Model#

class pharmpy.model.Model(name='', parameters=Parameters(), random_variables=, statements=, dataset=None, datainfo=Empty DataFrame Columns: [name, type, scale, continuous, categories, unit, drop, datatype, descriptor] Index: [], dependent_variables={y: 1}, observation_transformation=None, estimation_steps=EstimationSteps(), parent_model=None, initial_individual_estimates=None, filename_extension='', value_type='PREDICTION', description='', internals=None)[source]#

Bases: Immutable

The Pharmpy model class

Attributes Summary

`datainfo`	Definitions of model statements
`dataset`	Dataset connected to model
`dependent_variables`	The dependent variables of the model mapped to the corresponding DVIDs
`description`	A free text description of the model
`estimation_steps`	Definitions of estimation steps
`filename_extension`	Filename extension of model file
`initial_individual_estimates`	Initial estimates for individual parameters
`internals`	Internal data for tool specific part of model
`model_code`	Model type specific code
`name`	Name of the model
`observation_transformation`	Transformation to be applied to the observation data
`parameters`	Definitions of population parameters
`parent_model`	Name of parent model
`random_variables`	Definitions of random variables
`statements`	Definitions of model statements
`value_type`	The type of the model value (dependent variables)

Methods Summary

`create`(name[, parameters, random_variables, ...])
`from_dict`(d)
`has_same_dataset_as`(other)	Check if this model has the same dataset as another model
`parse_model`(path)	Create a model object by parsing a model file of any supported type
`parse_model_from_string`(code)	Create a model object by parsing a string with model code of any supported type
`replace`(**kwargs)
`to_dict`()
`update_source`()	Update source code of the model.
`write_files`([path, force])	Write all extra files needed for a specific external format.

Attributes Documentation

datainfo#

Definitions of model statements

See pharmpy.Statements

dataset#: Dataset connected to model

dependent_variables#: The dependent variables of the model mapped to the corresponding DVIDs

description#: A free text description of the model

estimation_steps#

Definitions of estimation steps

See pharmpy.EstimationSteps

filename_extension#: Filename extension of model file

initial_individual_estimates#: Initial estimates for individual parameters

internals#: Internal data for tool specific part of model

model_code#: Model type specific code

name#: Name of the model

observation_transformation#: Transformation to be applied to the observation data

parameters#

Definitions of population parameters

See pharmpy.Parameters

parent_model#: Name of parent model

random_variables#

Definitions of random variables

See pharmpy.RandomVariables

statements#

Definitions of model statements

See pharmpy.Statements

value_type#

The type of the model value (dependent variables)

By default this is set to ‘PREDICTION’ to mean that the model outputs a prediction. It could optionally be set to ‘LIKELIHOOD’ or ‘-2LL’ to let the model output the likelihood or -2*log(likelihood) of the prediction. If set to a symbol this variable can be used to change the type for different records. The model would then set this symbol to 0 for a prediction value, 1 for likelihood and 2 for -2ll.

Methods Documentation

classmethod create(name, parameters=None, random_variables=None, statements=None, dataset=None, datainfo=Empty DataFrame Columns: [name, type, scale, continuous, categories, unit, drop, datatype, descriptor] Index: [], dependent_variables=None, observation_transformation=None, estimation_steps=None, parent_model=None, initial_individual_estimates=None, filename_extension='', value_type='PREDICTION', description='', internals=None)[source]#

classmethod from_dict(d)[source]#

has_same_dataset_as(other)[source]#

Check if this model has the same dataset as another model

Parameters:: other (Model) – Another model
Returns:: bool – True if both models have the same dataset

static parse_model(path)[source]#

Create a model object by parsing a model file of any supported type

Parameters:: path (Path or str) – Path to a model file
Returns:: Model – A model object

static parse_model_from_string(code)[source]#

Create a model object by parsing a string with model code of any supported type

Parameters:: code (str) – Model code
Returns:: Model – A model object

replace(**kwargs)[source]#

to_dict()[source]#

update_source()[source]#: Update source code of the model. If any paths need to be changed or added (e.g. for a NONMEM model with an updated dataset) they will be replaced with DUMMYPATH

write_files(path=None, force=False)[source]#: Write all extra files needed for a specific external format.