hpcflow 0.1.15__py3-none-any.whl → 0.2.0a271__py3-none-any.whl

hpcflow/sdk/core/utils.py

@@ -0,0 +1,1207 @@
+"""
+Miscellaneous utilities.
+"""
+
+from __future__ import annotations
+from collections import Counter
+from asyncio import events
+import contextvars
+import contextlib
+import copy
+import enum
+import functools
+import hashlib
+from itertools import accumulate, islice
+from importlib import resources
+import json
+import keyword
+import os
+from pathlib import Path, PurePath
+import random
+import re
+import socket
+import string
+import subprocess
+from datetime import datetime, timedelta, timezone
+import sys
+import traceback
+from typing import Literal, cast, overload, TypeVar, TYPE_CHECKING
+import fsspec  # type: ignore
+import numpy as np
+
+from ruamel.yaml import YAML
+from ruamel.yaml.error import MarkedYAMLError
+from watchdog.utils.dirsnapshot import DirectorySnapshot
+
+from hpcflow.sdk.core.errors import (
+    ContainerKeyError,
+    InvalidIdentifier,
+    MissingVariableSubstitutionError,
+    YAMLError,
+)
+from hpcflow.sdk.log import TimeIt
+from hpcflow.sdk.utils.deferred_file import DeferredFileWriter
+
+if TYPE_CHECKING:
+    from collections.abc import Callable, Hashable, Iterable, Mapping, Sequence
+    from contextlib import AbstractContextManager
+    from types import ModuleType
+    from typing import Any, IO, Iterator
+    from typing_extensions import TypeAlias
+    from numpy.typing import NDArray
+    from ..typing import PathLike
+
+T = TypeVar("T")
+T2 = TypeVar("T2")
+T3 = TypeVar("T3")
+TList: TypeAlias = "T | list[TList]"
+TD = TypeVar("TD", bound="Mapping[str, Any]")
+E = TypeVar("E", bound=enum.Enum)
+
+
+def make_workflow_id() -> str:
+    """
+    Generate a random ID for a workflow.
+    """
+    length = 12
+    chars = string.ascii_letters + "0123456789"
+    return "".join(random.choices(chars, k=length))
+
+
+def get_time_stamp() -> str:
+    """
+    Get the current time in standard string form.
+    """
+    return datetime.now(timezone.utc).astimezone().strftime("%Y.%m.%d_%H:%M:%S_%z")
+
+
+def get_duplicate_items(lst: Iterable[T]) -> list[T]:
+    """Get a list of all items in an iterable that appear more than once, assuming items
+    are hashable.
+
+    Examples
+    --------
+    >>> get_duplicate_items([1, 1, 2, 3])
+    [1]
+
+    >>> get_duplicate_items([1, 2, 3])
+    []
+
+    >>> get_duplicate_items([1, 2, 3, 3, 3, 2])
+    [2, 3]
+
+    """
+    return [x for x, y in Counter(lst).items() if y > 1]
+
+
+def check_valid_py_identifier(name: str) -> str:
+    """Check a string is (roughly) a valid Python variable identifier and return it.
+
+    The rules are:
+        1. `name` must not be empty
+        2. `name` must not be a Python keyword
+        3. `name` must begin with an alphabetic character, and all remaining characters
+           must be alphanumeric.
+
+    Notes
+    -----
+    The following attributes are passed through this function on object initialisation:
+        - `ElementGroup.name`
+        - `Executable.label`
+        - `Parameter.typ`
+        - `TaskObjective.name`
+        - `TaskSchema.method`
+        - `TaskSchema.implementation`
+        - `Loop.name`
+
+    """
+    try:
+        trial_name = name[1:].replace("_", "")  # "internal" underscores are allowed
+    except TypeError:
+        raise InvalidIdentifier(name) from None
+    except KeyError as e:
+        raise KeyError(f"unexpected name type {name}") from e
+    if (
+        not name
+        or not (name[0].isalpha() and ((trial_name[1:] or "a").isalnum()))
+        or keyword.iskeyword(name)
+    ):
+        raise InvalidIdentifier(name)
+
+    return name
+
+
+@overload
+def group_by_dict_key_values(  # type: ignore[overload-overlap]
+    lst: list[dict[T, T2]], key: T
+) -> list[list[dict[T, T2]]]: ...
