xtgeo 4.14.0__cp311-cp311-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl

xtgeo/grid3d/_grid_import_xtgcpgeom.py

@@ -0,0 +1,375 @@
+"""Private module, Grid Import private functions for xtgeo based formats."""
+
+from __future__ import annotations
+
+import json
+from struct import unpack
+from typing import TYPE_CHECKING, Literal, Tuple, TypedDict, Union, overload
+
+import h5py
+import numpy as np
+from typing_extensions import NotRequired, TypeAlias
+
+import xtgeo.common.sys as xsys
+from xtgeo.common import null_logger
+from xtgeo.grid3d._gridprop_import_xtgcpprop import _read_filelike
+
+logger = null_logger(__name__)
+
+if TYPE_CHECKING:
+    from xtgeo.io._file import FileWrapper
+
+LMIN: TypeAlias = Literal["min"]
+LMAX: TypeAlias = Literal["max"]
+IJKRANGE: TypeAlias = Tuple[
+    Union[int, LMIN],
+    Union[int, LMAX],
+    Union[int, LMIN],
+    Union[int, LMAX],
+    Union[int, LMIN],
+    Union[int, LMAX],
+]
+
+
+class _RESULT(TypedDict):
+    coordsv: np.ndarray
+    zcornsv: np.ndarray
+    actnumsv: np.ndarray
+    subgrids: NotRequired[dict[str, range]]
+
+
+class __REQUIRED_(TypedDict):
+    xshift: np.ndarray
+    xscale: np.ndarray
+    yshift: np.ndarray
+    yscale: np.ndarray
+    zshift: np.ndarray
+    zscale: np.ndarray
+    subgrids: NotRequired[dict[str, int]]
+
+
+class _META(TypedDict):
+    _required_: __REQUIRED_
+
+
+@overload
+def convert_subgrids(sdict: dict[str, int]) -> dict[str, range]: ...
+
+
+@overload
+def convert_subgrids(sdict: None) -> None: ...
+
+
+def convert_subgrids(sdict: dict[str, int] | None) -> dict[str, range] | None:
+    # flake8: noqa: F841
+    """Convert a simplified ordered dictionary to
+        subgrid required by Grid.
+
+    The simplified dictionary is on the form
+    {"name1": 3, "name2": 5}
+
+    Note that the input must be an dict!
+
+    """
+    if sdict is None:
+        return None
+
+    if not isinstance(sdict, dict):
+        raise ValueError("Input sdict is not an dict")
+
+    start = 1
+    return {name: range(start, start := start + nsub) for name, nsub in sdict.items()}
+
+
+def handle_metadata(
+    result: _RESULT,
+    meta: _META,
+    ncol: int,
+    nrow: int,
+    nlay: int,
+) -> None:
+    # meta _optional_ *may* contain xshift, xscale etc which in case must be taken
+    # into account
+    coordsv = result["coordsv"]
+    zcornsv = result["zcornsv"]
+    actnumsv = result["actnumsv"]
+    nncol = ncol + 1
+    nnrow = nrow + 1
+    nnlay = nlay + 1
+    req = meta["_required_"]
+    shi = req["xshift"]
+    coordsv[0::3] = np.where(shi != 0, coordsv[0::3] + shi, coordsv[0::3])
+    sca = req["xscale"]
+    coordsv[0::3] = np.where(sca != 1, coordsv[0::3] * sca, coordsv[0::3])
+    shi = req["yshift"]
+    coordsv[1::3] = np.where(shi != 0, coordsv[1::3] + shi, coordsv[1::3])
+    sca = req["yscale"]
+    coordsv[1::3] = np.where(sca != 1, coordsv[1::3] * sca, coordsv[1::3])
+    shi = req["zshift"]
+    coordsv[2::3] = np.where(shi != 0, coordsv[2::3] + shi, coordsv[2::3])
+    sca = req["zscale"]
+    coordsv[2::3] = np.where(sca != 1, coordsv[2::3] * sca, coordsv[2::3])
+
+    result["coordsv"] = coordsv.reshape((nncol, nnrow, 6)).astype(np.float64)
+    result["zcornsv"] = zcornsv.reshape((nncol, nnrow, nnlay, 4)).astype(np.float32)
+    result["actnumsv"] = actnumsv.reshape((ncol, nrow, nlay)).astype(np.int32)
+
+    if "subgrids" in req:
+        result["subgrids"] = convert_subgrids(req["subgrids"])
+
+
+def import_xtgcpgeom(
+    mfile: FileWrapper,
+    mmap: bool = False,
+) -> _RESULT:
+    """Using pure python for experimental grid geometry import."""
