Source code for geoh5py.shared.utils

#  Copyright (c) 2024 Mira Geoscience Ltd.
#
#  This file is part of geoh5py.
#
#  geoh5py is free software: you can redistribute it and/or modify
#  it under the terms of the GNU Lesser General Public License as published by
#  the Free Software Foundation, either version 3 of the License, or
#  (at your option) any later version.
#
#  geoh5py 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 Lesser General Public License for more details.
#
#  You should have received a copy of the GNU Lesser General Public License
#  along with geoh5py.  If not, see <https://www.gnu.org/licenses/>.

from __future__ import annotations

import warnings
from abc import ABC
from contextlib import contextmanager
from io import BytesIO
from pathlib import Path
from typing import TYPE_CHECKING, Any, Callable
from uuid import UUID

import h5py
import numpy as np

if TYPE_CHECKING:
    from ..workspace import Workspace
    from .entity import Entity


KEY_MAP = {
    "cells": "Cells",
    "color_map": "Color map",
    "concatenated_attributes": "Attributes",
    "concatenated_object_ids": "Concatenated object IDs",
    "layers": "Layers",
    "metadata": "Metadata",
    "octree_cells": "Octree Cells",
    "options": "options",
    "prisms": "Prisms",
    "property_groups": "PropertyGroups",
    "property_group_ids": "Property Group IDs",
    "surveys": "Surveys",
    "trace": "Trace",
    "trace_depth": "TraceDepth",
    "u_cell_delimiters": "U cell delimiters",
    "v_cell_delimiters": "V cell delimiters",
    "values": "Data",
    "vertices": "Vertices",
    "z_cell_delimiters": "Z cell delimiters",
    "INVALID": "Invalid",
    "INTEGER": "Integer",
    "FLOAT": "Float",
    "TEXT": "Text",
    "BOOLEAN": "Boolean",
    "REFERENCED": "Referenced",
    "FILENAME": "Filename",
    "BLOB": "Blob",
    "VECTOR": "Vector",
    "DATETIME": "DateTime",
    "GEOMETRIC": "Geometric",
    "MULTI_TEXT": "Multi-Text",
    "UNKNOWN": "Unknown",
    "OBJECT": "Object",
    "CELL": "Cell",
    "VERTEX": "Vertex",
    "FACE": "Face",
    "GROUP": "Group",
}
INV_KEY_MAP = {value: key for key, value in KEY_MAP.items()}

PNG_KWARGS = {"format": "PNG", "compress_level": 9}
JPG_KWARGS = {"format": "JPEG", "quality": 85}
TIF_KWARGS = {"format": "TIFF"}

PILLOW_ARGUMENTS = {
    "1": PNG_KWARGS,
    "L": PNG_KWARGS,
    "P": PNG_KWARGS,
    "RGB": PNG_KWARGS,
    "RGBA": PNG_KWARGS,
    "CMYK": JPG_KWARGS,
    "YCbCr": JPG_KWARGS,
    "I": TIF_KWARGS,
    "F": TIF_KWARGS,
}