+
+
+@overload
+def group_by_dict_key_values(lst: list[TD], key: str) -> list[list[TD]]: ...
+
+
+def group_by_dict_key_values(lst: list, key):
+    """Group a list of dicts according to specified equivalent key-values.
+
+    Parameters
+    ----------
+    lst : list of dict
+        The list of dicts to group together.
+    key : key value
+        Dicts that have identical values for all of these keys will be grouped together
+        into a sub-list.
+
+    Returns
+    -------
+    grouped : list of list of dict
+
+    Examples
+    --------
+    >>> group_by_dict_key_values([{'a': 1}, {'a': 2}, {'a': 1}], 'a')
+    [[{'a': 1}, {'a': 1}], [{'a': 2}]]
+
+    """
+
+    grouped = [[lst[0]]]
+    for lst_item in lst[1:]:
+        for group_idx, group in enumerate(grouped):
+            try:
+                is_vals_equal = lst_item[key] == group[0][key]
+
+            except KeyError:
+                # dicts that do not have the `key` will be in their own group:
+                is_vals_equal = False
+
+            if is_vals_equal:
+                grouped[group_idx].append(lst_item)
+                break
+
+        if not is_vals_equal:
+            grouped.append([lst_item])
+
+    return grouped
+
+
+def swap_nested_dict_keys(dct: dict[T, dict[T2, T3]], inner_key: T2):
+    """Return a copy where top-level keys have been swapped with a second-level inner key.
+
+    Examples:
+    ---------
+    >>> swap_nested_dict_keys(
+        dct={
+            'p1': {'format': 'direct', 'all_iterations': True},
+            'p2': {'format': 'json'},
+            'p3': {'format': 'direct'},
+        },
+        inner_key="format",
+    )
+    {
+        "direct": {"p1": {"all_iterations": True}, "p3": {}},
+        "json": {"p2": {}},
+    }
+
+    """
+    out: dict[T3, dict[T, dict[T2, T3]]] = {}
+    for k, v in copy.deepcopy(dct or {}).items():
+        out.setdefault(v.pop(inner_key), {})[k] = v
+    return out
+
+
+def _ensure_int(path_comp: Any, cur_data: Any, cast_indices: bool) -> int:
+    """
+    Helper for get_in_container() and set_in_container()
+    """
+    if isinstance(path_comp, int):
+        return path_comp
+    if not cast_indices:
+        raise TypeError(
+            f"Path component {path_comp!r} must be an integer index "
+            f"since data is a sequence: {cur_data!r}."
+        )
+    try:
+        return int(path_comp)
+    except (TypeError, ValueError) as e:
+        raise TypeError(
+            f"Path component {path_comp!r} must be an integer index "
+            f"since data is a sequence: {cur_data!r}."
+        ) from e
+
+
+def get_in_container(
+    cont, path: Sequence, cast_indices: bool = False, allow_getattr: bool = False
+):
+    """
+    Follow a path (sequence of indices of appropriate type) into a container to obtain
+    a "leaf" value. Containers can be lists, tuples, dicts,
+    or any class (with `getattr()`) if ``allow_getattr`` is True.
+    """
+    cur_data = cont
+    err_msg = (
+        "Data at path {path_comps!r} is not a sequence, but is of type "
+        "{cur_data_type!r} and so sub-data cannot be extracted."
