# Copyright 2025, BRGM
#
# This file is part of Rameau.
#
# Rameau is free software: you can redistribute it and/or modify it under the
# terms of the GNU General Public License as published by the Free Software
# Foundation, either version 3 of the License, or (at your option) any later
# version.
#
# Rameau is distributed in the hope that it will be useful, but WITHOUT ANY
# WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A
# PARTICULAR PURPOSE. See the GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License along with
# Rameau. If not, see <https://www.gnu.org/licenses/>.
#
"""
rameau model.
"""
from __future__ import annotations
from typing import Union, Optional, Literal, List
import datetime
import functools
import pandas as pd
import numpy as np
from rameau.core.settings.spinup_settings import SpinupSettings
from rameau.wrapper import CModel # type: ignore
from rameau.core.settings import (
SimulationSettings,
ForecastSettings,
OptimizationSettings,
OutputSettings
)
from rameau.core.inputs import InputCollection
from rameau.core import states
from rameau.core.states import StatesCollection, States
from rameau.core.tree import Tree
from rameau.core.simulation import (
Simulation,
OptiSimulation,
ForecastSimulation
)
from rameau.core._utils import _check_literal
from rameau.core._abstract_wrapper import AbstractWrapper
from rameau.core._utils import _build_type, wrap_property
from rameau.core._descriptor import _DerivedTypeDecriptor
from rameau._typing import MethodType, TransformationType, ObjFunctionType
from copy import copy
def overload_sim(method):
@functools.wraps(method)
def factory(self, *args, **kwargs):
sim = method(self, *args, **kwargs)
sim.optimization_settings = copy(self.optimization_settings)
sim._forecast_settings = copy(self.forecast_settings)
return sim
return factory
def overload_opti_sim(method):
@functools.wraps(method)
def factory(self, *args, **kwargs):
sim = method(self, *args, **kwargs)
sim._forecast_settings = copy(self.forecast_settings)
return sim
return factory
def overload_forecast_sim(method):
@functools.wraps(method)
def factory(self, *args, **kwargs):
sim = method(self, *args, **kwargs)
sim.optimization_settings = copy(self.optimization_settings)
return sim
return factory
[docs]
class Model(AbstractWrapper):
"""Define a rameau model.
Parameters
----------
tree: `dict` or `Tree`
Watershed connection tree.
inputs: `dict` or `inputs.InputCollection`
Model input data.
init_states: `dict` or `states.StatesCollection`, optional
Model initial states.
simulation_settings: `dict` or `SimulationSettings`, optional
Settings related to a simulation run.
See `SimulationSettings` for details.
optimization_settings: `dict` or `OptimizationSettings`, optional
Settings related to an optimisation run.
See `OptimizationSettings` for details.
forecast_settings: `dict` or `ForecastSettings`, optional
Settings related to a forecast run.
See `ForecastSettings` for details.
output_settings: `dict` or `OutputSettings`, optional
Settings related to the simulation outputs
See `OutputSettings` for details.
Returns
-------
`Model`
Examples
--------
Constructing model from `Tree` and `inputs.InputCollection`.
>>> data = np.array([0.1, 0.2, 0.3])
>>> model = rm.Model(
... tree=rm.Tree(watersheds=[{}]),
... inputs=rm.inputs.InputCollection(rainfall=data, pet=data)
... )
>>> model.inputs.rainfall.data
array([[0.1],
[0.2],
[0.3]], dtype=float32)
Constructing model from `dict`.
>>> model = rm.Model(
... tree=dict(watersheds=[{}]),
... inputs=dict(rainfall=data, pet=data)
... )
>>> model.inputs.rainfall.data
array([[0.1],
[0.2],
[0.3]], dtype=float32)
"""
_computed_attributes = (
"tree", "inputs", "init_states", "simulation_settings",
"optimization_settings", "forecast_settings",
"output_settings"
)
_c_class = CModel
tree: Tree = _DerivedTypeDecriptor(
0, Tree, "Watershed connection tree of the model."
) #type: ignore
simulation_settings: SimulationSettings = _DerivedTypeDecriptor(
0, SimulationSettings, "Settings related to a simulation run."
) #type: ignore
optimization_settings: OptimizationSettings = _DerivedTypeDecriptor(
0, OptimizationSettings, "Settings related to an optimization run."
) #type: ignore
forecast_settings: ForecastSettings = _DerivedTypeDecriptor(
0, ForecastSettings, "Settings related to an forecast run."
) #type: ignore
output_settings: OutputSettings = _DerivedTypeDecriptor(
0, OutputSettings, "Settings related to simulation outputs."
) #type: ignore
init_states: StatesCollection = _DerivedTypeDecriptor(
0, states.StatesCollection, "Model initial states."