[docs] @contextmanager def fetch_active_workspace(workspace: Workspace | None, mode: str = "r"): """ Open a workspace in the requested 'mode'. If receiving an opened Workspace instead, merely return the given workspace. :param workspace: A Workspace class :param mode: Set the h5 read/write mode :return h5py.File: Handle to an opened Workspace. """ if ( workspace is None or getattr(workspace, "_geoh5") and mode in workspace.geoh5.mode ): try: yield workspace finally: pass else: if getattr(workspace, "_geoh5"): warnings.warn( f"Closing the workspace in mode '{workspace.geoh5.mode}' " f"and re-opening in mode '{mode}'." ) workspace.close() try: yield workspace.open(mode=mode) finally: workspace.close()
[docs] @contextmanager def fetch_h5_handle(file: str | h5py.File | Path, mode: str = "r") -> h5py.File: """ Open in read+ mode a geoh5 file from string. If receiving a file instead of a string, merely return the given file. :param file: Name or handle to a geoh5 file. :param mode: Set the h5 read/write mode :return h5py.File: Handle to an opened h5py file. """ if isinstance(file, h5py.File): try: yield file finally: pass else: if Path(file).suffix != ".geoh5": raise ValueError("Input h5 file must have a 'geoh5' extension.") h5file = h5py.File(file, mode) try: yield h5file finally: h5file.close()
[docs] def match_values(vec_a, vec_b, collocation_distance=1e-4) -> np.ndarray: """ Find indices of matching values between two arrays, within collocation_distance. :param: vec_a, list or numpy.ndarray Input sorted values :param: vec_b, list or numpy.ndarray Query values :return: indices, numpy.ndarray Pairs of indices for matching values between the two arrays such that vec_a[ind[:, 0]] == vec_b[ind[:, 1]]. """ ind_sort = np.argsort(vec_a) ind = np.minimum( np.searchsorted(vec_a[ind_sort], vec_b, side="right"), vec_a.shape[0] - 1 ) nearests = np.c_[ind, ind - 1] match = np.where( np.abs(vec_a[ind_sort][nearests] - vec_b[:, None]) < collocation_distance ) indices = np.c_[ind_sort[nearests[match[0], match[1]]], match[0]] return indices
[docs] def merge_arrays( head, tail, replace="A->B", mapping=None, collocation_distance=1e-4, return_mapping=False, ) -> np.ndarray: """ Given two numpy.arrays of different length, find the matching values and append both arrays. :param: head, numpy.array of float First vector of shape(M,) to be appended. :param: tail, numpy.array of float Second vector of shape(N,) to be appended :param: mapping=None, numpy.ndarray of int Optional array where values from the head are replaced by the tail. :param: collocation_distance=1e-4, float Tolerance between matching values. :return: numpy.array shape(O,) Unique values from head to tail without repeats, within collocation_distance. """ if mapping is None: mapping = match_values(head, tail, collocation_distance=collocation_distance) if mapping.shape[0] > 0: if replace == "B->A": head[mapping[:, 0]] = tail[mapping[:, 1]] else: tail[mapping[:, 1]] = head[mapping[:, 0]] tail = np.delete(tail, mapping[:, 1]) if return_mapping: return np.r_[head, tail], mapping return np.r_[head, tail]
[docs] def clear_array_attributes(entity: Entity, recursive: bool = False): """ Clear all stashed values of attributes from an entity to free up memory. :param entity: Entity to clear attributes from. :param recursive: Clear attributes from children entities. """ if isinstance(entity.workspace.h5file, BytesIO): return for attribute in ["vertices", "cells", "values", "prisms", "layers"]: if hasattr(entity, attribute): setattr(entity, f"_{attribute}", None) if recursive: for child in entity.children: clear_array_attributes(child, recursive=recursive)