+    )
+    for idx, path_comp in enumerate(path):
+        if isinstance(cur_data, (list, tuple)):
+            cur_data = cur_data[_ensure_int(path_comp, cur_data, cast_indices)]
+        elif isinstance(cur_data, dict) or hasattr(cur_data, "__getitem__"):
+            try:
+                cur_data = cur_data[path_comp]
+            except KeyError:
+                raise ContainerKeyError(path=cast("list[str]", path[: idx + 1]))
+        elif allow_getattr:
+            try:
+                cur_data = getattr(cur_data, path_comp)
+            except AttributeError:
+                raise ValueError(
+                    err_msg.format(cur_data_type=type(cur_data), path_comps=path[:idx])
+                )
+        else:
+            raise ValueError(
+                err_msg.format(cur_data_type=type(cur_data), path_comps=path[:idx])
+            )
+    return cur_data
+
+
+def set_in_container(
+    cont, path: Sequence, value, ensure_path=False, cast_indices=False
+) -> None:
+    """
+    Follow a path (sequence of indices of appropriate type) into a container to update
+    a "leaf" value. Containers can be lists, tuples or dicts.
+    The "branch" holding the leaf to update must be modifiable.
+    """
+    if ensure_path:
+        num_path = len(path)
+        for idx in range(1, num_path):
+            try:
+                get_in_container(cont, path[:idx], cast_indices=cast_indices)
+            except (KeyError, ValueError):
+                set_in_container(
+                    cont=cont,
+                    path=path[:idx],
+                    value={},
+                    ensure_path=False,
+                    cast_indices=cast_indices,
+                )
+
+    sub_data = get_in_container(cont, path[:-1], cast_indices=cast_indices)
+    path_comp = path[-1]
+    if isinstance(sub_data, (list, tuple)):
+        path_comp = _ensure_int(path_comp, sub_data, cast_indices)
+    sub_data[path_comp] = value
+
+
+def get_relative_path(path1: Sequence[T], path2: Sequence[T]) -> Sequence[T]:
+    """Get relative path components between two paths.
+
+    Parameters
+    ----------
+    path1 : tuple of (str or int or float) of length N
+    path2 : tuple of (str or int or float) of length less than or equal to N
+
+    Returns
+    -------
+    relative_path : tuple of (str or int or float)
+        The path components in `path1` that are not in `path2`.
+
+    Raises
+    ------
+    ValueError
+        If the two paths do not share a common ancestor of path components, or if `path2`
+        is longer than `path1`.
+
+    Notes
+    -----
+    This function behaves like a simplified `PurePath(*path1).relative_to(PurePath(*path2))`
+    from the `pathlib` module, but where path components can include non-strings.
+
+    Examples
+    --------
+    >>> get_relative_path(('A', 'B', 'C'), ('A',))
+    ('B', 'C')
+
+    >>> get_relative_path(('A', 'B'), ('A', 'B'))
+    ()
+
+    """
+
+    len_path2 = len(path2)
+    if len(path1) < len_path2 or any(i != j for i, j in zip(path1[:len_path2], path2)):
+        raise ValueError(f"{path1!r} is not in the subpath of {path2!r}.")
+
+    return path1[len_path2:]
+
+
+def search_dir_files_by_regex(
+    pattern: str | re.Pattern[str], directory: str | os.PathLike = "."
+) -> list[str]:
+    """Search recursively for files in a directory by a regex pattern and return matching
+    file paths, relative to the given directory."""
+    dir_ = Path(directory)
+    return [
+        str(entry.relative_to(dir_))
+        for entry in dir_.rglob("*")
+        if re.search(pattern, entry.name)
+    ]
+
+
+class PrettyPrinter:
+    """
+    A class that produces a nice readable version of itself with ``str()``.
+    Intended to be subclassed.
+    """
+
+    def __str__(self) -> str:
+        lines = [self.__class__.__name__ + ":"]
+        for key, val in vars(self).items():
+            lines.extend(f"{key}: {val}".split("\n"))
+        return "\n    ".join(lines)
+
+
+_STRING_VARS_RE = re.compile(r"\<\<var:(.*?)(?:\[(.*)\])?\>\>")
+
+
+@TimeIt.decorator
+def substitute_string_vars(string: str, variables: dict[str, str]):
+    """
+    Scan ``string`` and substitute sequences like ``<<var:ABC>>`` with the value
+    looked up in the supplied dictionary (with ``ABC`` as the key).