) #type: ignore
def __init__(
self,
tree: Union[dict, Tree],
inputs: Union[dict, InputCollection],
init_states: Optional[
Union[List[Union[dict, States]], Union[dict, StatesCollection]]
] = None,
simulation_settings: Optional[Union[dict, SimulationSettings]] = None,
optimization_settings: Optional[Union[dict, OptimizationSettings]] = None,
forecast_settings: Optional[Union[dict, ForecastSettings]] = None,
output_settings: Optional[Union[dict, OutputSettings]] = None
) -> None:
self._init_c()
self.tree = _build_type(tree, Tree)
self.inputs = _build_type(inputs, InputCollection)
if init_states is not None:
if isinstance(init_states, list):
self.init_states = StatesCollection(states=init_states)
else:
self.init_states = _build_type(init_states, StatesCollection)
else:
self._m.set_default_init_states()
if simulation_settings is not None:
self.simulation_settings = _build_type(
simulation_settings, SimulationSettings
)
else:
self.simulation_settings = SimulationSettings()
if optimization_settings is not None:
self.optimization_settings = _build_type(
optimization_settings, OptimizationSettings
)
else:
self.optimization_settings = OptimizationSettings()
if forecast_settings is not None:
self.forecast_settings = _build_type(
forecast_settings, ForecastSettings
)
else:
self.forecast_settings = ForecastSettings()
if output_settings is not None:
self.output_settings = _build_type(
output_settings, OutputSettings
)
else:
self.output_settings = OutputSettings()
[docs]
@overload_sim
@wrap_property(Simulation)
def create_simulation( self, **kwargs) -> Simulation:
"""Create a simulation.
See `Model.create_simulation` for keyword argument descriptions.
Returns
-------
`simulation.Simulation`
"""
kwargs2 = {}
for key in SimulationSettings._computed_attributes:
if key in kwargs:
kwargs2[key] = kwargs[key]
else:
kwargs2[key] = getattr(self.simulation_settings, key)
opt = SimulationSettings(**kwargs2)
sim, err = self._m.create_simulation(opt._m)
if err.getInt(0) != 0:
raise RuntimeError(err.getString(0))
return sim
[docs]
@overload_sim
@wrap_property(Simulation)
def run_simulation(
self,
name: Optional[str] = None,
starting_date: Optional[datetime.datetime] = None,
ending_date: Optional[datetime.datetime] = None,
spinup_settings: Optional[Union[SpinupSettings, dict]] = None
) -> Simulation:
"""Start a simulation run.
Parameters
----------
name: `str`, optional
Name of the simulation. This name is used to name to the output
directory. If omitted, default is "simulation".
starting_date: `datetime.datetime`, optional
Starting date of the simulation. This date needs to be contained in the
rainfall input data. If None, default is the starting date of
the rainfall input data. The simulation run will start from this
date and will stop at the end of the meteorological input data or
at the simulation ending date.
ending_date: `datetime.datetime`, optional
Ending date of the simulation. This date needs to be contained in the
rainfall input data. If None, default is the ending date of
the rainfall input data.
spinup_settings: `dict` or `SpinupSettings`, optional
Spinup settings of the simulation. See `SpinupSettings` for details.
Returns
-------
`simulation.Simulation`
"""
attrs = {
"name":name, "starting_date":starting_date,
"ending_date":ending_date,
"spinup_settings":spinup_settings
}
kwargs = {}
for key, value in attrs.items():
if value is not None:
kwargs[key] = value
else:
kwargs[key] = getattr(self.simulation_settings, key)
opt = SimulationSettings(**kwargs)
sim, err = self._m.run_simulation(opt._m)
if err.getInt(0) != 0:
raise RuntimeError(err.getString(0))
return sim
[docs]
@overload_opti_sim
@wrap_property(OptiSimulation)
def run_optimization(
self,
maxit: Optional[int] = None,
starting_date: Optional[datetime.datetime] = None,
ending_date: Optional[datetime.datetime] = None,
method: Optional[MethodType] = None,
transformation: Optional[TransformationType] = None,
river_objective_function: Optional[ObjFunctionType] = None,
selected_watersheds: Optional[List[int]] = None,
verbose = None
) -> OptiSimulation:
"""Start an optimisation run.
Parameters
----------
maxit: `int`, optional
Number of iterations for the optimisation algorithm.
starting_date: `datetime.datetime`, optional
The date and time defining the start of the period to consider
in the input data for the optimisation run.
ending_date: `datetime.datetime`, optional
The date and time defining the end of the period to consider
in the input data for the optimisation run.
method: `str`, optional
The approach to use when several gauged watersheds need to
be considered in the optimisation run.
See `OptimizationSettings.method` for details.
transformation: `str`, optional
The function to apply to transform the observed and predicted river
flow (Q) before computing the objective function.
See `OptimizationSettings.transformation` for details.
river_objective_function: `str`, optional
The objective function to use to compare the observed and
predicted river flow.
See `OptimizationSettings.river_objective_function` for details.
selected_watersheds: `list` or `int`, optional
The indices of the watersheds to consider in the
optimisation run. The indices relate to those in the
sequence of watersheds specified in the `Tree`. If not
provided, all gauged watersheds are considered.
verbose: `bool`, optional
Whether to display information for each step of the
optimisation process. If not provided, no information is
displayed.
