modelpattern module

Created on Mon Sep 02 19:32:22 2013

This module defines a number of pattern used in PYFS. If a new function is intruduced in the model definition language it should added to the function names in funkname

All functions in the module modeluserfunk will be added to the language and incorporated in the Business Logic language

@author: Ib

class modelpattern.nterm(number, op, var, lag)

Bases: tuple

Create new instance of nterm(number, op, var, lag)

lag

Alias for field number 3

number

Alias for field number 0

op

Alias for field number 1

var

Alias for field number 2

modelpattern.udtrykre(funks=[])

modelpattern.find_frml(equations)

Takes at modeltext and returns a list with where each element is a string starting with FRML and ending with $ It do not check if it is a valid FRML statement

modelpattern.split_frml(frml)

Splits a string with a frml into a tuple with 4 parts:

0. The unsplit frml statement

1. FRML

2. <Frml name>

3. <the frml expression>

modelpattern.find_statements(a_model)

splits a modeltest into comments and statements

• a comment starts with ! and ends at lineend

• a statement starts with a name and ends with a $ all characters between are considerd part of the statement

The statement is not chekked for meaningfulness returns a list of tuppels (comment,command,<rest of statement>)

modelpattern.model_parse_old(equations, funks=[])

Takes a model returns a list of tupels. Each tupel contains:

the compleete formular:

FRML:

formular name:

the expression:

list of terms from the expression:

The purpose of this function is to make model analysis faster. this is 20 times faster than looping over espressions in a model

modelpattern.model_parse(equations, funks=[])

Takes a model returns a list of tupels. Each tupel contains:

the compleete formular:

FRML:

formular name:

the expression:

list of terms from the expression:

The purpose of this function is to make model analysis faster. this is 20 times faster than looping over espressions in a model

This new model_parse handels lags of -0 or +0 which ocours in some models from world bank.

modelpattern.list_extract(equations, silent=True)

creates lists used in a model

returns a dictonary with the lists if a list is defined several times, the first definition is used

modelpattern.check_syntax_model(equations, test=True)

cheks if equations have syntax errors by calling the python compile.parse

modelpattern.udtryk_parse(udtryk, funks=[])

returns a list of terms from an expression ie: lhs=rhs $ or just an expression like x+b

modelpattern.kw_frml_name(frml_name0, kw, default=None)

find keywords and associated value from string ‘<kw=xxx,res=kdkdk>’

modelpattern.f1()

modelpattern.f2()