+
+    Default values for the substitution can be supplied like:
+    ``<<var:ABC[default=XYZ]>>``
+
+    Examples
+    --------
+    >>> substitute_string_vars("abc <<var:def>> ghi", {"def": "123"})
+    "abc 123 def"
+    """
+
+    def var_repl(match_obj: re.Match[str]) -> str:
+        kwargs: dict[str, str] = {}
+        var_name: str = match_obj[1]
+        kwargs_str: str | None = match_obj[2]
+        if kwargs_str:
+            for i in kwargs_str.split(","):
+                k, v = i.split("=")
+                kwargs[k.strip()] = v.strip()
+        try:
+            out = str(variables[var_name])
+        except KeyError:
+            if "default" in kwargs:
+                out = kwargs["default"]
+                print(
+                    f"Using default value ({out!r}) for workflow template string "
+                    f"variable {var_name!r}."
+                )
+            else:
+                raise MissingVariableSubstitutionError(var_name, variables)
+        return out
+
+    return _STRING_VARS_RE.sub(
+        repl=var_repl,
+        string=string,
+    )
+
+
+@TimeIt.decorator
+def read_YAML_str(
+    yaml_str: str,
+    typ="safe",
+    variables: dict[str, str] | Literal[False] | None = None,
+    source: str | None = None,
+) -> Any:
+    """Load a YAML string. This will produce basic objects.
+
+    Parameters
+    ----------
+    yaml_str:
+        The YAML string to parse.
+    typ:
+        Load type passed to the YAML library.
+    variables:
+        String variables to substitute in `yaml_str`. Substitutions will be attempted if
+        the file looks to contain variable references (like "<<var:name>>"). If set to
+        `False`, no substitutions will occur.
+    source:
+        Used to document the source of the YAML string if raising a parsing error.
+        Typically, this should be a string that starts with "from ...", e.g.
+        "from the file path '/path/to/bad/file'".
+    """
+    if variables is not False and "<<var:" in yaml_str:
+        yaml_str = substitute_string_vars(yaml_str, variables=variables or {})
+    yaml = YAML(typ=typ)
+    try:
+        return yaml.load(yaml_str)
+    except MarkedYAMLError as err:  # includes `ScannerError` and `ParserError`
+        source_str = f"{source} " if source else ""
+        raise YAMLError(
+            f"The YAML string {source_str}is not formatted correctly."
+        ) from err
+
+
+@TimeIt.decorator
+def read_YAML_file(
+    path: PathLike, typ="safe", variables: dict[str, str] | Literal[False] | None = None
+) -> Any:
+    """Load a YAML file. This will produce basic objects.
+
+    Parameters
+    ----------
+    path:
+        Path to the YAML file to parse.
+    typ:
+        Load type passed to the YAML library.
+    variables:
+        String variables to substitute in the file given by `path`. Substitutions will be
+        attempted if the file looks to contain variable references (like "<<var:name>>").
+        If set to `False`, no substitutions will occur.
+    """
+    with fsspec.open(path, "rt") as f:
+        yaml_str: str = f.read()
+    return read_YAML_str(yaml_str, typ=typ, variables=variables, source=f"from {path!r}")
+
+
+def write_YAML_file(obj, path: str | Path, typ: str = "safe") -> None:
+    """Write a basic object to a YAML file."""
+    yaml = YAML(typ=typ)
+    with Path(path).open("wt") as fp:
+        yaml.dump(obj, fp)
+
+
+def read_JSON_string(
+    json_str: str, variables: dict[str, str] | Literal[False] | None = None
+) -> Any:
+    """Load a JSON string. This will produce basic objects.
+
+    Parameters
+    ----------
+    json_str:
+        The JSON string to parse.
+    variables:
+        String variables to substitute in `json_str`. Substitutions will be attempted if
+        the file looks to contain variable references (like "<<var:name>>"). If set to
+        `False`, no substitutions will occur.
+    """
+    if variables is not False and "<<var:" in json_str:
+        json_str = substitute_string_vars(json_str, variables=variables or {})
+    return json.loads(json_str)
+
+
+def read_JSON_file(path, variables: dict[str, str] | Literal[False] | None = None) -> Any:
+    """Load a JSON file. This will produce basic objects.