Returns
-------
`simulation.OptiSimulation`
"""
attrs = {
"maxit":maxit, "starting_date":starting_date,
"ending_date":ending_date, "method":method,
"transformation":transformation,
"river_objective_function":river_objective_function,
"selected_watersheds":selected_watersheds,
"verbose":verbose
}
kwargs = {}
for key, value in attrs.items():
if value is not None:
kwargs[key] = value
else:
val = getattr(self.optimization_settings, key)
if not (key == 'selected_watersheds' and val == []):
kwargs[key] = val
opt = OptimizationSettings(**kwargs)
if opt.maxit > 0:
sim, err = self._m.run_optimization(opt._m)
if err.getInt(0) != 0:
raise RuntimeError(err.getString(0))
return sim
else:
raise RuntimeError("0 iterations defined (maxit=0).")
[docs]
@overload_forecast_sim
@wrap_property(ForecastSimulation)
def run_forecast(
self,
emission_date: Optional[datetime.datetime] = None,
scope: Optional[datetime.timedelta] = None,
year_members: Optional[Union[List[int], tuple[int]]] = None,
correction: Optional[Literal["no", "halflife"]] = None,
pumping_date: Optional[datetime.datetime] = None,
quantiles_output: Optional[bool] = None,
quantiles: Optional[Union[List[int], tuple[int]]] = None,
norain: Optional[bool] = None
) -> ForecastSimulation:
"""Start a forecast run.
Parameters
----------
emission_date: `datetime.datetime`, optional
The date and time on which to issue a forecast.
scope: `datetime.timedelta`, optional
The duration for which to run the forecast. If not provided,
set to one day.
year_members: `list` or `int`, optional
The years to consider to form the forecast ensemble members.
If not provided, all years in record are considered.
correction: `str`, optional
The approach to use to correct the initial conditions
before issuing a forecast. See `ForecastSettings` for details.
pumping_date: `datetime.datetime`, optional
quantiles_output: `bool`, optional
Whether to reduce the forecast ensemble members to specific
climatology quantiles. If not provided, all years in record
or years specified via the ``year_members`` parameter are
considered. The quantiles can be chosen via the ``quantiles``
parameter.
quantiles: `list` or `int`, optional
The climatology percentiles to include in the forecast ensemble
members. Only considered if ``quantiles_output`` is set to
`True`. By default, the percentiles computed are 10, 20, 50,
80, and 90.
norain: `bool`, optional
Whether to include an extra ensemble member corresponding to
a completely rain-free year. By default, this member is not
included in the forecast output.
Returns
-------
`simulation.ForecastSimulation`
"""
attrs = {
"emission_date":emission_date, "scope":scope,
"year_members":year_members,
"correction":correction, "pumping_date":pumping_date,
"quantiles_output":quantiles_output, "quantiles":quantiles,
"norain":norain
}
kwargs = {}
for key, value in attrs.items():
if value is not None:
kwargs[key] = value
else:
val = getattr(self.forecast_settings, key)
if not (key == 'year_members' and val == []):
kwargs[key] = val
fcast = ForecastSettings(**kwargs)
sim, err = self._m.run_forecast(fcast._m)
if err.getInt(0) != 0:
raise RuntimeError(err.getString(0))
return sim
[docs]
@classmethod
def from_toml(cls, path: str, load_inputs: bool = True) -> Model:
"""Load a model from a TOML file.
Parameters
----------
path: `str`
TOML file path.
load_inputs: `bool`
Whether to read and load in memory input data from text files.
If False, inputs will be read during simulation creation, if necessary.
Default is True.
"""
model = cls.__new__(cls)
model._m = CModel()
err = model._m.from_toml(path, load_inputs)
if err.getInt(0) != 0:
raise RuntimeError(err.getString(0))
return model
[docs]
def to_toml(
self,
path: str,
tree: Optional[Tree] = None,
) -> None:
"""
Dump the model to a TOML file.
Parameters
----------
path: `str`
TOML file path.
tree: `Tree`, optional
The `Tree` object to write in the TOML file. If None, write the
`Tree` object associated with the `tree` model attribute.
"""
if tree is None:
tree = self.tree
err = self._m.to_toml(path, tree._m)
if err.getInt(0) != 0:
raise RuntimeError(err.getString(0))
def _input_to_dataframe(self, data, nan_nodata=False):
d = data.data
if np.size(d) == 0:
return pd.DataFrame()
df = pd.DataFrame(
data=d, index=data.dates,
columns=range(1, d.shape[1] + 1)
)
df.index.name = "dates"
df.columns.name = "zones"
if nan_nodata:
df = df.where(df != data.nodata, np.nan)
return df
@property
@wrap_property(InputCollection)
def inputs(self) -> InputCollection:
"""Model input data.
Returns
-------
`inputs.InputCollection`
"""
return self._m.getInputs()
@inputs.setter
def inputs(self, v: InputCollection) -> None:
e = self._m.setInputs(v._m)
if e.getInt(0) != 0:
raise RuntimeError(e.getString(0))
