def build_endogenous_variables(self, code):
"""Return code to store endogenous variable values.
Parameters
==========
code : string
Code script containing endogenous model variables
Returns
=======
variables : string
Python code to insert endogenous variable values into a Dictionary
See also
========
FSIC.parser.code.identify_variables()
"""
from FSIC.parser.code import identify_variables
variables = identify_variables(code)
variables = variables['endogenous']
variables = [
('values[\'' + v.replace('self.', '') + '\'] = ' + v + '[period]')
for v in variables
]
variables = '\n'.join(variables)
return variables
def build_results(self, code):
"""Return code to return model results as a DataFrame.
Parameters
==========
code : string
Code script containing all model variables
Returns
=======
results : string
Python code to generate a results DataFrame
See also
========
FSIC.parser.code.identify_variables()
"""
from FSIC.parser.code import identify_variables
variables = identify_variables(code)
variables = variables['endogenous'] + variables['exogenous']
results = 'results = DataFrame({' + '\n\t' + (',\n\t'.join(
['\'' + v.replace('self.', '') + '\': ' + v
for v in variables]) + '})')
return results
def build_initialise(self, code):
"""Return code to initialise the variables of a model.
Parameters
==========
code : string
Code script containing all model variables
Returns
=======
initialise : string
Python code to initialise model variables
See also
========
FSIC.parser.code.identify_variables()
"""
from FSIC.parser.code import identify_variables
variables = identify_variables(code)
variables = variables['endogenous'] + variables['exogenous']
initialise = [
v + (' = Series(default, '
'index=self.full_span, '
'dtype=dtype)') for v in variables
]
initialise = '\n'.join(initialise)
return initialise
def build_results(self, code):
"""Return code to return model results as a DataFrame.
Parameters
==========
code : string
Code script containing all model variables
Returns
=======
results : string
Python code to generate a results DataFrame
See also
========
FSIC.parser.code.identify_variables()
"""
from FSIC.parser.code import identify_variables
variables = identify_variables(code)
variables = variables['endogenous'] + variables['exogenous']
results = 'results = DataFrame({' + '\n\t' + (
',\n\t'.join(
['\'' + v.replace('self.', '') + '\': ' + v
for v in variables]) + '})')
return results
def build_endogenous_variables(self, code):
"""Return code to store endogenous variable values.
Parameters
==========
code : string
Code script containing endogenous model variables
Returns
=======
variables : string
Python code to insert endogenous variable values into a Dictionary
See also
========
FSIC.parser.code.identify_variables()
"""
from FSIC.parser.code import identify_variables
variables = identify_variables(code)
variables = variables['endogenous']
variables = [('values[\'' +
v.replace('self.', '') +
'\'] = ' + v + '[period]')
for v in variables]
variables = '\n'.join(variables)
return variables
def build_initialise(self, code):
"""Return code to initialise the variables of a model.
Parameters
==========
code : string
Code script containing all model variables
Returns
=======
initialise : string
Python code to initialise model variables
See also
========
FSIC.parser.code.identify_variables()
"""
from FSIC.parser.code import identify_variables
variables = identify_variables(code)
variables = variables['endogenous'] + variables['exogenous']
initialise = [v + (' = Series(default, '
'index=self.full_span, '
'dtype=dtype)')
for v in variables]
initialise = '\n'.join(initialise)
return initialise
def make_graph(equations, warn=True):
"""Return `equations` as a NetworkX DiGraph.
Parameters
==========
equations : list of strings
List of equations to reorder, one equation per element
warn : boolean
If `True`, print a warning if there is more than one equation with the
same endogenous variable
Returns
=======
G : NetworkX DiGraph object
Directed graph representation of `equations`. Nodes signifying
endogenous variables have, as an attribute, the associated equation
string(s). If `warn` is `True`, this function prints a warning if
multiple equations are found for a single endogenous variable.
"""
from FSIC.parser.code import identify_variables
G = nx.DiGraph()
# Initialise dictionary to store equations
node_equations = {}
# Loop by equation
for e in equations:
v = identify_variables(e, suffix=r'\[.+?\]')
# Extract endogenous variable (should only be one)
n = v['endogenous']
if len(n) != 1:
raise ValueError('Expected just one endogenous variable')
n = n[0]
# Extract exogenous variable(s) and add edges
x = v['exogenous']
for term in x:
G.add_edge(term, n)
# Store equation with the endogenous variable as the key
if n in node_equations:
if warn:
warnings.warn(
'An endogenous variable appears as the left hand-side '
'variable in more than one equation',
Warning)
node_equations[n] = node_equations[n] + [e]
else:
node_equations[n] = [e]
# Add equations to node attributes
nx.set_node_attributes(G, 'equations', node_equations)
# Return
return G