Package PyDSTool :: Module ModelConstructor'

Module ModelConstructor'

Model Constructor classes.

   Instantiate abstract Model Specifications into concrete simulation code.

   Robert Clewley, September 2005.


Overview of steps that ModelConstructor takes:

1. Build Generators for each vector field
 * Take flattened spec and select a compatible Generator type and language
   * include check that compatibleGen selection actually exists,
      and is compatible with specType/domain, and with cts/discrete
      time variables)
   * presently, each v.f. must have all the same Gen type
 * Map things like x^y and x**y power specs -> pow(x,y)
    and abs() function -> fabs() in C.

2. Given the specific target lang/Generator with a ModelSpec vector field
 * Associate additional events, code inserts, reused vars
 * Choose which domains map to domain verification events (certainly none that
    are infinite -- although semi-infinite can be checked in one direction)
 * All non-infinite domain ends would have an event created for them, but
    default is for them to be inactive.

 The 'mspec' argument to the GeneratorConstructor class must be a complete
 ModelSpec (up to introduction of global references including definitions of
 external inputs).

 The resulting model output from getModel contains the ModelSpec mspec in
 order to use its structure in resolving information about the relationship
 between variables.

Classes

[hide private]

Descriptor
Abstract class for model and generator descriptors

EvMapping
Event mapping class, for use by makeModelInfoEntry and, when instantiated, the Model class.

GDescriptor
All-in-one descriptor class for single Generators, and information necessary to be able to build a Model object using a ModelConstructor call -- i.e.

GeneratorConstructor

MDescriptor
All-in-one descriptor class for hybrid model definitions and information necessary to be able to build a Model object using a ModelConstructor call.

ModelConstructor

Functions

[hide private]

embed(gen, icdict=None, name=None, tdata=None, make_copy=True)
Only use this function for building non-hybrid models with single Generators.

source code

makeEvMapping(mappingDict, varnames, parnames)

source code

makeModelInfo(arg)

source code

makeModelInfoEntry(dsi, allModelNames=None, swmap_list=None, globcon_list=None, nonevent_reasons=None)
Create an entry for the modelInfo attribute of a Model or Generator, already wrapped in a dsInterface object.

source code

processReused(sourcenames, auxvarnames, flatspec, registry, FScompatibleNames, FScompatibleNamesInv)
Find and process reused terms in abstract specification.

source code

validateTransitionName(name, special_reasons)

source code

Variables

[hide private]

API = API_class()

Continuous = Continuous Domain

Discrete = Discrete Domain

Inf = inf

LargestInt32 = 2147483647

NAMESEP = '.'

NaN = nan

_all_complex = (<type 'complex'>, <type 'numpy.complexfloating...

_all_float = (<type 'float'>, <type 'numpy.floating'>, <type '...

_all_int = (<type 'int'>, <type 'numpy.integer'>, <type 'numpy...

_all_numpy_complex = (<type 'numpy.complex128'>, <type 'numpy....

_all_numpy_float = (<type 'numpy.float64'>, <type 'numpy.float...

_all_numpy_int = (<type 'numpy.int32'>, <type 'numpy.int32'>, ...

_complex_types = (<type 'complex'>, <type 'numpy.complexfloati...

_float_types = (<type 'float'>, <type 'numpy.floating'>)

_int_types = (<type 'int'>, <type 'numpy.integer'>)

_num_equivtype = {<type 'float'>: <type 'numpy.float64'>, <typ...

_num_maxmin = {<type 'numpy.int32'>: [-2147483648, 2147483647]...

_num_name2equivtypes = {'float': (<type 'float'>, <type 'numpy...

_num_name2type = {'float': <type 'numpy.float64'>, 'int': <typ...

_num_type2name = {<type 'float'>: 'float', <type 'int'>: 'int'...

_num_types = (<type 'float'>, <type 'int'>, <type 'numpy.float...

_pytypefromtype = {<type 'numpy.int32'>: <type 'int'>, <type '...

_real_types = (<type 'int'>, <type 'numpy.integer'>, <type 'fl...

_seq_types = (<type 'list'>, <type 'tuple'>, <type 'numpy.ndar...

isfinite = <ufunc 'isfinite'>

mathNameMap = {'Abs': 'abs', 'Acos': 'acos', 'Asin': 'asin', '...

null_predicate = null_predicate_class(None)

targetLangs = ['c', 'python', 'matlab']

Function Details

[hide private]

embed(gen, icdict=None, name=None, tdata=None, make_copy=True)

source code

Only use this function for building non-hybrid models with single Generators. Otherwise, use the ModelConstructor class.

