"""States component."""
import copy
import os
import numpy as np
import pandas as pd
from .decorators import apply_to_each_input
from .base_spatial import SpatialComponent
from .plot_utils import show_slice_static, show_slice_interactive
from .parse_utils import read_ecl_bin
from .utils import get_single_path, get_multout_paths
FULL_STATE_KEYS = ('PRESSURE', 'RS', 'SGAS', 'SOIL', 'SWAT')
[docs]
class States(SpatialComponent):
"""States component of geological model."""
def __init__(self, *args, **kwargs):
self._dates = kwargs['dates'] if 'dates' in kwargs else pd.to_datetime([])
super().__init__(*args, **{k:v for k, v in kwargs.items() if k != 'dates'})
[docs]
def copy(self):
cp = super().copy()
cp.dates = copy.deepcopy(self.dates)
return cp
def _dump_hdf5_group(self, grp, compression, state, **kwargs):
super()._dump_hdf5_group(grp, compression, state, **kwargs)
grp = grp[self.class_name] if self.class_name in grp else grp.create_group(self.class_name)
grp.attrs['DATES'] = self.dates.astype(np.int64)
@property
def dates(self):
"""Dates for which states are available.
Returns
-------
pandas.DatetimeIndex
Dates.
"""
return self._dates
@dates.setter
def dates(self, value):
self._dates = value
@property
def n_timesteps(self):
"""Effective number of timesteps."""
if not self.attributes:
return 0
return np.min([x.shape[0] for _, x in self.items()])
[docs]
@apply_to_each_input
def apply(self, func, attr, *args, inplace=False, **kwargs):
"""Apply function to each timestamp of states attributes.
Parameters
----------
func : callable
A function to apply. Must accept data as its first argument.
attr : str, array-like
Attributes to get data from.
args : misc
Any additional positional arguments to ``func``.
inplace: bool
Modify сomponent inplace.
kwargs : misc
Any additional named arguments to ``func``.
Returns
-------
output : States
Transformed component.
"""
data = getattr(self, attr)
res = np.array([func(x, *args, **kwargs) for x in data])
if inplace:
setattr(self, attr, res)
return self
return res
@apply_to_each_input
def _to_spatial(self, attr):
"""Spatial order 'F' transformations."""
dimens = self.field.grid.dimens
self.pad_na(attr=attr)
return self.reshape(attr=attr, newshape=(-1,) + tuple(dimens),
order='F', inplace=True)
@apply_to_each_input
def _ravel(self, attr):
"""Ravel order 'F' transformations."""
return self.reshape(attr=attr, newshape=(self.n_timesteps, -1), order='F', inplace=False)
[docs]
@apply_to_each_input
def pad_na(self, attr, fill_na=0., inplace=True):
"""Add dummy cells into the state vector in the positions of non-active cells if necessary.
Parameters
----------
attr: str, array-like
Attributes to be padded with non-active cells.
actnum: array-like of type bool
Vector representing a mask of active and non-active cells.
fill_na: float
Value to be used as filler.
inplace: bool
Modify сomponent inplace.
Returns
-------
output : component if inplace else padded attribute.
"""
data = getattr(self, attr)
if np.prod(data.shape[1:]) == np.prod(self.field.grid.dimens):
return self if inplace else data
actnum = self.field.grid.actnum
if data.ndim > 2:
raise ValueError('Data should be raveled before padding.')
n_ts = data.shape[0]
actnum_ravel = actnum.ravel(order='F').astype(bool)
not_actnum_ravel = ~actnum_ravel
padded_data = np.empty(shape=(n_ts, actnum.size), dtype=float)
padded_data[..., actnum_ravel] = data
del data
padded_data[..., not_actnum_ravel] = fill_na
if inplace:
setattr(self, attr, padded_data)
return self
return padded_data
[docs]
@apply_to_each_input
def strip_na(self, attr):
"""Remove non-active cells from the state vector.
Parameters
----------
attr: str, array-like
Attributes to be stripped
actnum: array-like of type bool
Vector representing mask of active and non-active cells.
Returns
-------
output : stripped attribute.
Notes
-----
Outputs 1d array for each timestamp.
"""
data = self.ravel(attr)
actnum = self.field.grid.actnum
if data.shape[1] == np.sum(actnum):
return data
stripped_data = data[..., actnum.ravel(order='F')]
return stripped_data
def __getitem__(self, keys):
if isinstance(keys, str):
return super().__getitem__(keys)
out = self.__class__()
for attr, data in self.items():
data = data[keys].reshape((-1,) + data.shape[1:])
setattr(out, attr, data)
out.set_state(**self.state.as_dict())
return out
[docs]
def show_slice(self, attr, t=None, i=None, j=None, k=None, figsize=None, **kwargs):
"""Visualize slices of 4D states arrays. If no slice is specified, spatial slices
will be shown with interactive slider widgets.
Parameters
----------
attr : str
Attribute to show.
t : int or None, optional
Timestamp to show.
i : int or None, optional
Slice along x-axis to show.
j : int or None, optional
Slice along y-axis to show.
k : int or None, optional
Slice along z-axis to show.
figsize : array-like, optional
Output plot size.
kwargs : dict, optional
Additional keyword arguments for plot.
