults to 0 if not set
critical	A routine to communicate to the root of the server network that there is an issue
debug	Writes at a low level to the log file.
determine_failure_front
determine_failure_stress	uses the failure mode to compare to allowable stress
difference	a context manager that will allow you to dynamically change any information, then will change it back in a fail safe way.
display_results
error	Writes to log as a error
estimate_failure	uses a support vector machine to estimate stresses and returns a number zero or one to indicate failure, if prediction is set to True, otherwise calculates stress
estimate_stress	uses a support vector machine to estimate stresses and returns the ratio of the allowable stress, also known as the failure fracion if prediction is set to True, otherwise calculates stress
extract_message
fail_frac_criteria
fail_learning	optimizes stress until failure, given base_kw arguments for extra parameters
filter	This acts as the interface for logging.Filter Don't overwrite this, use add_fields instead.
from_cache
go_through_configurations	A generator that will go through all internal configurations up to a certain level if levels_to_descend is less than 0 ie(-1) it will go down, if it 0, None, or False it will only go through this configuration
hash	hash by parm or by input_kw, only input can be hashed by lookup as system properties can create a recursive loop and should be deterministic from input
hash_id	string for saving to persisting
hash_numeric_with
hash_with	Generates a hash for the object's dictionary representation, updated with additional keyword arguments.
info	Writes to log but with info category, these are important typically and inform about progress of process in general
init_with_material
input_attrs	Lists all input attributes for class
input_fields	no attr base types, no tuples, no lists, no dicts
installSTDLogger	We only want std logging to start
internal_configurations	go through all attributes determining which are configuration objects additionally this skip any configuration that start with an underscore (private variable)
mesh_section	caches section properties and mesh
message_with_identiy	converts to color and string via the termcolor library :type message: str :param message: a string convertable entity :type color: :param color: a color in [grey,red,green,yellow,blue,magenta,cyan,white]
msg	Writes to log.
numeric_fields	no tuples,lists, dicts, strings, or attr base types
observe_and_predict	uses the existing models to predict the row and measure the error
parent_configurations_cls	returns all subclasses that are a Configuration
plot_attributes	Lists all plot attributes for class
plot_mesh
pre_compile	an overrideable classmethod that executes when compiled, however will not execute as a subclass
prediction_dataframe
prediction_weights
random_force_input
record_stress	determines if stress record should be added to prediction_records
resetLog	reset log
resetSystemLogs	resets log on all internal instance LoggingMixins
reset_prediction
score_data	scores a dataframe
setattrs	sets attributes from a dictionary
signals_attributes	Lists all signals attributes for class
slack_notification
slot_refs	returns all slot references in this configuration
slots_attributes	Lists all slots attributes for class
solve_fail
solvers_attributes	Lists all signals attributes for class
subclasses	return all subclasses of components, including their subclasses :type out: :param out: out is to pass when the middle of a recursive operation, do not use it!
subcls_compile	reliably compiles this method even for subclasses, override this to compile functionality for subclass interfaces & mixins
table_fields	the table attributes corresponding to
trace_attributes	Lists all trace attributes for class
train_compare	Use the dataframe to train the models, and compare the results to the current models using train_frac to divide total samples into training and testing sets, unless train_full is set.
train_until_valid	trains the prediction models until the error is below the goal error
training_callback	when training is complete save the model to a pickle with ShapelySection_<hash>.pkl
transients_attributes	Lists all signals attributes for class
validate_class	A customizeable validator at the end of class creation in forge
warning	Writes to log as a warning

Attributes

A
Ao	outside area, over ride for hallow sections
Ixx
Ixy
Iyy
J
as_dict	returns values as they are in the class instance
attrs_fields
basis
cache_name
cache_path
classname	Shorthand for the classname
displayname
filename	A nice to have, good to override
identity	A customizeable property that will be in the log by default
input_as_dict	returns values as they are in the class instance, but converts classes inputs to their input_as_dict
input_hash
log_fmt
log_level
log_on
log_silo
logger
max_margin
max_rec_parm
mesh_size
meta_name
meta_path
min_mesh_area
near_margin
numeric_as_dict	recursively gets internal components numeric_as_dict as well as its own numeric values
numeric_hash
prediction_goal_error
prediction_records
save_threshold
section_cache
slack_webhook_url
table_hash
table_row_dict	returns values as they would be put in a table row from this instance ignoring any sub components
train_window
trained
unique_hash
x_bounds
y_bounds
name
shape
coarse
min_mesh_angle
min_mesh_size
goal_elements
material
prediction
max_records