+
+    Parameters
+    ----------
+    path:
+        Path to the JSON file to parse.
+    variables:
+        String variables to substitute in the file given by `path`. Substitutions will be
+        attempted if the file looks to contain variable references (like "<<var:name>>").
+        If set to `False`, no substitutions will occur.
+    """
+    with fsspec.open(path, "rt") as f:
+        json_str: str = f.read()
+    return read_JSON_string(json_str, variables=variables)
+
+
+def write_JSON_file(obj, path: str | Path) -> None:
+    """Write a basic object to a JSON file."""
+    with Path(path).open("wt") as fp:
+        json.dump(obj, fp)
+
+
+def get_item_repeat_index(
+    lst: Sequence[T],
+    *,
+    distinguish_singular: bool = False,
+    item_callable: Callable[[T], Hashable] | None = None,
+):
+    """Get the repeat index for each item in a list.
+
+    Parameters
+    ----------
+    lst : list
+        Must contain hashable items, or hashable objects that are returned via `callable`
+        called on each item.
+    distinguish_singular : bool
+        If True, items that are not repeated will have a repeat index of 0, and items that
+        are repeated will have repeat indices starting from 1.
+    item_callable : callable
+        If specified, comparisons are made on the output of this callable on each item.
+
+    Returns
+    -------
+    repeat_idx : list of int
+        Repeat indices of each item (see `distinguish_singular` for details).
+
+    """
+
+    idx: dict[Any, list[int]] = {}
+    if item_callable:
+        for i_idx, item in enumerate(lst):
+            idx.setdefault(item_callable(item), []).append(i_idx)
+    else:
+        for i_idx, item in enumerate(lst):
+            idx.setdefault(item, []).append(i_idx)
+
+    rep_idx = [0] * len(lst)
+    for v in idx.values():
+        start = len(v) > 1 if distinguish_singular else 0
+        for i_idx, i in enumerate(v, start):
+            rep_idx[i] = i_idx
+
+    return rep_idx
+
+
+def get_process_stamp() -> str:
+    """
+    Return a globally unique string identifying this process.
+
+    Note
+    ----
+    This should only be called once per process.
+    """
+    return "{} {} {}".format(
+        datetime.now(),
+        socket.gethostname(),
+        os.getpid(),
+    )
+
+
+_ANSI_ESCAPE_RE = re.compile(r"(\x9B|\x1B\[)[0-?]*[ -\/]*[@-~]")
+
+
+def remove_ansi_escape_sequences(string: str) -> str:
+    """
+    Strip ANSI terminal escape codes from a string.
+    """
+    return _ANSI_ESCAPE_RE.sub("", string)
+
+
+def get_md5_hash(obj) -> str:
+    """
+    Compute the MD5 hash of an object.
+    This is the hash of the JSON of the object (with sorted keys) as a hex string.
+    """
+    json_str = json.dumps(obj, sort_keys=True)
+    return hashlib.md5(json_str.encode("utf-8")).hexdigest()
+
+
+def get_nested_indices(
+    idx: int, size: int, nest_levels: int, raise_on_rollover: bool = False
+) -> list[int]:
+    """Generate the set of nested indices of length `n` that correspond to a global
+    `idx`.
+
+    Examples
+    --------
+    >>> for i in range(4**2): print(get_nest_index(i, nest_levels=2, size=4))
+    [0, 0]
+    [0, 1]
+    [0, 2]
+    [0, 3]
+    [1, 0]
+    [1, 1]
+    [1, 2]
+    [1, 3]
+    [2, 0]
+    [2, 1]
+    [2, 2]
+    [2, 3]
+    [3, 0]
+    [3, 1]
+    [3, 2]
+    [3, 3]
+
+    >>> for i in range(4**3): print(get_nested_indices(i, nest_levels=3, size=4))
+    [0, 0, 0]
+    [0, 0, 1]
+    [0, 0, 2]
+    [0, 0, 3]
+    [0, 1, 0]
+       ...
