from __future__ import annotations
import collections
import pathlib
import subprocess
import warnings
from copy import deepcopy
from datetime import datetime
from pathlib import Path
from typing import Any, Callable, DefaultDict, Iterable, Optional, Union, cast
import cftime
import dask
import jinja2
import numpy as np
import tomli
import tomli_w
import xarray as xr
import xugrid as xu
import imod
import imod.logging
import imod.mf6.exchangebase
from imod.logging import standard_log_decorator
from imod.mf6.gwfgwf import GWFGWF
from imod.mf6.gwfgwt import GWFGWT
from imod.mf6.gwtgwt import GWTGWT
from imod.mf6.ims import Solution, SolutionPresetModerate
from imod.mf6.interfaces.imodel import IModel
from imod.mf6.interfaces.isimulation import ISimulation
from imod.mf6.model import Modflow6Model
from imod.mf6.model_gwf import GroundwaterFlowModel
from imod.mf6.model_gwt import GroundwaterTransportModel
from imod.mf6.multimodel.exchange_creator_structured import ExchangeCreator_Structured
from imod.mf6.multimodel.exchange_creator_unstructured import (
ExchangeCreator_Unstructured,
)
from imod.mf6.multimodel.modelsplitter import create_partition_info, slice_model
from imod.mf6.out import open_cbc, open_conc, open_hds
from imod.mf6.package import Package
from imod.mf6.regrid.regrid_schemes import RegridMethodType
from imod.mf6.ssm import SourceSinkMixing
from imod.mf6.statusinfo import NestedStatusInfo
from imod.mf6.utilities.mask import _mask_all_models
from imod.mf6.utilities.regrid import _regrid_like
from imod.mf6.validation_context import ValidationContext
from imod.mf6.write_context import WriteContext
from imod.prepare.topsystem.default_allocation_methods import (
SimulationAllocationOptions,
SimulationDistributingOptions,
)
from imod.schemata import ValidationError
from imod.typing import GridDataArray, GridDataset
from imod.typing.grid import (
concat,
is_equal,
is_unstructured,
merge_partitions,
)
OUTPUT_FUNC_MAPPING: dict[str, Callable] = {
"head": open_hds,
"concentration": open_conc,
"budget-flow": open_cbc,
"budget-transport": open_cbc,
}
OUTPUT_MODEL_MAPPING: dict[
str, type[GroundwaterFlowModel] | type[GroundwaterTransportModel]
] = {
"head": GroundwaterFlowModel,
"concentration": GroundwaterTransportModel,
"budget-flow": GroundwaterFlowModel,
"budget-transport": GroundwaterTransportModel,
}
def get_models(simulation: Modflow6Simulation) -> dict[str, Modflow6Model]:
return {k: v for k, v in simulation.items() if isinstance(v, Modflow6Model)}
def get_packages(simulation: Modflow6Simulation) -> dict[str, Package]:
return {
pkg_name: pkg
for pkg_name, pkg in simulation.items()
if isinstance(pkg, Package)
}
[docs]
class Modflow6Simulation(collections.UserDict, ISimulation):
def _initialize_template(self):
loader = jinja2.PackageLoader("imod", "templates/mf6")
env = jinja2.Environment(loader=loader, keep_trailing_newline=True)
self._template = env.get_template("sim-nam.j2")
[docs]
def __init__(self, name):
super().__init__()
self.name = name
self.directory = None
self._initialize_template()
self._validation_context = ValidationContext()
def __setitem__(self, key, value):
super().__setitem__(key, value)
def update(self, *args, **kwargs):
for k, v in dict(*args, **kwargs).items():
self[k] = v
def time_discretization(self, times):
warnings.warn(
f"{self.__class__.__name__}.time_discretization() is deprecated. "
f"In the future call {self.__class__.__name__}.create_time_discretization().",
DeprecationWarning,
)
self.create_time_discretization(additional_times=times)
[docs]
def create_time_discretization(self, additional_times, validate: bool = True):
"""
Collect all unique times from model packages and additional given
`times`. These unique times are used as stress periods in the model. All
stress packages must have the same starting time. Function creates
TimeDiscretization object which is set to self["time_discretization"]
The time discretization in imod-python works as follows:
- The datetimes of all packages you send in are always respected
- Subsequently, the input data you use is always included fully as well
- All times are treated as starting times for the stress: a stress is
always applied until the next specified date
- For this reason, a final time is required to determine the length of
the last stress period
- Additional times can be provided to force shorter stress periods &
more detailed output
- Every stress has to be defined on the first stress period (this is a
modflow requirement)
Or visually (every letter a date in the time axes):
>>> recharge a - b - c - d - e - f
>>> river g - - - - h - - - - j
>>> times - - - - - - - - - - - i
>>> model a - b - c h d - e - f i
with the stress periods defined between these dates. I.e. the model
times are the set of all times you include in the model.
Parameters
----------
additional_times : str, datetime; or iterable of str, datetimes.
Times to add to the time discretization. At least one single time
should be given, which will be used as the ending time of the
simulation.
Note
----
To set the other parameters of the TimeDiscretization object, you have
to set these to the object after calling this function.