[docs] def compare_entities( # pylint: disable=too-many-branches object_a, object_b, ignore: list | None = None, decimal: int = 6 ) -> None: ignore_list = ["_workspace", "_children", "_visual_parameters"] if ignore is not None: for item in ignore: ignore_list.append(item) if isinstance(object_a, bytes): assert object_a == object_b, "Bytes values do not match." return for attr in object_a.__dict__.keys(): if attr in ignore_list: continue if isinstance(getattr(object_a, attr[1:]), ABC): compare_entities( getattr(object_a, attr[1:]), getattr(object_b, attr[1:]), ignore=ignore ) else: if isinstance(getattr(object_a, attr[1:]), np.ndarray): if getattr(object_b, attr[1:]) is None: raise ValueError( f"attr {attr[1:]} is None for object {object_b.name}" ) attr_a = getattr(object_a, attr[1:]).tolist() if len(attr_a) > 0 and isinstance(attr_a[0], str): assert all( a == b for a, b in zip( getattr(object_a, attr[1:]), getattr(object_b, attr[1:]) ) ), f"Error comparing attribute '{attr}'." else: np.testing.assert_array_almost_equal( attr_a, getattr(object_b, attr[1:]).tolist(), decimal=decimal, err_msg=f"Error comparing attribute '{attr}'.", ) elif isinstance(getattr(object_a, attr[1:]), float): np.testing.assert_almost_equal( getattr(object_a, attr[1:]), getattr(object_b, attr[1:]), decimal=decimal, err_msg=f"Error comparing attribute '{attr}'.", ) else: assert np.all( getattr(object_a, attr[1:]) == getattr(object_b, attr[1:]) ), f"Output attribute '{attr[1:]}' for {object_a} do not match input {object_b}"
[docs] def iterable(value: Any, checklen: bool = False) -> bool: """ Checks if object is iterable. Parameters ---------- value : Object to check for iterableness. checklen : Restrict objects with __iter__ method to len > 1. Returns ------- True if object has __iter__ attribute but is not string or dict type. """ only_array_like = (not isinstance(value, str)) & (not isinstance(value, dict)) if (hasattr(value, "__iter__")) & only_array_like: return not (checklen and (len(value) == 1)) return False
[docs] def iterable_message(valid: list[Any] | None) -> str: """Append possibly iterable valid: "Must be (one of): {valid}.".""" if valid is None: msg = "" elif iterable(valid, checklen=True): vstr = "'" + "', '".join(str(k) for k in valid) + "'" msg = f" Must be one of: {vstr}." else: msg = f" Must be: '{valid[0]}'." return msg
[docs] def is_uuid(value: str) -> bool: """Check if a string is UUID compliant.""" try: UUID(str(value)) return True except ValueError: return False
[docs] def entity2uuid(value: Any) -> UUID | Any: """Convert an entity to its UUID.""" if hasattr(value, "uid"): return value.uid return value
[docs] def uuid2entity(value: UUID, workspace: Workspace) -> Entity | Any: """Convert UUID to a known entity.""" if isinstance(value, UUID): if value in workspace.list_entities_name: return workspace.get_entity(value)[0] # Search for property groups for obj in workspace.objects: if getattr(obj, "property_groups", None) is not None: prop_group = [ prop_group for prop_group in getattr(obj, "property_groups") if prop_group.uid == value ] if prop_group: return prop_group[0] return None return value
[docs] def str2uuid(value: Any) -> UUID | Any: """Convert string to UUID""" if is_uuid(value): # TODO insert validation return UUID(str(value)) return value
[docs] def as_str_if_uuid(value: UUID | Any) -> str | Any: """Convert :obj:`UUID` to string used in geoh5.""" if isinstance(value, UUID): return "{" + str(value) + "}" return value
[docs] def bool_value(value: np.int8) -> bool: """Convert logical int8 to bool.""" return bool(value)
[docs] def as_str_if_utf8_bytes(value) -> str: """Convert bytes to string""" if isinstance(value, bytes): value = value.decode("utf-8") return value
[docs] def dict_mapper(val, string_funcs: list[Callable], *args, omit: dict | None = None): """ Recursion through nested dictionaries and applies mapping functions to values. :param val: Value (could be another dictionary) to apply transform functions. :param string_funcs: Functions to apply on values within the input dictionary. :param omit: Dictionary of functions to omit. :return val: Transformed values """ if isinstance(val, dict): for key, values in val.items(): short_list = string_funcs.copy() if omit is not None: short_list = [ fun for fun in string_funcs if fun not in omit.get(key, []) ] val[key] = dict_mapper(values, short_list) if isinstance(val, list): out = [] for elem in val: for fun in string_funcs: elem = fun(elem, *args) out += [elem] return out for fun in string_funcs: val = fun(val, *args) return val
