from pathlib import Path
import jinja2
import numpy as np
import imod
from imod.wq.pkgbase import BoundaryCondition
def _column_order(df, variable):
"""
Return ordered columns, and associated timeseries columns.
"""
if "time" in df:
assoc_columns = ["time", variable]
if "layer" in df:
columns = ["x", "y", "id", "time", "layer", variable]
else:
columns = ["x", "y", "id", "time", variable]
else:
assoc_columns = None
if "layer" in df:
columns = ["x", "y", variable, "layer"]
else:
columns = ["x", "y", variable]
return columns, assoc_columns
[docs]class Well(BoundaryCondition):
"""
The Well package is used to simulate a specified flux to individual cells
and specified in units of length3/time.
Parameters
----------
id_name: str or list of str
name of the well(s).
x: float or list of floats
x coordinate of the well(s).
y: float or list of floats
y coordinate of the well(s).
rate: float or list of floats.
pumping rate in the well(s).
layer: "None" or int, optional
layer from which the pumping takes place.
time: "None" or listlike of np.datetime64, datetime.datetime, pd.Timestamp,
cftime.datetime
time during which the pumping takes place. Only need to specify if model
is transient.
save_budget: bool, optional
is a flag indicating if the budget should be saved (IRIVCB).
Default is False.
"""
_pkg_id = "wel"
_template = jinja2.Template(
" {%- for time, timedict in wels.items() -%}"
" {%- for layer, value in timedict.items() %}\n"
" wel_p{{time}}_s{{system_index}}_l{{layer}} = {{value}}"
" {%- endfor -%}\n"
" {%- endfor -%}"
)
# TODO: implement well to concentration IDF and use ssm_template
# Ignored for now, since wells are nearly always extracting
[docs] def __init__(
self,
id_name,
x,
y,
rate,
layer=None,
time=None,
concentration=None,
save_budget=False,
):
super().__init__()
variables = {
"id_name": id_name,
"x": x,
"y": y,
"rate": rate,
"layer": layer,
"time": time,
"concentration": concentration,
}
variables = {k: np.atleast_1d(v) for k, v in variables.items() if v is not None}
length = max(map(len, variables.values()))
index = np.arange(1, length + 1)
self["index"] = index
for k, v in variables.items():
if v.size == index.size:
self[k] = ("index", v)
elif v.size == 1:
self[k] = ("index", np.full(length, v))
else:
raise ValueError(f"Length of {k} does not match other arguments")
self["save_budget"] = save_budget
def _max_active_n(self, varname, nlayer, nrow, ncol):
"""
Determine the maximum active number of cells that are active
during a stress period.
Parameters
----------
varname : str
name of the variable to use to calculate the maximum number of
active cells. Not used for well, here for polymorphism.
nlayer, nrow, ncol : int
"""
nmax = np.unique(self["id_name"]).size
if "layer" not in self.dataset.coords: # Then it applies to every layer
nmax *= nlayer
self._cellcount = nmax
self._ssm_cellcount = nmax
return nmax
def _compose_values_layer(self, directory, nlayer, name, time=None, compress=True):
values = {}
d = {"directory": directory, "name": name, "extension": ".ipf"}
if time is None:
if "layer" in self.dataset:
for layer in np.unique(self.dataset["layer"]):
layer = int(layer)
d["layer"] = layer
values[layer] = self._compose_path(d)
else:
values["$"] = self._compose_path(d)
else:
d["time"] = time
if "layer" in self.dataset:
# Since the well data is in long table format, it's the only
# input that has to be inspected.
select = np.argwhere((self.dataset["time"] == time).values)
for layer in np.unique(self.dataset["layer"].values[select]):
d["layer"] = layer
values[layer] = self._compose_path(d)
else:
values["?"] = self._compose_path(d)
if "layer" in self.dataset:
# Compose does not accept non-integers, so use 0, then replace
d["layer"] = 0
if np.unique(self.dataset["layer"].values).size == nlayer:
token_path = imod.util.compose(d).as_posix()
token_path = token_path.replace("_l0", "_l$")
values = {"$": token_path}
elif compress:
range_path = imod.util.compose(d).as_posix()
range_path = range_path.replace("_l0", "_l:")
# TODO: temporarily disable until imod-wq is fixed
values = self._compress_idflayers(values, range_path)
return values
def _compose_values_time(self, directory, name, globaltimes, nlayer):
# TODO: rename to _compose_values_timelayer?
values = {"?": self._compose_values_layer(directory, nlayer=nlayer, name=name)}
return values
def _render(self, directory, globaltimes, system_index, nlayer):
d = {"system_index": system_index}
d["wels"] = self._compose_values_time(directory, "rate", globaltimes, nlayer)
return self._template.render(d)
def _render_ssm(self, directory, globaltimes, nlayer):
if "concentration" in self.dataset.data_vars:
d = {"pkg_id": self._pkg_id}
name = "concentration"
if "species" in self.dataset["concentration"].coords:
concentration = {}
for species in self.dataset["concentration"]["species"].values:
concentration[species] = self._compose_values_time(
directory, f"{name}_c{species}", globaltimes, nlayer=nlayer
)
else:
concentration = {
1: self._compose_values_time(
directory, name, globaltimes, nlayer=nlayer
)
}
d["concentration"] = concentration
return self._ssm_template.render(d)
else:
return ""
def _save_layers(self, df, directory, name, variable):
d = {"directory": directory, "name": name, "extension": ".ipf"}
d["directory"].mkdir(exist_ok=True, parents=True)
if "time" in df:
itype = "timeseries"
else:
itype = None
columns, assoc_columns = _column_order(df, variable)
path = self._compose_path(d)
df = df[columns]
if "layer" in df:
for layer, layerdf in df.groupby("layer"):
# Ensure different IDs per layer are not overwritten.
layerdf["id"] = f"{name}_l{layer}/" + layerdf["id"].astype(str)
imod.ipf.save(
path=path, df=layerdf, itype=itype, assoc_columns=assoc_columns
)
else:
imod.ipf.save(path=path, df=df, itype=itype, assoc_columns=assoc_columns)
return
def save(self, directory):
directory = Path(directory)
all_species = [None]
if "concentration" in self.dataset.data_vars:
if "species" in self.dataset["concentration"].coords:
all_species = self.dataset["concentration"]["species"].values
df = self.dataset.to_dataframe().rename(columns={"id_name": "id"})
self._save_layers(df, directory, "rate", "rate")
# Loop over species if applicable
if "concentration" in self.dataset:
for species in all_species:
if species is not None:
ds = self.dataset.sel(species=species)
else:
ds = self.dataset
df = ds.to_dataframe().rename(columns={"id_name": "id"})
name = "concentration"
if species is not None:
name = f"{name}_c{species}"
self._save_layers(df, directory, name, "concentration")
return
def _pkgcheck(self, ibound=None):
# TODO: implement
pass
def repeat_stress(self, stress_repeats, use_cftime=False):
raise NotImplementedError(
"Well does not support repeated stresses: time-varying data is "
"saved into associated IPF files. Set explicit timeseries intead."
)