Example
-------
>>> simulation = imod.mf6.Modflow6Simulation("example")
>>> simulation.create_time_discretization(times=["2000-01-01", "2000-01-02"])
>>> # Set number of timesteps
>>> simulation["time_discretization"]["n_timesteps"] = 5
"""
self.use_cftime = any(
model._use_cftime()
for model in self.values()
if isinstance(model, Modflow6Model)
)
times = [
imod.util.time.to_datetime_internal(time, self.use_cftime)
for time in additional_times
]
for model in self.values():
if isinstance(model, Modflow6Model):
times.extend(model._yield_times())
# np.unique also sorts
times = np.unique(np.hstack(times))
duration = imod.util.time.timestep_duration(times, self.use_cftime) # type: ignore
# Generate time discretization, just rely on default arguments
# Probably won't be used that much anyway?
timestep_duration = xr.DataArray(
duration, coords={"time": np.array(times)[:-1]}, dims=("time",)
)
self["time_discretization"] = imod.mf6.TimeDiscretization(
timestep_duration=timestep_duration, validate=validate
)
def render(self, write_context: WriteContext):
"""Renders simulation namefile"""
d: dict[str, Any] = {}
models = []
solutiongroups = []
for key, value in self.items():
if isinstance(value, Modflow6Model):
model_name_file = pathlib.Path(
write_context.root_directory / pathlib.Path(f"{key}", f"{key}.nam")
).as_posix()
models.append((value.model_id, model_name_file, key))
elif isinstance(value, Package):
if value._pkg_id == "tdis":
d["tdis6"] = f"{key}.tdis"
elif value._pkg_id == "ims":
slnnames = value["modelnames"].values
modeltypes = set()
for name in slnnames:
try:
modeltypes.add(type(self[name]))
except KeyError:
raise KeyError(f"model {name} of {key} not found")
if len(modeltypes) > 1:
raise ValueError(
"Only a single type of model allowed in a solution"
)
solutiongroups.append(("ims6", f"{key}.ims", slnnames))
d["models"] = models
if len(models) > 1:
d["exchanges"] = self.get_exchange_relationships()
d["solutiongroups"] = [solutiongroups]
return self._template.render(d)
[docs]
@standard_log_decorator()
def write(
self,
directory=".",
binary=True,
validate: bool = True,
use_absolute_paths=False,
):
"""
Write Modflow6 simulation, including assigned groundwater flow and
transport models.
Parameters
----------
directory: str, pathlib.Path
Directory to write Modflow 6 simulation to.
use_binary: ({True, False}, optional)
Whether to write time-dependent input for stress packages as binary
files, which are smaller in size, or more human-readable text files.
validate: ({True, False}, optional)
Whether to validate the Modflow6 simulation, including models, at
write. If True, erronous model input will throw a
``ValidationError``.
use_absolute_paths: ({True, False}, optional)
True if all paths written to the mf6 inputfiles should be absolute.
"""
# create write context
write_context = WriteContext(directory, binary, use_absolute_paths)
self._validation_context.validate = validate
if self.is_split():
self._validation_context.strict_well_validation = False
# Check models for required content
for key, model in self.items():
# skip timedis, exchanges
if isinstance(model, Modflow6Model):
model._model_checks(key)
# Generate GWF-GWT exchanges
if gwfgwt_exchanges := self._generate_gwfgwt_exchanges():
self["gwtgwf_exchanges"] = gwfgwt_exchanges
directory = pathlib.Path(directory)
directory.mkdir(exist_ok=True, parents=True)
# Write simulation namefile
mfsim_content = self.render(write_context)
mfsim_path = directory / "mfsim.nam"
with open(mfsim_path, "w") as f:
f.write(mfsim_content)
# Write time discretization file
globaltimes = self["time_discretization"]["time"].values
self["time_discretization"]._write(
pkgname="time_discretization",
globaltimes=globaltimes,
write_context=write_context,
)
# Write individual models
status_info = NestedStatusInfo("Simulation validation status")
for key, value in self.items():
model_write_context = write_context.copy_with_new_write_directory(
write_context.simulation_directory
)
# skip timedis, exchanges
if isinstance(value, Modflow6Model):
status_info.add(
value._write(
modelname=key,
globaltimes=globaltimes,
write_context=model_write_context,
validate_context=self._validation_context,
)
)
elif isinstance(value, Package):
if value._pkg_id == "ims":
ims_write_context = write_context.copy_with_new_write_directory(
write_context.simulation_directory
)
value._write(key, globaltimes, ims_write_context)
elif isinstance(value, list):
for exchange in value:
if isinstance(exchange, imod.mf6.exchangebase.ExchangeBase):
exchange._write(
exchange.package_name(), globaltimes, write_context
)
if status_info.has_errors():
raise ValidationError("\n" + status_info.to_string())
self.directory = directory
[docs]
def run(self, mf6path: Union[str, Path] = "mf6") -> None:
"""
Run Modflow 6 simulation. This method runs a subprocess calling
``mf6path``. This argument is set to ``mf6``, which means the Modflow 6
executable is expected to be added to your PATH environment variable.
:doc:`See this writeup how to add Modflow 6 to your PATH on Windows </examples/mf6/index>`
Note that the ``write`` method needs to be called before this method is
called.
Parameters
----------
mf6path: Union[str, Path]
Path to the Modflow 6 executable. Defaults to calling ``mf6``.
Examples
--------
Make sure you write your model first
>>> simulation.write(path/to/model)
>>> simulation.run()
"""
if self.directory is None:
raise RuntimeError(f"Simulation {self.name} has not been written yet.")
with imod.util.cd(self.directory):
result = subprocess.run(mf6path, capture_output=True)
if result.returncode != 0:
raise RuntimeError(
f"Simulation {self.name}: {mf6path} failed to run with returncode "
f"{result.returncode}, and error message:\n\n{result.stdout.decode()} "
)
[docs]
def open_head(
self,
dry_nan: bool = False,
simulation_start_time: Optional[np.datetime64] = None,
time_unit: Optional[str] = "d",
) -> GridDataArray:
"""
Open heads of finished simulation, requires that the ``run`` method has
been called.