[docs] def box_intersect(extent_a: np.ndarray, extent_b: np.ndarray) -> bool: """ Compute the intersection of two axis-aligned bounding extents defined by their arrays of minimum and maximum bounds in N-D space. :param extent_a: First extent or shape (2, N) :param extent_b: Second extent or shape (2, N) :return: Logic if the box extents intersect along all dimensions. """ for extent in [extent_a, extent_b]: if not isinstance(extent, np.ndarray) or extent.ndim != 2: raise TypeError("Input extents must be 2D numpy.ndarrays.") if extent.shape[0] != 2 or not np.all(extent[0, :] <= extent[1, :]): raise ValueError( "Extents must be of shape (2, N) containing the minimum and maximum " "bounds in nd-space on the first and second row respectively." ) for comp_a, comp_b in zip(extent_a.T, extent_b.T): min_ext = max(comp_a[0], comp_b[0]) max_ext = min(comp_a[1], comp_b[1]) if min_ext > max_ext: return False return True
[docs] def mask_by_extent( locations: np.ndarray, extent: np.ndarray, inverse: bool = False ) -> np.ndarray: """ Find indices of locations within a rectangular extent. :param locations: shape(*, 3) or shape(*, 2) Coordinates to be evaluated. :param extent: shape(2, 2) Limits defined by the South-West and North-East corners. Extents can also be provided as 3D coordinates with shape(2, 3) defining the top and bottom limits. :param inverse: Return the complement of the mask extent. :returns: Array of bool for the locations inside or outside the box extent. """ if not isinstance(extent, np.ndarray) or extent.ndim != 2: raise ValueError("Input 'extent' must be a 2D array-like.") if not isinstance(locations, np.ndarray) or locations.ndim != 2: raise ValueError( "Input 'locations' must be an array-like of shape(*, 3) or (*, 2)." ) indices = np.ones(locations.shape[0], dtype=bool) for loc, lim in zip(locations.T, extent.T): indices &= (lim[0] <= loc) & (loc <= lim[1]) if inverse: return ~indices return indices
[docs] def get_attributes(entity, omit_list=(), attributes=None): """Extract the attributes of an object with omissions.""" if attributes is None: attributes = {} for key in vars(entity): if key not in omit_list: if key[0] == "_": key = key[1:] attr = getattr(entity, key) attributes[key] = attr return attributes
[docs] def xy_rotation_matrix(angle: float) -> np.ndarray: """ Rotation matrix about the z-axis. :param angle: Rotation angle in radians. :return rot: Rotation matrix. """ return np.array( [ [np.cos(angle), -np.sin(angle), 0.0], [np.sin(angle), np.cos(angle), 0.0], [0.0, 0.0, 1.0], ] )
[docs] def yz_rotation_matrix(angle: float) -> np.ndarray: """ Rotation matrix about the x-axis. :param angle: Rotation angle in radians. :return: rot: Rotation matrix. """ return np.array( [ [1, 0, 0], [0, np.cos(angle), -np.sin(angle)], [0, np.sin(angle), np.cos(angle)], ] )
[docs] def dip_points(points: np.ndarray, dip: float, rotation: float = 0) -> np.ndarray: """ Rotate points about the x-axis by the dip angle and then about the z-axis by the rotation angle. :param points: an array of points to rotate :param dip: the dip angle in radians :param rotation: the rotation angle in radians :return: the rotated points """ # Assert points is a numpy array containing 3D points if not isinstance(points, np.ndarray) and points.ndim != 2 and points.shape[1] != 3: raise TypeError("Input points must be a 2D numpy array of shape (N, 3).") # rotate the points about the z-axis by the inverse rotation angle points = xy_rotation_matrix(-rotation) @ points.T # Rotate points with the dip angle points = yz_rotation_matrix(dip) @ points # Rotate back the points to initial orientation points = xy_rotation_matrix(rotation) @ points return points.T