+    [3, 2, 3]
+    [3, 3, 0]
+    [3, 3, 1]
+    [3, 3, 2]
+    [3, 3, 3]
+    """
+    if raise_on_rollover and idx >= size**nest_levels:
+        raise ValueError(
+            f"`idx` ({idx}) is greater than or equal to  size**nest_levels` "
+            f"({size**nest_levels})."
+        )
+
+    return [(idx // (size ** (nest_levels - (i + 1)))) % size for i in range(nest_levels)]
+
+
+def ensure_in(item: T, lst: list[T]) -> int:
+    """Get the index of an item in a list and append the item if it is not in the
+    list."""
+    # TODO: add tests
+    try:
+        return lst.index(item)
+    except ValueError:
+        lst.append(item)
+        return len(lst) - 1
+
+
+def list_to_dict(
+    lst: Sequence[Mapping[T, T2]], exclude: Iterable[T] | None = None
+) -> dict[T, list[T2]]:
+    """
+    Convert a list of dicts to a dict of lists.
+    """
+    # TODO: test
+    exc = frozenset(exclude or ())
+    dct: dict[T, list[T2]] = {k: [] for k in lst[0] if k not in exc}
+    for d in lst:
+        for k, v in d.items():
+            if k not in exc:
+                dct[k].append(v)
+    return dct
+
+
+def bisect_slice(selection: slice, len_A: int) -> tuple[slice, slice]:
+    """Given two sequences (the first of which of known length), get the two slices that
+    are equivalent to a given slice if the two sequences were combined."""
+
+    if selection.start < 0 or selection.stop < 0 or selection.step < 0:
+        raise NotImplementedError("Can't do negative slices yet.")
+
+    A_idx = selection.indices(len_A)
+    B_start = selection.start - len_A
+    if len_A != 0 and B_start < 0:
+        B_start = B_start % selection.step
+    if len_A > selection.stop:
+        B_stop = B_start
+    else:
+        B_stop = selection.stop - len_A
+
+    return slice(*A_idx), slice(B_start, B_stop, selection.step)
+
+
+def replace_items(lst: list[T], start: int, end: int, repl: list[T]) -> list[T]:
+    """Replaced a range of items in a list with items in another list."""
+    # Convert to actual indices for our safety checks; handles end-relative addressing
+    real_start, real_end, _ = slice(start, end).indices(len(lst))
+    if real_end <= real_start:
+        raise ValueError(
+            f"`end` ({end}) must be greater than or equal to `start` ({start})."
+        )
+    if real_start >= len(lst):
+        raise ValueError(f"`start` ({start}) must be less than length ({len(lst)}).")
+    if real_end > len(lst):
+        raise ValueError(
+            f"`end` ({end}) must be less than or equal to length ({len(lst)})."
+        )
+
+    lst = list(lst)
+    lst[start:end] = repl
+    return lst
+
+
+def flatten(
+    lst: list[int] | list[list[int]] | list[list[list[int]]],
+) -> tuple[list[int], tuple[list[int], ...]]:
+    """Flatten an arbitrarily (but of uniform depth) nested list and return shape
+    information to enable un-flattening.
+
+    Un-flattening can be performed with the :py:func:`reshape` function.
+
+    lst
+        List to be flattened. Each element must contain all lists or otherwise all items
+        that are considered to be at the "bottom" of the nested structure (e.g. integers).
+        For example, `[[1, 2], [3]]` is permitted and flattens to `[1, 2, 3]`, but
+        `[[1, 2], 3]` is not permitted because the first element is a list, but the second
+        is not.