The data is lazily read per timestep and automatically converted into
(dense) xr.DataArrays or xu.UgridDataArrays, for DIS and DISV
respectively. The conversion is done via the information stored in the
Binary Grid file (GRB).
Parameters
----------
dry_nan: bool, default value: False.
Whether to convert dry values to NaN.
simulation_start_time : Optional datetime
The time and date correpsonding to the beginning of the simulation.
Use this to convert the time coordinates of the output array to
calendar time/dates. time_unit must also be present if this argument is present.
time_unit: Optional str
The time unit MF6 is working in, in string representation.
Only used if simulation_start_time was provided.
Admissible values are:
ns -> nanosecond
ms -> microsecond
s -> second
m -> minute
h -> hour
d -> day
w -> week
Units "month" or "year" are not supported, as they do not represent unambiguous timedelta values durations.
Returns
-------
head: Union[xr.DataArray, xu.UgridDataArray]
Examples
--------
Make sure you write and run your model first
>>> simulation.write(path/to/model)
>>> simulation.run()
Then open heads:
>>> head = simulation.open_head()
"""
return self._open_output(
"head",
dry_nan=dry_nan,
simulation_start_time=simulation_start_time,
time_unit=time_unit,
)
[docs]
def open_transport_budget(
self,
species_ls: Optional[list[str]] = None,
simulation_start_time: Optional[np.datetime64] = None,
time_unit: Optional[str] = "d",
) -> GridDataArray | GridDataset:
"""
Open transport budgets of finished simulation, requires that the ``run``
method has been called.
The data is lazily read per timestep and automatically converted into
(dense) xr.DataArrays or xu.UgridDataArrays, for DIS and DISV
respectively. The conversion is done via the information stored in the
Binary Grid file (GRB).
Parameters
----------
species_ls: list of strings, default value: None.
List of species names, which will be used to concatenate the
concentrations along the ``"species"`` dimension, in case the
simulation has multiple species and thus multiple transport models.
If None, transport model names will be used as species names.
Returns
-------
budget: Dict[str, xr.DataArray|xu.UgridDataArray]
DataArray contains float64 data of the budgets, with dimensions ("time",
"layer", "y", "x").
"""
return self._open_output(
"budget-transport",
species_ls=species_ls,
simulation_start_time=simulation_start_time,
time_unit=time_unit,
merge_to_dataset=True,
flowja=False,
)
[docs]
def open_flow_budget(
self,
flowja: bool = False,
simulation_start_time: Optional[np.datetime64] = None,
time_unit: Optional[str] = "d",
) -> GridDataArray | GridDataset:
"""
Open flow budgets of finished simulation, requires that the ``run``
method has been called.
The data is lazily read per timestep and automatically converted into
(dense) xr.DataArrays or xu.UgridDataArrays, for DIS and DISV
respectively. The conversion is done via the information stored in the
Binary Grid file (GRB).
The ``flowja`` argument controls whether the flow-ja-face array (if
present) is returned in grid form as "as is". By default
``flowja=False`` and the array is returned in "grid form", meaning:
* DIS: in right, front, and lower face flow. All flows are placed in
the cell.
* DISV: in horizontal and lower face flow.the horizontal flows are
placed on the edges and the lower face flow is placed on the faces.
When ``flowja=True``, the flow-ja-face array is returned as it is found in
the CBC file, with a flow for every cell to cell connection. Additionally,
a ``connectivity`` DataArray is returned describing for every cell (n) its
connected cells (m).
Parameters
----------
flowja: bool, default value: False
Whether to return the flow-ja-face values "as is" (``True``) or in a
grid form (``False``).
Returns
-------
budget: Dict[str, xr.DataArray|xu.UgridDataArray]
DataArray contains float64 data of the budgets, with dimensions ("time",
"layer", "y", "x").
Examples
--------
Make sure you write and run your model first
>>> simulation.write(path/to/model)
>>> simulation.run()
Then open budgets:
>>> budget = simulation.open_flow_budget()
Check the contents:
>>> print(budget.keys())
Get the drainage budget, compute a time mean for the first layer:
>>> drn_budget = budget["drn]
>>> mean = drn_budget.sel(layer=1).mean("time")
"""
return self._open_output(
"budget-flow",
flowja=flowja,
simulation_start_time=simulation_start_time,
time_unit=time_unit,
merge_to_dataset=True,
)
[docs]
def open_concentration(
self,
species_ls: Optional[list[str]] = None,
dry_nan: bool = False,
simulation_start_time: Optional[np.datetime64] = None,
time_unit: Optional[str] = "d",
) -> GridDataArray:
"""
Open concentration of finished simulation, requires that the ``run``
method has been called.
The data is lazily read per timestep and automatically converted into
(dense) xr.DataArrays or xu.UgridDataArrays, for DIS and DISV
respectively. The conversion is done via the information stored in the
Binary Grid file (GRB).
Parameters
----------
species_ls: list of strings, default value: None.
List of species names, which will be used to concatenate the
concentrations along the ``"species"`` dimension, in case the
simulation has multiple species and thus multiple transport models.
If None, transport model names will be used as species names.
dry_nan: bool, default value: False.
Whether to convert dry values to NaN.