Parameters:

• name (str)

• shape (Polygon)

• coarse (bool)

• min_mesh_angle (float)

• min_mesh_size (float)

• goal_elements (float)

• material (Material)

• failure_mode (str)

• prediction (bool)

• prediction_goal_error (float)

• max_records (list)

property Ao

outside area, over ride for hallow sections

add_fields(record)

Overwrite this to modify logging fields

add_prediction_record(record, extra_add=True, mult_sigma=1, target_items=1000)

adds a record to the prediction records, and calcultes the average and variance of the data :type record: :param record: a dict of the record :type extra_add: :param extra_add: if true, the record is added to the prediction records even if the data is inbounds :returns: a boolean indicating if the record was out of bounds of current data (therefor should be added)

property as_dict

returns values as they are in the class instance

basis_expand(expand_values=[0.9, 0.75, 0.5, 0.1, 0.01], Nparm=4, est=True, normalize=True)

run combinations of parameters and permutations of weights against the basis values to populate the stress records, by default using estimation logic to speed up the process

check_and_retrain(records, min_rec=None)

Checks if more data than threshold to train or if error is sufficiently low to ignore retraining, or if more data already exists than window size (no training)

check_out_of_domain(record, extra_margin=1, target_items=1000)

checks if the record is in bounds of the current data

classmethod check_ref_slot_type(sys_key)

recursively checks class slots for the key, and returns the slot type

Return type:

list

Parameters:

sys_key (str)

check_symmetric(precision=3, Nincr=180)

checks if the section is symmetric about the x and y axis, by finding the intersection of

property classname

Shorthand for the classname

classmethod cls_compile()

compiles this class, override this to compile functionality for this class

classmethod collect_all_attributes()

collects all the attributes for a system

collect_inst_attributes(**kw)

collects all the attributes for a system

classmethod compile_classes()

compiles all subclass functionality

copy_config_at_state(level=None, levels_deep=-1, changed=None, **kw)

copy the system at the current state recrusively to a certain level, by default copying everything :type levels_deep: int :param levels_deep: how many levels deep to copy, -1 is all :type level: :param level: the current level, defaults to 0 if not set

Parameters:

• levels_deep (int)

• changed (dict | None)

critical(*args)

A routine to communicate to the root of the server network that there is an issue

debug(*args)

Writes at a low level to the log file… usually this should be detailed messages about what exactly is going on

determine_failure_stress(stress_obj)

uses the failure mode to compare to allowable stress

difference(**kwargs)

a context manager that will allow you to dynamically change any information, then will change it back in a fail safe way.

with self.difference(name=’new_name’, value = new_value) as new_config:

#do stuff with config, ok to fail

you may not access any “private” variable that starts with an _ as in _whatever

difference is useful for saving slight differences in configuration in conjunction with solve you might create wrappers for eval, or implement a strategy pattern.

only attributes may be changed.

#TODO: allow recursive operation with sub comps or systems. #TODO: make a full system copy so the system can be reverted later

error(error, msg='')

Writes to log as a error

estimate_failure(n=0, vx=0, vy=0, mxx=0, myy=0, mzz=0)

uses a support vector machine to estimate stresses and returns a number zero or one to indicate failure, if prediction is set to True, otherwise calculates stress

Return type:

float

estimate_stress(n=0, vx=0, vy=0, mxx=0, myy=0, mzz=0, value=False, calc_margin=2, min_est_records=100, calc_every=25, pre_train_margin=2, force_calc=False)

uses a support vector machine to estimate stresses and returns the ratio of the allowable stress, also known as the failure fracion if prediction is set to True, otherwise calculates stress

Return type:

float

fail_learning(X, parm, base_kw, mult=1)

optimizes stress until failure, given base_kw arguments for extra parameters

property filename

A nice to have, good to override

filter(record)

This acts as the interface for logging.Filter Don’t overwrite this, use add_fields instead.

go_through_configurations(level=0, levels_to_descend=-1, parent_level=0, only_inst=True, **kw)

A generator that will go through all internal configurations up to a certain level if levels_to_descend is less than 0 ie(-1) it will go down, if it 0, None, or False it will only go through this configuration

Returns:

level,config

hash(*args, **input_kw)

hash by parm or by input_kw, only input can be hashed by lookup as system properties can create a recursive loop and should be deterministic from input

hash_id()

string for saving to persisting

Return type:

str

hash_with(**input_kw)

Generates a hash for the object’s dictionary representation, updated with additional keyword arguments. :type **input_kw: :param **input_kw: Arbitrary keyword arguments to update the object’s dictionary representation.