+    #
+    offset = 36
+    with _read_filelike(mfile.file, size=offset) as f:
+        # unpack header
+        swap, magic, nformat, ncol, nrow, nlay = unpack("= i i i q q q", f)
+
+    nncol = ncol + 1
+    nnrow = nrow + 1
+    nnlay = nlay + 1
+
+    if swap != 1 or magic != 1301:
+        raise ValueError(f"Error, swap magic are {swap} {magic}, expected is 1 1301")
+
+    # subformat processing, indicating number of bytes per datatype
+    # here, 844 is native XTGeo (float64, float32, int32)
+    if nformat not in (444, 844, 841, 881, 884):
+        raise ValueError(f"The subformat value {nformat} is not valid")
+
+    coordfmt, zcornfmt, actnumfmt = [int(nbyte) for nbyte in str(nformat)]
+
+    dtype_coordsv = f"float{coordfmt * 8}"
+    dtype_zcornsv = f"float{zcornfmt * 8}"
+    dtype_actnumv = f"int{actnumfmt * 8}"
+
+    ncoord = nncol * nnrow * 6
+    nzcorn = nncol * nnrow * nnlay * 4
+    nactnum = ncol * nrow * nlay
+
+    # read numpy arrays from file
+    coordsv = xsys.npfromfile(
+        mfile.file,
+        dtype=dtype_coordsv,
+        count=ncoord,
+        offset=offset,
+        mmap=mmap,
+    )
+
+    newoffset = offset + ncoord * coordfmt
+    zcornsv = xsys.npfromfile(
+        mfile.file,
+        dtype=dtype_zcornsv,
+        count=nzcorn,
+        offset=newoffset,
+        mmap=mmap,
+    )
+
+    newoffset += nzcorn * zcornfmt
+    actnumsv = xsys.npfromfile(
+        mfile.file,
+        dtype=dtype_actnumv,
+        count=nactnum,
+        offset=newoffset,
+        mmap=mmap,
+    )
+
+    newoffset += nactnum * actnumfmt
+    result = _RESULT(coordsv=coordsv, zcornsv=zcornsv, actnumsv=actnumsv)
+
+    # read metadata which will be at position offet + nfloat*narr +13
+    with _read_filelike(mfile.file, seek=newoffset + 13) as f:
+        meta = json.loads(f, object_pairs_hook=dict)
+
+    handle_metadata(result, meta, ncol, nrow, nlay)
+    return result
+
+
+def import_hdf5_cpgeom(
+    mfile: FileWrapper,
+    ijkrange: IJKRANGE | None = None,
+    zerobased: bool = False,
+) -> _RESULT:
+    """Experimental grid geometry import using hdf5."""
+    #
+    with h5py.File(mfile.name, "r") as h5h:
+        grp = h5h["CornerPointGeometry"]
+
+        idcode = grp.attrs["format-idcode"]
+        provider = grp.attrs["provider"]
+        if idcode != 1301:
+            raise ValueError(f"Wrong id code: {idcode}")
+        logger.info("Provider is %s", provider)
+
+        jmeta = grp.attrs["metadata"]
+        meta = json.loads(jmeta, object_pairs_hook=dict)
+
+        req = meta["_required_"]
+        ncol = req["ncol"]
+        nrow = req["nrow"]
+        nlay = req["nlay"]
+
+        if ijkrange is not None:
+            incoord, inzcorn, inactnum, ncol, nrow, nlay = _partial_read(
+                h5h, req, ijkrange, zerobased
+            )
+            if (
+                "subgrids" in meta["_required_"]
+                and meta["_required_"]["subgrids"] is not None
+            ):
+                *_, k1, k2 = ijkrange
+                meta["_required_"]["subgrids"] = filter_subgrids_partial(
+                    meta["_required_"]["subgrids"],
+                    k1,
+                    k2,
+                    nlay,
+                    zerobased,
+                )
+        else:
+            incoord = grp["coord"][:, :, :]
+            inzcorn = grp["zcorn"][:, :, :, :]
+            inactnum = grp["actnum"][:, :, :]
+
+    result = _RESULT(
+        coordsv=incoord.astype("float64"),
+        zcornsv=inzcorn.astype("float32"),
+        actnumsv=inactnum.astype("float32"),
+    )
+
+    handle_metadata(result, meta, ncol, nrow, nlay)
+    return result
+
+
+def filter_subgrids_partial(
+    subgrids: dict[str, int],
+    k1: int | LMIN,
+    k2: int | LMAX,
+    nlay: int,
+    zerobased: bool,
+) -> dict[str, int]:
+    """
+    Filters and truncates the subgrids of the global grid so that they
+    refer to the filtered grid.