Returns
-------
concentration: Union[xr.DataArray, xu.UgridDataArray]
Examples
--------
Make sure you write and run your model first
>>> simulation.write(path/to/model)
>>> simulation.run()
Then open concentrations:
>>> concentration = simulation.open_concentration()
"""
return self._open_output(
"concentration",
species_ls=species_ls,
dry_nan=dry_nan,
simulation_start_time=simulation_start_time,
time_unit=time_unit,
)
def _open_output(self, output: str, **settings) -> GridDataArray | GridDataset:
"""
Opens output of one or multiple models.
Parameters
----------
output: str
Output variable name to open
**settings:
Extra settings that need to be passed through to the respective
output function.
"""
modeltype = OUTPUT_MODEL_MAPPING[output]
modelnames = self.get_models_of_type(modeltype._model_id).keys()
if len(modelnames) == 0:
modeltype = OUTPUT_MODEL_MAPPING[output]
raise ValueError(
f"Could not find any models of appropriate type for {output}, "
f"make sure a model of type {modeltype} is assigned to simulation."
)
if output in ["head", "budget-flow"]:
return self._open_single_output(list(modelnames), output, **settings)
elif output in ["concentration", "budget-transport"]:
return self._concat_species(output, **settings)
else:
raise RuntimeError(
f"Unexpected error when opening {output} for {modelnames}"
)
return
def _open_single_output(
self, modelnames: list[str], output: str, **settings
) -> GridDataArray | GridDataset:
"""
Open single output, e.g. concentration of single species, or heads. This
can be output of partitioned models that need to be merged.
"""
if len(modelnames) == 0:
modeltype = OUTPUT_MODEL_MAPPING[output]
raise ValueError(
f"Could not find any models of appropriate type for {output}, "
f"make sure a model of type {modeltype} is assigned to simulation."
)
elif len(modelnames) == 1:
modelname = next(iter(modelnames))
return self._open_single_output_single_model(modelname, output, **settings)
elif self.is_split():
if "budget" in output:
return self._merge_budgets(modelnames, output, **settings)
else:
return self._merge_states(modelnames, output, **settings)
raise ValueError("error in _open_single_output")
def _merge_states(
self, modelnames: list[str], output: str, **settings
) -> GridDataArray:
state_partitions = []
for modelname in modelnames:
state_partitions.append(
self._open_single_output_single_model(modelname, output, **settings)
)
return merge_partitions(state_partitions)
def _merge_and_assign_exchange_budgets(self, cbc: GridDataset) -> GridDataset:
"""
Merge and assign exchange budgets to cell by cell budgets:
cbc[[gwf-gwf_1, gwf-gwf_3]] to cbc[gwf-gwf]
"""
exchange_names = [
key
for key in cast(Iterable[str], cbc.keys())
if (("gwf-gwf" in key) or ("gwt-gwt" in key))
]
exchange_budgets = cbc[exchange_names].to_array().sum(dim="variable")
cbc = cbc.drop_vars(exchange_names)
# "gwf-gwf" or "gwt-gwt"
exchange_key = exchange_names[0].split("_")[1]
cbc[exchange_key] = exchange_budgets
return cbc
def _pad_missing_variables(self, cbc_per_partition: list[GridDataset]) -> None:
"""
Boundary conditions can be missing in certain partitions, as do their
budgets, in which case we manually assign an empty grid of nans.
"""
dims_per_unique_key = {
key: cbc[key].dims for cbc in cbc_per_partition for key in cbc.keys()
}
for cbc in cbc_per_partition:
missing_keys = set(dims_per_unique_key.keys()) - set(cbc.keys())
for missing in missing_keys:
missing_dims = dims_per_unique_key[missing]
missing_coords = {dim: cbc.coords[dim] for dim in missing_dims}
shape = tuple([len(missing_coords[dim]) for dim in missing_dims])
chunks = (1,) + shape[1:]
missing_data = dask.array.full(shape, np.nan, chunks=chunks)
missing_grid = xr.DataArray(
missing_data, dims=missing_dims, coords=missing_coords
)
if isinstance(cbc, xu.UgridDataset):
missing_grid = xu.UgridDataArray(
missing_grid,
grid=cbc.ugrid.grid,
)
cbc[missing] = missing_grid
def _merge_budgets(
self, modelnames: list[str], output: str, **settings
) -> GridDataset:
if settings["flowja"] is True:
raise ValueError("``flowja`` cannot be set to True when merging budgets.")
cbc_per_partition = []
for modelname in modelnames:
cbc = self._open_single_output_single_model(modelname, output, **settings)