NB The supplied Generator is *copied* into the model unless optional make_copy argument is False.

makeModelInfoEntry(dsi, allModelNames=None, swmap_list=None, globcon_list=None, nonevent_reasons=None)

source code

Create an entry for the modelInfo attribute of a Model or Generator, already wrapped in a dsInterface object. Specify the list of non-event based reasons which can be generated by global consistency checks.

processReused(sourcenames, auxvarnames, flatspec, registry, FScompatibleNames, FScompatibleNamesInv)

source code

Find and process reused terms in abstract specification. To avoid RHS specs depending on auxiliary variables, temp variables will be declared in FuncSpec.py and used in both the RHS and auxiliary variables in the target language specification.

Variables Details

[hide private]

_all_complex

Value:

(<type 'complex'>,
 <type 'numpy.complexfloating'>,
 <type 'numpy.complex128'>,
 <type 'numpy.complex64'>,
 <type 'numpy.complex128'>)

_all_float

Value:

(<type 'float'>,
 <type 'numpy.floating'>,
 <type 'numpy.float64'>,
 <type 'numpy.float32'>,
 <type 'numpy.float64'>)

_all_int

Value:

(<type 'int'>,
 <type 'numpy.integer'>,
 <type 'numpy.int32'>,
 <type 'numpy.int32'>,
 <type 'numpy.int8'>,
 <type 'numpy.int16'>,
 <type 'numpy.int32'>,
 <type 'numpy.int64'>)

_all_numpy_complex

Value:

(<type 'numpy.complex128'>,
 <type 'numpy.complex64'>,
 <type 'numpy.complex128'>)

_all_numpy_float

Value:

(<type 'numpy.float64'>,
 <type 'numpy.float32'>,
 <type 'numpy.float64'>)

_all_numpy_int

Value:

(<type 'numpy.int32'>,
 <type 'numpy.int32'>,
 <type 'numpy.int8'>,
 <type 'numpy.int16'>,
 <type 'numpy.int32'>,
 <type 'numpy.int64'>)

_complex_types

Value:

(<type 'complex'>, <type 'numpy.complexfloating'>)

_num_equivtype

Value:

{<type 'float'>: <type 'numpy.float64'>,
 <type 'int'>: <type 'numpy.int32'>,
 <type 'numpy.integer'>: <type 'numpy.int32'>,
 <type 'numpy.floating'>: <type 'numpy.float64'>,
 <type 'numpy.int8'>: <type 'numpy.int32'>,
 <type 'numpy.int16'>: <type 'numpy.int32'>,
 <type 'numpy.int32'>: <type 'numpy.int32'>,
 <type 'numpy.int32'>: <type 'numpy.int32'>,
...

_num_maxmin

Value:

{<type 'numpy.int32'>: [-2147483648, 2147483647],
 <type 'numpy.float64'>: [-inf, inf]}

_num_name2equivtypes

Value:

{'float': (<type 'float'>,
           <type 'numpy.floating'>,
           <type 'numpy.float64'>,
           <type 'numpy.float32'>,
           <type 'numpy.float64'>),
 'int': (<type 'int'>,
         <type 'numpy.integer'>,
         <type 'numpy.int32'>,
...

_num_name2type

Value:

{'float': <type 'numpy.float64'>, 'int': <type 'numpy.int32'>}

_num_type2name

Value:

{<type 'float'>: 'float',
 <type 'int'>: 'int',
 <type 'numpy.integer'>: 'int',
 <type 'numpy.floating'>: 'float',
 <type 'numpy.int8'>: 'int',
 <type 'numpy.int16'>: 'int',
 <type 'numpy.int32'>: 'int',
 <type 'numpy.int32'>: 'int',
...

_num_types

Value:

(<type 'float'>,
 <type 'int'>,
 <type 'numpy.floating'>,
 <type 'numpy.integer'>)

_pytypefromtype

Value:

{<type 'numpy.int32'>: <type 'int'>,
 <type 'numpy.float64'>: <type 'float'>}

_real_types

Value:

(<type 'int'>,
 <type 'numpy.integer'>,
 <type 'float'>,
 <type 'numpy.floating'>)

_seq_types

Value:

(<type 'list'>, <type 'tuple'>, <type 'numpy.ndarray'>)

mathNameMap

Value:

{'Abs': 'abs',
 'Acos': 'acos',
 'Asin': 'asin',
 'Atan': 'atan',
 'Atan2': 'atan2',
 'Betavariate': 'betavariate',
 'Ceil': 'ceil',
 'Choice': 'choice',
...