+
+    >>> filter_subgrids_partial(
+    ...   {"subgrid1": 4, "subgrid2": 1, "subgrid3":9},
+    ...   4,
+    ...   5,
+    ...   12,
+    ...   True
+    ... )
+    dict([('subgrid2', 1), ('subgrid3', 1)])
+
+    Args:
+        subgrids: The dict of subgrids.
+        k1: Start of subgrid layers (can be "min" to mean 0 or 1 dependent on zerobased)
+        k2: End of subgrid layers (cna be "max" to mean nlay or nlay -1
+            dependent on zerobased.
+        nlay: Original number of layers
+        zerobased: Whether indexing starts with 0 or 1.
+
+    Returns:
+        New Orderedict which is the subgrids dictionary with out of range
+        subgrids removed or truncated.
+
+    """
+    if k1 == "min":
+        k1 = 0 if zerobased else 1
+
+    if k2 == "max":
+        k2 = nlay - 1 if zerobased else nlay
+
+    # convert k1 and k2 to zero based
+    if not zerobased:
+        k1 -= 1
+        k2 -= 1
+
+    partial_subgrid: dict[str, int] = {}
+    start = 0
+    for key, value in subgrids.items():
+        end = value + start
+        partial_start = max(start, k1)
+        partial_end = min(end, k2 + 1)
+        if partial_end - partial_start > 0:
+            partial_subgrid[key] = partial_end - partial_start
+        start = end
+    return partial_subgrid
+
+
+def _partial_read(
+    h5h: dict[str, np.ndarray],
+    req: dict[str, int],
+    ijkrange: tuple[
+        int | LMIN,
+        int | LMAX,
+        int | LMIN,
+        int | LMAX,
+        int | LMIN,
+        int | LMAX,
+    ],
+    zerobased: bool,
+) -> tuple[np.ndarray, np.ndarray, np.ndarray, int, int, int]:
+    """Read a partial IJ range."""
+    ncol = req["ncol"]
+    nrow = req["nrow"]
+    nlay = req["nlay"]
+
+    if len(ijkrange) != 6:
+        raise ValueError("The ijkrange list must have 6 elements")
+
+    i1, i2, j1, j2, k1, k2 = ijkrange
+
+    if i1 == "min":
+        i1 = 0 if zerobased else 1
+    if j1 == "min":
+        j1 = 0 if zerobased else 1
+    if k1 == "min":
+        k1 = 0 if zerobased else 1
+
+    if i2 == "max":
+        i2 = ncol - 1 if zerobased else ncol
+    if j2 == "max":
+        j2 = nrow - 1 if zerobased else nrow
+    if k2 == "max":
+        k2 = nlay - 1 if zerobased else nlay
+
+    if not zerobased:
+        i1 -= 1
+        i2 -= 1
+        j1 -= 1
+        j2 -= 1
+        k1 -= 1
+        k2 -= 1
+
+    ncol2 = i2 - i1 + 1
+    nrow2 = j2 - j1 + 1
+    nlay2 = k2 - k1 + 1
+
+    if (
+        ncol2 < 1
+        or ncol2 > ncol
+        or nrow2 < 1
+        or nrow2 > nrow
+        or nlay2 < 1
+        or nlay2 > nlay
+    ):
+        raise ValueError("The ijkrange spesification exceeds boundaries.")
+
+    nncol2 = ncol2 + 1
+    nnrow2 = nrow2 + 1
+    nnlay2 = nlay2 + 1
+
+    dset = h5h["CornerPointGeometry/coord"]
+    cv = dset[i1 : i1 + nncol2, j1 : j1 + nnrow2, :]
+
+    dset = h5h["CornerPointGeometry/zcorn"]
+    zv = dset[i1 : i1 + nncol2, j1 : j1 + nnrow2, k1 : k1 + nnlay2, :]
+
+    dset = h5h["CornerPointGeometry/actnum"]
+    av = dset[i1 : i1 + ncol2, j1 : j1 + nrow2, k1 : k1 + nlay2]
+
+    return cv, zv, av, ncol2, nrow2, nlay2

xtgeo/grid3d/_grid_refine.py

@@ -0,0 +1,258 @@
+"""Private module for refinement of a grid."""