# Merge and assign exchange budgets to dataset
# FUTURE: Refactor to insert these exchange budgets in horizontal
# flows.
cbc = self._merge_and_assign_exchange_budgets(cbc)
if not is_unstructured(cbc):
cbc = cbc.where(self[modelname].domain, other=np.nan)
cbc_per_partition.append(cbc)
self._pad_missing_variables(cbc_per_partition)
return merge_partitions(cbc_per_partition)
def _concat_species(
self, output: str, species_ls: Optional[list[str]] = None, **settings
) -> GridDataArray | GridDataset:
# groupby flow model, to somewhat enforce consistent transport model
# ordening. Say:
# F = Flow model, T = Transport model
# a = species "a", b = species "b"
# 1 = partition 1, 2 = partition 2
# then this:
# F1Ta1 F1Tb1 F2Ta2 F2Tb2 -> F1: [Ta1, Tb1], F2: [Ta2, Tb2]
# F1Ta1 F2Tb1 F1Ta1 F2Tb2 -> F1: [Ta1, Tb1], F2: [Ta2, Tb2]
tpt_models_per_flow_model = self._get_transport_models_per_flow_model()
all_tpt_names = list(tpt_models_per_flow_model.values())
# [[Ta_1, Tb_1], [Ta_2, Tb_2]] -> [[Ta_1, Ta_2], [Tb_1, Tb_2]]
# [[Ta, Tb]] -> [[Ta], [Tb]]
tpt_names_per_species = list(zip(*all_tpt_names))
if self.is_split():
# [[Ta_1, Tb_1], [Ta_2, Tb_2]] -> [Ta, Tb]
unpartitioned_modelnames = [
tpt_name.rpartition("_")[0] for tpt_name in all_tpt_names[0]
]
else:
# [[Ta, Tb]] -> [Ta, Tb]
unpartitioned_modelnames = all_tpt_names[0]
if not species_ls:
species_ls = unpartitioned_modelnames
if len(species_ls) != len(tpt_names_per_species):
raise ValueError(
"species_ls does not equal the number of transport models, "
f"expected length {len(tpt_names_per_species)}, received {species_ls}"
)
if len(species_ls) == 1:
return self._open_single_output(
list(tpt_names_per_species[0]), output, **settings
)
# Concatenate species
outputs = []
for species, tpt_names in zip(species_ls, tpt_names_per_species):
output_data = self._open_single_output(list(tpt_names), output, **settings)
output_data = output_data.assign_coords(species=species)
outputs.append(output_data)
return concat(outputs, dim="species")
def _open_single_output_single_model(
self, modelname: str, output: str, **settings
) -> GridDataArray | GridDataset:
"""
Opens single output of single model
Parameters
----------
modelname: str
Name of groundwater model from which output should be read.
output: str
Output variable name to open.
**settings:
Extra settings that need to be passed through to the respective
output function.
"""
open_func = OUTPUT_FUNC_MAPPING[output]
expected_modeltype = OUTPUT_MODEL_MAPPING[output]
if self.directory is None:
raise RuntimeError(f"Simulation {self.name} has not been written yet.")
model_path = self.directory / modelname
# Get model
model = self[modelname]
if not isinstance(model, expected_modeltype):
raise TypeError(
f"{modelname} not a {expected_modeltype}, instead got {type(model)}"
)
# Get output file path
oc_key = model._get_pkgkey("oc")
oc_pkg = model[oc_key]
# Ensure "-transport" and "-flow" are stripped from "budget"
oc_output = output.split("-")[0]
output_path = oc_pkg._get_output_filepath(model_path, oc_output)
# Force path to always include simulation directory.
output_path = self.directory / output_path
grb_path = self._get_grb_path(modelname)
if not output_path.exists():
raise RuntimeError(
f"Could not find output in {output_path}, check if you already ran simulation {self.name}"
)
return open_func(output_path, grb_path, **settings)
def _get_flow_modelname_coupled_to_transport_model(
self, transport_modelname: str
) -> str:
"""
Get name of flow model coupled to transport model, throws error if
multiple flow models are couple to 1 transport model.
"""
exchanges = self.get_exchange_relationships()
coupled_flow_models = [
i[2]
for i in exchanges
if (i[3] == transport_modelname) & (i[0] == "GWF6-GWT6")
]
if len(coupled_flow_models) != 1:
raise ValueError(
f"Exactly one flow model must be coupled to transport model {transport_modelname}, got: {coupled_flow_models}"
)
return coupled_flow_models[0]
def _get_grb_path(self, modelname: str) -> Path:
"""
Finds appropriate grb path belonging to modelname. Grb files are not
written for transport models, so this method always returns a path to a
flowmodel. In case of a transport model, it returns the path to the grb
file its coupled flow model.
"""
if self.directory is None:
raise ValueError("Directory not set")
model = self[modelname]
# Get grb path
if isinstance(model, GroundwaterTransportModel):
flow_model_name = self._get_flow_modelname_coupled_to_transport_model(
modelname
)
flow_model_path = self.directory / flow_model_name
else:
flow_model_path = self.directory / modelname
diskey = model._get_diskey()
dis_id = model[diskey]._pkg_id
return flow_model_path / f"{diskey}.{dis_id}.grb"
[docs]
@standard_log_decorator()
def dump(
self,
directory=".",
validate: bool = True,
mdal_compliant: bool = False,
crs=None,
) -> None:
"""
Dump simulation to files. Writes a model definition as .TOML file, which
points to data for each package. Each package is stored as a separate
NetCDF. Structured grids are saved as regular NetCDFs, unstructured
grids are saved as UGRID NetCDF. Structured grids are always made GDAL
compliant, unstructured grids can be made MDAL compliant optionally.
Parameters
----------
directory: str or Path, optional
directory to dump simulation into. Defaults to current working directory.
validate: bool, optional
Whether to validate simulation data. Defaults to True.
mdal_compliant: bool, optional
Convert data with
:func:`imod.prepare.spatial.mdal_compliant_ugrid2d` to MDAL
compliant unstructured grids. Defaults to False.
crs: Any, optional
Anything accepted by rasterio.crs.CRS.from_user_input
Requires ``rioxarray`` installed.