+
+    """
+
+    def _flatten(
+        lst: list[int] | list[list[int]] | list[list[list[int]]], depth=0
+    ) -> list[int]:
+        out: list[int] = []
+        for item in lst:
+            if isinstance(item, list):
+                out.extend(_flatten(item, depth + 1))
+                all_lens[depth].append(len(item))
+            else:
+                out.append(item)
+        return out
+
+    def _get_max_depth(lst: list[int] | list[list[int]] | list[list[list[int]]]) -> int:
+        val: Any = lst
+        max_depth = 0
+        while isinstance(val, list):
+            max_depth += 1
+            try:
+                val = val[0]
+            except IndexError:
+                # empty list, assume this is max depth
+                break
+        return max_depth
+
+    max_depth = _get_max_depth(lst) - 1
+    all_lens: tuple[list[int], ...] = tuple([] for _ in range(max_depth))
+
+    return _flatten(lst), all_lens
+
+
+def reshape(lst: Sequence[T], lens: Sequence[Sequence[int]]) -> list[TList[T]]:
+    """
+    Reverse the destructuring of the :py:func:`flatten` function.
+    """
+
+    def _reshape(lst: list[T2], lens: Sequence[int]) -> list[list[T2]]:
+        lens_acc = [0, *accumulate(lens)]
+        return [lst[lens_acc[idx] : lens_acc[idx + 1]] for idx in range(len(lens))]
+
+    result: list[TList[T]] = list(lst)
+    for lens_i in lens[::-1]:
+        result = cast("list[TList[T]]", _reshape(result, lens_i))
+
+    return result
+
+
+@overload
+def remap(
+    lst: list[int], mapping_func: Callable[[Sequence[int]], Sequence[T]]
+) -> list[T]: ...
+
+
+@overload
+def remap(
+    lst: list[list[int]], mapping_func: Callable[[Sequence[int]], Sequence[T]]
+) -> list[list[T]]: ...
+
+
+@overload
+def remap(
+    lst: list[list[list[int]]], mapping_func: Callable[[Sequence[int]], Sequence[T]]
+) -> list[list[list[T]]]: ...
+
+
+def remap(lst, mapping_func):
+    """
+    Apply a mapping to a structure of lists with ints (typically indices) as leaves to
+    get a structure of lists with some objects as leaves.
+
+    Parameters
+    ----------
+    lst: list[int] | list[list[int]] | list[list[list[int]]]
+        The structure to remap.
+    mapping_func: Callable[[Sequence[int]], Sequence[T]]
+        The mapping function from sequences of ints to sequences of objects.
+
+    Returns
+    -------
+    list[T] | list[list[T]] | list[list[list[T]]]
+        Nested list structure in same form as input, with leaves remapped.
+    """
+    x, y = flatten(lst)
+    return reshape(mapping_func(x), y)
+
+
+_FSSPEC_URL_RE = re.compile(r"(?:[a-z0-9]+:{1,2})+\/\/")
+
+
+def is_fsspec_url(url: str) -> bool:
+    """
+    Test if a URL appears to be one that can be understood by fsspec.
+    """
+    return bool(_FSSPEC_URL_RE.match(url))
+
+
+class JSONLikeDirSnapShot(DirectorySnapshot):
+    """
+    Overridden DirectorySnapshot from watchdog to allow saving and loading from JSON.
+
+    Parameters
+    ----------
+    root_path: str
+        Where to take the snapshot based at.
+    data: dict[str, list]
+        Serialised snapshot to reload from.
+        See :py:meth:`to_json_like`.
+    """
+
+    def __init__(
+        self,
+        root_path: str | None = None,
+        data: dict[str, list] | None = None,
+        use_strings: bool = False,
+    ):
+        """
+        Create an empty snapshot or load from JSON-like data.
+        """
+
+        #: Where to take the snapshot based at.
+        self.root_path = root_path
+        self._stat_info: dict[bytes | str, os.stat_result] = {}
+        self._inode_to_path: dict[tuple[int, int], bytes | str] = {}
+
+        if data:
+            assert root_path
+            for name, item in data.items():
+                # add root path
+                full_name = str(PurePath(root_path) / PurePath(name))
+                item = [int(i) for i in item] if use_strings else item
+                stat_dat, inode_key = item[:-2], item[-2:]
+                self._stat_info[full_name] = os.stat_result(stat_dat)
+                self._inode_to_path[tuple(inode_key)] = full_name
+
+    def take(self, *args, **kwargs) -> None:
+        """Take the snapshot."""