+
+from __future__ import annotations
+
+from typing import TYPE_CHECKING
+
+import numpy as np
+
+import xtgeo._internal as _internal  # type: ignore
+from xtgeo.common import XTGeoDialog, null_logger
+from xtgeo.grid3d import _gridprop_op1
+
+xtg = XTGeoDialog()
+logger = null_logger(__name__)
+
+if TYPE_CHECKING:
+    from xtgeo.grid3d import Grid, GridProperty
+
+
+def refine(
+    self: Grid,
+    refine_col: int | dict[int, int],
+    refine_row: int | dict[int, int],
+    refine_layer: int | dict[int, int],
+    zoneprop: GridProperty | None = None,
+) -> Grid:
+    """Refine in all direction, proportionally.
+
+    See details in caller.
+    """
+    self._set_xtgformat1()
+    self.make_zconsistent()
+
+    if isinstance(refine_col, int):
+        refine_factor_column = [refine_col] * self.dimensions[0]
+    else:
+        refine_factor_column = [1] * self.dimensions[0]
+        for col, factor in refine_col.items():
+            if 0 < col < self.dimensions[0] and isinstance(factor, int) and factor > 0:
+                refine_factor_column[col - 1] = factor
+            else:
+                xtg.warning("Invalid refine_col item {col}:{factor}")
+
+    if isinstance(refine_row, int):
+        refine_factor_row = [refine_row] * self.dimensions[1]
+    else:
+        refine_factor_row = [1] * self.dimensions[1]
+        for row, factor in refine_row.items():
+            if 0 < row < self.dimensions[1] and isinstance(factor, int) and factor > 0:
+                refine_factor_row[row - 1] = factor
+            else:
+                xtg.warning("Invalid refine_row item {row}:{factor}")
+
+    refine_factor_layer_dict = {}
+    # case 1 rfactor as scalar value.
+    if isinstance(refine_layer, int):
+        if self.subgrids:
+            subgrids = self.get_subgrids()
+            for i, _ in enumerate(self.subgrids.keys()):
+                refine_factor_layer_dict[i + 1] = refine_layer
+        else:
+            refine_factor_layer_dict[0] = refine_layer
+            subgrids = {}
+            subgrids[1] = self.nlay
+
+    # case 2 rfactor is a dict
+    else:
+        refine_factor_layer_dict = dict(
+            sorted(refine_layer.items())
+        )  # redefined to ordered
+        # 2a: zoneprop is present
+        if zoneprop is not None:
+            oldsubgrids = None
+            if self.subgrids:
+                oldsubgrids = self.get_subgrids()
+
+            subgrids = self.subgrids_from_zoneprop(zoneprop)
+
+            if oldsubgrids and subgrids.values() != oldsubgrids.values():
+                xtg.warn(
+                    "Subgrid definitions from zone property do not match existing "
+                    "subgrids. Proceeding with new subgrid definitions from zone "
+                    "property."
+                )
+
+        # 2b: zoneprop is not present
+        elif zoneprop is None and self.subgrids:
+            subgrids = self.get_subgrids()
+
+        elif zoneprop is None and not self.subgrids:
+            raise ValueError(
+                "You gave in a dict, but no zoneprops and "
+                "subgrids are not present in the grid"
+            )
+        else:
+            raise ValueError("Some major unexpected issue in routine...")
+
+    if len(subgrids) != len(refine_factor_layer_dict):
+        raise RuntimeError("Subgrids and refinements: different definition!")
+
+    self.set_subgrids(subgrids)
+
+    # Now, based on dict, give a value per subgrid for key, val in rfactor
+    newsubgrids = {}
+    newnlay = 0
+    for (_x, rfi), (snam, sran) in zip(
+        refine_factor_layer_dict.items(), subgrids.items()
+    ):
+        newsubgrids[snam] = sran * rfi
+        newnlay += newsubgrids[snam]
+
+    logger.debug("New layers: %s", newnlay)
+
+    refine_factor_layer = []
+
+    for (_, rfi), (_, arr) in zip(
+        refine_factor_layer_dict.items(), self.subgrids.items()
+    ):
+        for _ in range(len(arr)):
+            refine_factor_layer.append(rfi)
+
+    self._set_xtgformat2()
+
+    refine_factor_column = np.array(refine_factor_column, dtype=np.int8)
+    refine_factor_row = np.array(refine_factor_row, dtype=np.int8)
+    refine_factor_layer = np.array(refine_factor_layer, dtype=np.int8)
+
+    if refine_factor_column.sum() > self.dimensions[0]:
+        grid_cpp = _internal.grid3d.Grid(self)
+        ref_coordsv, ref_zcornsv, ref_actnumsv = grid_cpp.refine_columns(