"""
directory = pathlib.Path(directory)
directory.mkdir(parents=True, exist_ok=True)
toml_content: DefaultDict[str, dict] = collections.defaultdict(dict)
for key, value in self.items():
cls_name = type(value).__name__
if isinstance(value, Modflow6Model):
model_toml_path = value.dump(
directory, key, validate, mdal_compliant, crs
)
toml_content[cls_name][key] = model_toml_path.relative_to(
directory
).as_posix()
elif key in ["gwtgwf_exchanges", "split_exchanges"]:
toml_content[key] = collections.defaultdict(list)
for exchange_package in self[key]:
exchange_type, filename, _, _ = exchange_package.get_specification()
exchange_class_short = type(exchange_package).__name__
path = f"{filename}.nc"
exchange_package.dataset.to_netcdf(directory / path)
toml_content[key][exchange_class_short].append(path)
else:
path = f"{key}.nc"
value.dataset.to_netcdf(directory / path)
toml_content[cls_name][key] = path
with open(directory / f"{self.name}.toml", "wb") as f:
tomli_w.dump(toml_content, f)
return
@staticmethod
def from_file(toml_path):
classes = {
item_cls.__name__: item_cls
for item_cls in (
GroundwaterFlowModel,
GroundwaterTransportModel,
imod.mf6.TimeDiscretization,
imod.mf6.Solution,
imod.mf6.GWFGWF,
imod.mf6.GWFGWT,
imod.mf6.GWTGWT,
)
}
toml_path = pathlib.Path(toml_path)
with open(toml_path, "rb") as f:
toml_content = tomli.load(f)
simulation = Modflow6Simulation(name=toml_path.stem)
for key, entry in toml_content.items():
if key not in ["gwtgwf_exchanges", "split_exchanges"]:
item_cls = classes[key]
for name, filename in entry.items():
path = toml_path.parent / filename
simulation[name] = item_cls.from_file(path)
else:
simulation[key] = []
for exchange_class, exchange_list in entry.items():
item_cls = classes[exchange_class]
for filename in exchange_list:
path = toml_path.parent / filename
simulation[key].append(item_cls.from_file(path))
return simulation
def get_exchange_relationships(self):
result = []
if "gwtgwf_exchanges" in self:
for exchange in self["gwtgwf_exchanges"]:
result.append(exchange.get_specification())
# exchange for splitting models
if self.is_split():
for exchange in self["split_exchanges"]:
result.append(exchange.get_specification())
return result
def get_models_of_type(self, model_id) -> dict[str, IModel]:
return {
k: v
for k, v in self.items()
if isinstance(v, Modflow6Model) and (v.model_id == model_id)
}
def get_models(self):
return {k: v for k, v in self.items() if isinstance(v, Modflow6Model)}
[docs]
def clip_box(
self,
time_min: Optional[cftime.datetime | np.datetime64 | str] = None,
time_max: Optional[cftime.datetime | np.datetime64 | str] = None,
layer_min: Optional[int] = None,
layer_max: Optional[int] = None,
x_min: Optional[float] = None,
x_max: Optional[float] = None,
y_min: Optional[float] = None,
y_max: Optional[float] = None,
states_for_boundary: Optional[dict[str, GridDataArray]] = None,
) -> Modflow6Simulation:
"""
Clip a simulation by a bounding box (time, layer, y, x).
Slicing intervals may be half-bounded, by providing None:
* To select 500.0 <= x <= 1000.0:
``clip_box(x_min=500.0, x_max=1000.0)``.
* To select x <= 1000.0: ``clip_box(x_min=None, x_max=1000.0)``
or ``clip_box(x_max=1000.0)``.
* To select x >= 500.0: ``clip_box(x_min = 500.0, x_max=None.0)``
or ``clip_box(x_min=1000.0)``.
Parameters
----------
time_min: optional
time_max: optional
layer_min: optional, int
layer_max: optional, int
x_min: optional, float
x_max: optional, float
y_min: optional, float
y_max: optional, float
states_for_boundary : optional, Dict[pkg_name:str, boundary_values:Union[xr.DataArray, xu.UgridDataArray]]
Returns
-------
clipped : Simulation
"""
if self.is_split():
raise RuntimeError(
"Unable to clip simulation. Clipping can only be done on simulations that haven't been split."
+ "Therefore clipping should be done before splitting the simulation."
)
if not self.has_one_flow_model():
raise ValueError(
"Unable to clip simulation. Clipping can only be done on simulations that have a single flow model ."
)
for model_name, model in self.get_models().items():
supported, error_with_object = model.is_clipping_supported()
if not supported:
raise ValueError(
f"simulation cannot be clipped due to presence of package '{error_with_object}' in model '{model_name}'"
)
clipped = type(self)(name=self.name)
for key, value in self.items():
state_for_boundary = (
None if states_for_boundary is None else states_for_boundary.get(key)
)
if isinstance(value, Modflow6Model):
clipped[key] = value.clip_box(
time_min=time_min,
time_max=time_max,
layer_min=layer_min,
layer_max=layer_max,
x_min=x_min,
x_max=x_max,
y_min=y_min,
y_max=y_max,
state_for_boundary=state_for_boundary,
)
elif isinstance(value, Package):
clipped[key] = value.clip_box(
time_min=time_min,
time_max=time_max,
layer_min=layer_min,
layer_max=layer_max,
x_min=x_min,
x_max=x_max,
y_min=y_min,
y_max=y_max,
)
else:
raise ValueError(f"object of type {type(value)} cannot be clipped.")
return clipped
[docs]
def split(self, submodel_labels: GridDataArray) -> Modflow6Simulation:
"""
Split a simulation in different partitions using a submodel_labels array.
The submodel_labels array defines how a simulation will be split. The array should have the same topology as
the domain being split i.e. similar shape as a layer in the domain. The values in the array indicate to
which partition a cell belongs. The values should be zero or greater.
The method return a new simulation containing all the split models and packages
"""
if self.is_split():
raise RuntimeError(
"Unable to split simulation. Splitting can only be done on simulations that haven't been split."
)
if not self.has_one_flow_model():
raise ValueError(
"splitting of simulations with more (or less) than 1 flow model currently not supported."