Returns:

The hash value of the updated dictionary.

Raises:

Any exceptions raised by deepdiff.DeepHash if hashing fails. –

property identity

A customizeable property that will be in the log by default

info(*args)

Writes to log but with info category, these are important typically and inform about progress of process in general

property input_as_dict

returns values as they are in the class instance, but converts classes inputs to their input_as_dict

classmethod input_attrs()

Lists all input attributes for class

classmethod input_fields(add_ign_types=None)

no attr base types, no tuples, no lists, no dicts

Parameters:

add_ign_types (list | None)

installSTDLogger()

We only want std logging to start

internal_configurations(check_config=True, use_dict=True, none_ok=False)

go through all attributes determining which are configuration objects additionally this skip any configuration that start with an underscore (private variable)

Return type:

dict

mesh_section()

caches section properties and mesh

message_with_identiy(message, color=None)

converts to color and string via the termcolor library :type message: str :param message: a string convertable entity :type color: :param color: a color in [grey,red,green,yellow,blue,magenta,cyan,white]

Parameters:

message (str)

msg(*args, lvl=5)

Writes to log… this should be for raw data or something… least priorty

property numeric_as_dict

recursively gets internal components numeric_as_dict as well as its own numeric values

classmethod numeric_fields()

no tuples,lists, dicts, strings, or attr base types

observe_and_predict(row)

uses the existing models to predict the row and measure the error

classmethod parent_configurations_cls()

returns all subclasses that are a Configuration

Return type:

list

classmethod plot_attributes()

Lists all plot attributes for class

Return type:

Dict[str, Attribute]

classmethod pre_compile()

an overrideable classmethod that executes when compiled, however will not execute as a subclass

record_stress(stress_dict)

determines if stress record should be added to prediction_records

resetLog()

reset log

resetSystemLogs(reseted=None)

resets log on all internal instance LoggingMixins

score_data(df)

scores a dataframe

setattrs(dict)

sets attributes from a dictionary

classmethod signals_attributes()

Lists all signals attributes for class

Return type:

Dict[str, Attribute]

classmethod slot_refs(recache=False)

returns all slot references in this configuration

classmethod slots_attributes(attr_type=False)

Lists all slots attributes for class

Return type:

Dict[str, Attribute]

classmethod solvers_attributes()

Lists all signals attributes for class

Return type:

Dict[str, Attribute]

classmethod subclasses(out=None)

return all subclasses of components, including their subclasses :type out: :param out: out is to pass when the middle of a recursive operation, do not use it!

classmethod subcls_compile()

reliably compiles this method even for subclasses, override this to compile functionality for subclass interfaces & mixins

classmethod table_fields()

the table attributes corresponding to

property table_row_dict

returns values as they would be put in a table row from this instance ignoring any sub components

classmethod trace_attributes()

Lists all trace attributes for class

Return type:

Dict[str, Attribute]

train_compare(df, test_frac=2, train_full=False, min_rec=250)

Use the dataframe to train the models, and compare the results to the current models using train_frac to divide total samples into training and testing sets, unless train_full is set.

Parameters:

• df – dataframe to train with

• test_frac – N/train_frac will be size of the training window

• train_full – boolean to use full training data

Returns:

trained models

train_until_valid(print_interval=50, max_iter=1000, est=False)

trains the prediction models until the error is below the goal error

training_callback(models)

when training is complete save the model to a pickle with ShapelySection_<hash>.pkl

classmethod transients_attributes()

Lists all signals attributes for class

Return type:

Dict[str, Attribute]

classmethod validate_class()

A customizeable validator at the end of class creation in forge

warning(*args)

Writes to log as a warning