"""
if np.all([t is None, i is None, j is None, k is None]):
show_slice_interactive(self, attr, figsize=figsize, **kwargs)
else:
show_slice_static(self, attr, t=t, i=i, j=j, k=k, figsize=figsize, **kwargs)
return self
def _read_buffer(self, path_or_buffer, attr, **kwargs):
super()._read_buffer(path_or_buffer, attr, **kwargs)
return self.reshape(attr=attr, newshape=(1, -1))
def _load_ecl_binary(self, path_to_results, attrs, basename, logger=None, **kwargs):
"""Load states from binary ECLIPSE results files.
Parameters
----------
path_to_results : str
Path to the folder with precomputed results of hydrodynamical simulation
attrs : list or str
Keyword names to be loaded
logger : logger
Logger for messages.
**kwargs : dict, optional
Any kwargs to be passed to load method.
Returns
-------
states : States
States with loaded attributes.
"""
if attrs is None:
attrs = FULL_STATE_KEYS
rsspec_path = get_single_path(path_to_results, basename + '.RSSPEC')
if rsspec_path is not None:
self._load_ecl_rsspec(rsspec_path, logger=logger)
unifout_path = get_single_path(path_to_results, basename + '.UNRST', logger)
multout_paths = get_multout_paths(path_to_results, basename)
if unifout_path is not None:
return self._load_ecl_bin_unifout(unifout_path, attrs=attrs, logger=logger, **kwargs)
if multout_paths is not None:
return self._load_ecl_bin_multout(multout_paths, attrs=attrs, logger=logger, **kwargs)
if logger is not None:
logger.warning('The results in "%s" were not found!' % path_to_results)
return self
raise FileNotFoundError('The results in "%s" were not found!' % path_to_results)
def _load_ecl_rsspec(self, path, logger, subset=None, **kwargs):
_ = kwargs
data = read_ecl_bin(path, attrs=['TIME', 'ITIME'], sequential=True, subset=subset,
logger=logger)
dates = pd.to_datetime([])
if len(data['ITIME'][0]) > 1:
dates = pd.to_datetime(
[f'{row[3]}-{row[2]}-{row[1]}' for row in data['ITIME']]
)
else:
dates = pd.DatetimeIndex(
[pd.to_datetime(self.field.meta['START']) +
pd.Timedelta(v[0], 'days') for v in data['TIME']])
self._dates = dates
return self
def _load_hdf5_group(self, grp, attrs, raise_errors, logger, subset):
grp_tmp = grp[self.class_name]
if 'DATES' in grp.attrs:
self._dates = pd.to_datetime(grp_tmp.attrs['DATES'])
return super()._load_hdf5_group(grp, attrs, raise_errors, logger, subset)
def _load_ecl_bin_unifout(self, path, attrs, logger, subset=None, **kwargs):
"""Load states from .UNRST binary file.
Parameters
----------
path: str
Path to the .UNRST file.
attrs: list or str
Keyword names to be loaded from the file.
kwargs : dict, optional
Any kwargs to be passed to load method.
Returns
-------
states : States
States with loaded attributes.
"""
_ = kwargs
if isinstance(attrs, str):
attrs = [attrs]
states = read_ecl_bin(path, attrs, sequential=True, subset=subset, logger=logger)
for attr, x in states.items():
setattr(self, attr, np.array(x))
self.state.binary_attributes.append(attr)
return self
def _load_ecl_bin_multout(self, paths, attrs, logger, subset=None, **kwargs):
"""Load states from .X____ binary files.
Parameters
----------
paths: list
List of paths to .X____ files
attrs: list or str
Keyword names to be loaded from the files.
kwargs : dict, optional
Any kwargs to be passed to load method.
Returns
-------
states : States
States with loaded attributes.
"""
_ = kwargs
if isinstance(attrs, str):
attrs = [attrs]
states = {}
logger.info('Start reading X files.')
def is_in_subset(x):
fmt = os.path.splitext(x)[1]
timestep = int(fmt.lstrip('.X'))
criteria = timestep in subset
return criteria
paths = filter(is_in_subset if subset is not None else None, paths)
for path in paths:
state = read_ecl_bin(path, attrs, logger=logger)
for attr, x in state.items():
if attr not in states:
states[attr] = [x]
else:
states[attr].append(x)
logger.info('Finish reading X files.')
states = {attr: np.stack(x) for attr, x in states.items()}
for attr, x in states.items():
setattr(self, attr, x)
self.state.binary_attributes.append(attr)
return self
def _make_data_dump(self, attr, fmt=None, actnum=None, float_dtype=None, **kwargs):
"""Prepare data for dump."""
if fmt.upper() == 'ASCII':
data = self.ravel(attr=attr)
return data[0]
if fmt.upper() == 'HDF5':
if attr.upper() == 'DATES':
return self.dates.astype(np.int64)
if actnum is None:
data = self.ravel(attr=attr)
else:
data = self.strip_na(attr=attr)
return data if float_dtype is None else data.astype(float_dtype)
return super()._make_data_dump(attr, fmt=fmt, **kwargs)