)
transport_models = self.get_models_of_type("gwt6")
flow_models = self.get_models_of_type("gwf6")
if not any(flow_models) and not any(transport_models):
raise ValueError("a simulation without any models cannot be split.")
original_models = {**flow_models, **transport_models}
for model_name, model in original_models.items():
supported, error_with_object = model.is_splitting_supported()
if not supported:
raise ValueError(
f"simulation cannot be split due to presence of package '{error_with_object}' in model '{model_name}'"
)
original_packages = get_packages(self)
partition_info = create_partition_info(submodel_labels)
exchange_creator: ExchangeCreator_Unstructured | ExchangeCreator_Structured
if is_unstructured(submodel_labels):
exchange_creator = ExchangeCreator_Unstructured(
submodel_labels, partition_info
)
else:
exchange_creator = ExchangeCreator_Structured(
submodel_labels, partition_info
)
new_simulation = imod.mf6.Modflow6Simulation(f"{self.name}_partioned")
for package_name, package in {**original_packages}.items():
new_simulation[package_name] = deepcopy(package)
for model_name, model in original_models.items():
solution_name = self.get_solution_name(model_name)
new_simulation[solution_name].remove_model_from_solution(model_name)
for submodel_partition_info in partition_info:
new_model_name = f"{model_name}_{submodel_partition_info.id}"
new_simulation[new_model_name] = slice_model(
submodel_partition_info, model
)
new_simulation[solution_name].add_model_to_solution(new_model_name)
exchanges: list[Any] = []
for flow_model_name, flow_model in flow_models.items():
exchanges += exchange_creator.create_gwfgwf_exchanges(
flow_model_name, flow_model.domain.layer
)
if any(transport_models):
for tpt_model_name in transport_models:
exchanges += exchange_creator.create_gwtgwt_exchanges(
tpt_model_name, flow_model_name, model.domain.layer
)
new_simulation._add_modelsplit_exchanges(exchanges)
new_simulation._update_buoyancy_packages()
new_simulation._set_flow_exchange_options()
new_simulation._set_transport_exchange_options()
new_simulation._update_ssm_packages()
new_simulation._filter_inactive_cells_from_exchanges()
return new_simulation
[docs]
def regrid_like(
self,
regridded_simulation_name: str,
target_grid: GridDataArray,
validate: bool = True,
) -> "Modflow6Simulation":
"""
This method creates a new simulation object. The models contained in the new simulation are regridded versions
of the models in the input object (this).
Time discretization and solver settings are copied.
Parameters
----------
regridded_simulation_name: str
name given to the output simulation
target_grid: xr.DataArray or xu.UgridDataArray
discretization onto which the models in this simulation will be regridded
validate: bool
set to true to validate the regridded packages
Returns
-------
a new simulation object with regridded models
"""
return _regrid_like(self, regridded_simulation_name, target_grid, validate)
def _add_modelsplit_exchanges(self, exchanges_list: list[GWFGWF]) -> None:
if not self.is_split():
self["split_exchanges"] = []
self["split_exchanges"].extend(exchanges_list)
def _set_flow_exchange_options(self) -> None:
# collect some options that we will auto-set
for exchange in self["split_exchanges"]:
if isinstance(exchange, GWFGWF):
model_name_1 = exchange.dataset["model_name_1"].values[()]
model_1 = self[model_name_1]
exchange.set_options(
save_flows=model_1["oc"].is_budget_output,
dewatered=model_1["npf"].is_dewatered,
variablecv=model_1["npf"].is_variable_vertical_conductance,
xt3d=model_1["npf"].get_xt3d_option(),
newton=model_1.is_use_newton(),
)
def _set_transport_exchange_options(self) -> None:
for exchange in self["split_exchanges"]:
if isinstance(exchange, GWTGWT):
model_name_1 = exchange.dataset["model_name_1"].values[()]
model_1 = self[model_name_1]
advection_key = model_1._get_pkgkey("adv")
dispersion_key = model_1._get_pkgkey("dsp")
scheme = None
xt3d_off = None
xt3d_rhs = None
if advection_key is not None:
scheme = model_1[advection_key].dataset["scheme"].values[()]
if dispersion_key is not None:
xt3d_off = model_1[dispersion_key].dataset["xt3d_off"].values[()]
xt3d_rhs = model_1[dispersion_key].dataset["xt3d_rhs"].values[()]
exchange.set_options(
save_flows=model_1["oc"].is_budget_output,
adv_scheme=scheme,
dsp_xt3d_off=xt3d_off,
dsp_xt3d_rhs=xt3d_rhs,
)
def _filter_inactive_cells_from_exchanges(self) -> None:
for ex in self["split_exchanges"]:
for i in [1, 2]:
self._filter_inactive_cells_exchange_domain(ex, i)
def _filter_inactive_cells_exchange_domain(self, ex: GWFGWF, i: int) -> None:
"""Filters inactive cells from one exchange domain inplace"""
modelname = ex[f"model_name_{i}"].values[()]
domain = self[modelname].domain
layer = ex.dataset["layer"] - 1
id = ex.dataset[f"cell_id{i}"] - 1
if is_unstructured(domain):
exchange_cells = {
"layer": layer,
"mesh2d_nFaces": id,
}
else:
exchange_cells = {
"layer": layer,
"y": id.sel({f"cell_dims{i}": f"row_{i}"}),
"x": id.sel({f"cell_dims{i}": f"column_{i}"}),
}
exchange_domain = domain.isel(exchange_cells)
active_exchange_domain = exchange_domain.where(exchange_domain.values > 0)
active_exchange_domain = active_exchange_domain.dropna("index")
ex.dataset = ex.dataset.sel(index=active_exchange_domain["index"])
def get_solution_name(self, model_name: str) -> Optional[str]:
for k, v in self.items():
if isinstance(v, Solution):
if model_name in v.dataset["modelnames"]:
return k
return None
def __repr__(self) -> str:
typename = type(self).__name__
INDENT = " "
attrs = [
f"{typename}(",
f"{INDENT}name={repr(self.name)},",
f"{INDENT}directory={repr(self.directory)}",
]
items = [
f"{INDENT}{repr(key)}: {type(value).__name__},"
for key, value in self.items()
]