+            refine_factor_column
+        )
+        self._coordsv = ref_coordsv
+        self._zcornsv = ref_zcornsv
+        self._actnumsv = ref_actnumsv.astype(np.int32)
+        self._ncol = int(refine_factor_column.sum())
+
+    if refine_factor_row.sum() > self.dimensions[1]:
+        grid_cpp = _internal.grid3d.Grid(self)
+        ref_coordsv, ref_zcornsv, ref_actnumsv = grid_cpp.refine_rows(refine_factor_row)
+        self._coordsv = ref_coordsv
+        self._zcornsv = ref_zcornsv
+        self._actnumsv = ref_actnumsv.astype(np.int32)
+        self._nrow = int(refine_factor_row.sum())
+
+    if refine_factor_layer.sum() > self.dimensions[2]:
+        grid_cpp = _internal.grid3d.Grid(self)
+        ref_zcornsv, ref_actnumsv = grid_cpp.refine_vertically(refine_factor_layer)
+        self._zcornsv = ref_zcornsv
+        self._actnumsv = ref_actnumsv.astype(np.int32)
+        self._nlay = newnlay
+        if self.subgrids is None or len(self.subgrids) <= 1:
+            self.subgrids = None
+        else:
+            self.set_subgrids(newsubgrids)
+
+    # Check if grid has any properties
+    if self._props and self._props.props and len(self._props.props) > 0:
+        for prop in self._props.props:
+            _gridprop_op1.refine(
+                prop, refine_factor_column, refine_factor_row, refine_factor_layer
+            )
+
+    return self
+
+
+def refine_vertically(
+    self: Grid,
+    rfactor: int | dict[int, int],
+    zoneprop: GridProperty | None = None,
+) -> Grid:
+    """Refine vertically, proportionally.
+
+    See details in caller.
+    """
+    self._set_xtgformat1()
+    self.make_zconsistent()
+
+    rfactord = {}
+
+    # case 1 rfactor as scalar value.
+    if isinstance(rfactor, int):
+        if self.subgrids:
+            subgrids = self.get_subgrids()
+            for i, _ in enumerate(self.subgrids.keys()):
+                rfactord[i + 1] = rfactor
+        else:
+            rfactord[0] = rfactor
+            subgrids = {}
+            subgrids[1] = self.nlay
+
+    # case 2 rfactor is a dict
+    else:
+        rfactord = dict(sorted(rfactor.items()))  # redefined to ordered
+        # 2a: zoneprop is present
+        if zoneprop is not None:
+            oldsubgrids = None
+            if self.subgrids:
+                oldsubgrids = self.get_subgrids()
+
+            subgrids = self.subgrids_from_zoneprop(zoneprop)
+
+            if oldsubgrids and subgrids.values() != oldsubgrids.values():
+                xtg.warn("ISSUES!!!")
+
+        # 2b: zoneprop is not present
+        elif zoneprop is None and self.subgrids:
+            subgrids = self.get_subgrids()
+
+        elif zoneprop is None and not self.subgrids:
+            raise ValueError(
+                "You gave in a dict, but no zoneprops and "
+                "subgrids are not present in the grid"
+            )
+        else:
+            raise ValueError("Some major unexpected issue in routine...")
+
+    if len(subgrids) != len(rfactord):
+        raise RuntimeError("Subgrids and refinements: different definition!")
+
+    self.set_subgrids(subgrids)
+
+    # Now, based on dict, give a value per subgrid for key, val in rfactor
+    newsubgrids = {}
+    newnlay = 0
+    for (_x, rfi), (snam, sran) in zip(rfactord.items(), subgrids.items()):
+        newsubgrids[snam] = sran * rfi
+        newnlay += newsubgrids[snam]
+
+    logger.debug("New layers: %s", newnlay)
+
+    refine_factors = []
+
+    for (_, rfi), (_, arr) in zip(rfactord.items(), self.subgrids.items()):
+        for _ in range(len(arr)):
+            refine_factors.append(rfi)
+
+    self._set_xtgformat2()
+    grid_cpp = _internal.grid3d.Grid(self)
+    refine_factors = np.array(refine_factors, dtype=np.int8)
+    ref_zcornsv, ref_actnumsv = grid_cpp.refine_vertically(refine_factors)
+
+    # update instance:
+    self._nlay = newnlay
+    self._zcornsv = ref_zcornsv
+    self._actnumsv = ref_actnumsv.astype(np.int32)
+
+    if self.subgrids is None or len(self.subgrids) <= 1:
+        self.subgrids = None
+    else:
+        self.set_subgrids(newsubgrids)
+
+    # Check if grid has any properties
+    if self._props and self._props.props and len(self._props.props) > 0:
+        for prop in self._props.props:
+            _gridprop_op1.refine(prop, 1, 1, refine_factors)
+
+    return self