def get_var(backend_model, var, dims=None, sparse=False):
"""
Return output for variable `var` as a pandas.Series (1d),
pandas.Dataframe (2d), or xarray.DataArray (3d and higher).
Parameters
----------
var : variable name as string, e.g. 'resource'
dims : list, optional
indices as strings, e.g. ('loc_techs', 'timesteps');
if not given, they are auto-detected
sparse : bool, optional; default = False
If extracting Pyomo Param data, the output sparse array includes inputs
the user left as NaN replaced with the default value for that Param.
"""
try:
var_container = getattr(backend_model, var)
except AttributeError:
raise exceptions.BackendError("Variable {} inexistent.".format(var))
if not dims:
if var + "_index" == var_container.index_set().name:
dims = [i.name for i in var_container.index_set().set_tuple]
else:
dims = [var_container.index_set().name]
if sparse:
result = pd.DataFrame.from_dict(var_container.extract_values_sparse(),
orient="index")
else:
result = pd.DataFrame.from_dict(var_container.extract_values(),
orient="index")
if result.empty:
raise exceptions.BackendError("Variable {} has no data.".format(var))
result = result[0] # Get the only column in the dataframe
if len(dims) > 1:
result.index = pd.MultiIndex.from_tuples(result.index, names=dims)
if len(result.index.names) == 1:
result = result.sort_index()
result.index.name = dims[0]
elif len(result.index.names) == 2:
# if len(dims) is 2, we already have a well-formed DataFrame
result = result.unstack(level=0)
result = result.sort_index()
else: # len(dims) >= 3
result = xr.DataArray.from_series(result)
return result
def get_var(backend_model, var, dims=None, sparse=False, expr=False):
"""
Return output for variable `var` as a pandas.Series (1d),
pandas.Dataframe (2d), or xarray.DataArray (3d and higher).
Parameters
----------
var : variable name as string, e.g. 'resource'
dims : list, optional
indices as strings, e.g. ('loc_techs', 'timesteps');
if not given, they are auto-detected
sparse : bool, optional; default = False
If extracting Pyomo Param data, the output sparse array includes inputs
the user left as NaN replaced with the default value for that Param.
expr : bool, optional
If True, treat var as a pyomo expression, which requires calculating
the result of the expression before translating into nd data structure
"""
try:
var_container = getattr(backend_model, var)
except AttributeError:
raise exceptions.BackendError("Variable {} inexistent.".format(var))
if not dims:
if var + "_index" == var_container.index_set().name:
dims = [i.name for i in var_container.index_set().subsets()]
else:
dims = [var_container.index_set().name]
if sparse and not expr:
if invalid(var_container.default()):
result = pd.Series(var_container._data).apply(
lambda x: po.value(x) if not invalid(x) else np.nan)
else:
result = pd.Series(var_container.extract_values_sparse())
else:
if expr:
result = pd.Series(var_container._data).apply(po.value)
else:
result = pd.Series(var_container.extract_values())
if result.empty:
raise exceptions.BackendError("Variable {} has no data.".format(var))
result = result.rename_axis(index=dims)
return xr.DataArray.from_series(result)