# Place the emtpy dict on the same line. Looks silly otherwise.
if items:
content = attrs + ["){"] + items + ["}"]
else:
content = attrs + ["){}"]
return "\n".join(content)
def _get_transport_models_per_flow_model(self) -> dict[str, list[str]]:
flow_models = self.get_models_of_type("gwf6")
transport_models = self.get_models_of_type("gwt6")
# exchange for flow and transport
result = collections.defaultdict(list)
for flow_model_name in flow_models:
flow_model = self[flow_model_name]
for tpt_model_name in transport_models:
tpt_model = self[tpt_model_name]
if is_equal(tpt_model.domain, flow_model.domain):
result[flow_model_name].append(tpt_model_name)
return result
def _generate_gwfgwt_exchanges(self) -> list[GWFGWT]:
exchanges = []
flow_transport_mapping = self._get_transport_models_per_flow_model()
for flow_name, tpt_models_of_flow_model in flow_transport_mapping.items():
if len(tpt_models_of_flow_model) > 0:
for transport_model_name in tpt_models_of_flow_model:
exchanges.append(GWFGWT(flow_name, transport_model_name))
return exchanges
def _update_ssm_packages(self) -> None:
flow_transport_mapping = self._get_transport_models_per_flow_model()
for flow_name, tpt_models_of_flow_model in flow_transport_mapping.items():
flow_model = self[flow_name]
for tpt_model_name in tpt_models_of_flow_model:
tpt_model = self[tpt_model_name]
ssm_key = tpt_model._get_pkgkey("ssm")
if ssm_key is not None:
old_ssm_package = tpt_model.pop(ssm_key)
state_variable_name = old_ssm_package.dataset[
"auxiliary_variable_name"
].values[0]
ssm_package = SourceSinkMixing.from_flow_model(
flow_model, state_variable_name, is_split=self.is_split()
)
if ssm_package is not None:
tpt_model[ssm_key] = ssm_package
def _update_buoyancy_packages(self) -> None:
flow_transport_mapping = self._get_transport_models_per_flow_model()
for flow_name, tpt_models_of_flow_model in flow_transport_mapping.items():
flow_model = self[flow_name]
flow_model.update_buoyancy_package(tpt_models_of_flow_model)
def is_split(self) -> bool:
return "split_exchanges" in self.keys()
def has_one_flow_model(self) -> bool:
flow_models = self.get_models_of_type("gwf6")
return len(flow_models) == 1
def mask_all_models(
self,
mask: GridDataArray,
):
"""
This function applies a mask to all models in a simulation, provided they use
the same discretization. The method parameter "mask" is an idomain-like array.
Masking will overwrite idomain with the mask where the mask is 0 or -1.
Where the mask is 1, the original value of idomain will be kept.
Masking will update the packages accordingly, blanking their input where needed,
and is therefore not a reversible operation.
Parameters
----------
mask: xr.DataArray, xu.UgridDataArray of ints
idomain-like integer array. 1 sets cells to active, 0 sets cells to inactive,
-1 sets cells to vertical passthrough
"""
_mask_all_models(self, mask)
[docs]
@classmethod
@standard_log_decorator()
def from_imod5_data(
cls,
imod5_data: dict[str, dict[str, GridDataArray]],
period_data: dict[str, dict[str, GridDataArray]],
allocation_options: SimulationAllocationOptions,
distributing_options: SimulationDistributingOptions,
times: list[datetime],
regridder_types: dict[str, RegridMethodType] = {},
) -> "Modflow6Simulation":
"""
Imports a GroundwaterFlowModel (GWF) from the data in an IMOD5 project file.
It adds the packages for which import from imod5 is supported.
Some packages (like OC) must be added manually later.
Parameters
----------
imod5_data: dict[str, dict[str, GridDataArray]]
dictionary containing the arrays mentioned in the project file as xarray datasets,
under the key of the package type to which it belongs
allocation_options: SimulationAllocationOptions
object containing the allocation options per package type.
If you want a package to have a different allocation option,
then it should be imported separately
distributing_options: SimulationDistributingOptions
object containing the conductivity distribution options per package type.
If you want a package to have a different allocation option,
then it should be imported separately
times: list[datetime]
time discretization of the model to be imported.
regridder_types: dict[str, RegridMethodType]
the key is the package name. The value is the RegridMethodType
object containing the settings for regridding the package with the
specified key
Returns
-------
"""
simulation = Modflow6Simulation("imported_simulation")
simulation._validation_context.strict_well_validation = False
# import GWF model,
groundwaterFlowModel = GroundwaterFlowModel.from_imod5_data(
imod5_data,
period_data,
allocation_options,
distributing_options,
times,
regridder_types,
)
simulation["imported_model"] = groundwaterFlowModel
# generate ims package
solution = SolutionPresetModerate(
["imported_model"],
print_option="all",
)
simulation["ims"] = solution
simulation.create_time_discretization(additional_times=times)
return simulation