+        super().__init__(*args, **kwargs)
+
+    def to_json_like(self, use_strings: bool = False) -> dict[str, Any]:
+        """Export to a dict that is JSON-compatible and can be later reloaded.
+
+        The last two integers in `data` for each path are the keys in
+        `self._inode_to_path`.
+
+        """
+        # first key is the root path:
+        root_path = next(iter(self._stat_info))
+
+        # store efficiently:
+        inode_invert = {v: k for k, v in self._inode_to_path.items()}
+        data: dict[str, list] = {
+            str(PurePath(cast("str", k)).relative_to(cast("str", root_path))): [
+                str(i) if use_strings else i for i in [*v, *inode_invert[k]]
+            ]
+            for k, v in self._stat_info.items()
+        }
+
+        return {
+            "root_path": root_path,
+            "data": data,
+            "use_strings": use_strings,
+        }
+
+
+def open_file(filename: str | Path):
+    """Open a file or directory using the default system application."""
+    if sys.platform == "win32":
+        os.startfile(filename)
+    else:
+        opener = "open" if sys.platform == "darwin" else "xdg-open"
+        subprocess.call([opener, filename])
+
+
+@overload
+def get_enum_by_name_or_val(enum_cls: type[E], key: None) -> None: ...
+
+
+@overload
+def get_enum_by_name_or_val(enum_cls: type[E], key: str | int | float | E) -> E: ...
+
+
+def get_enum_by_name_or_val(
+    enum_cls: type[E], key: str | int | float | E | None
+) -> E | None:
+    """Retrieve an enum by name or value, assuming uppercase names and integer values."""
+    if key is None or isinstance(key, enum_cls):
+        return key
+    elif isinstance(key, (int, float)):
+        return enum_cls(int(key))  # retrieve by value
+    elif isinstance(key, str):
+        try:
+            return cast("E", getattr(enum_cls, key.upper()))  # retrieve by name
+        except AttributeError:
+            pass
+    raise ValueError(f"Unknown enum key or value {key!r} for class {enum_cls!r}")
+
+
+_PARAM_SPLIT_RE = re.compile(r"^((?:\w+\.)*)(\w+)(?:\[(\w+)\])?((?:\.\w+)*)$")
+
+
+def split_param_label(
+    param_path: str,
+) -> tuple[str, str] | tuple[str, None] | tuple[None, None]:
+    """Split a parameter path into the path and the label, if present."""
+    m = _PARAM_SPLIT_RE.match(param_path)
+    if not m:
+        return (None, None)
+
+    clean_path = m.group(1) + m.group(2) + m.group(4)
+    bracket_value = m.group(3)
+    return (clean_path, bracket_value)
+
+
+def process_string_nodes(data: T, str_processor: Callable[[str], str]) -> T:
+    """Walk through a nested data structure and process string nodes using a provided
+    callable."""
+
+    if isinstance(data, dict):
+        return cast(
+            "T", {k: process_string_nodes(v, str_processor) for k, v in data.items()}
+        )
+
+    elif isinstance(data, (list, tuple, set, frozenset)):
+        _data = (process_string_nodes(i, str_processor) for i in data)
+        if isinstance(data, tuple):
+            return cast("T", tuple(_data))
+        elif isinstance(data, set):
+            return cast("T", set(_data))
+        elif isinstance(data, frozenset):
+            return cast("T", frozenset(_data))
+        else:
+            return cast("T", list(_data))
+
+    elif isinstance(data, str):
+        return cast("T", str_processor(data))
+
+    return data
+
+
+def linspace_rect(
+    start: Sequence[float],
+    stop: Sequence[float],
+    num: Sequence[int],
+    include: Sequence[str] | None = None,
+    **kwargs,
+) -> NDArray:
+    """Generate a linear space around a rectangle.
+
+    Parameters
+    ----------
+    start
+        Two start values; one for each dimension of the rectangle.
+    stop
+        Two stop values; one for each dimension of the rectangle.
+    num
+        Two number values; one for each dimension of the rectangle.
+    include
+        If specified, include only the specified edges. Choose from "top", "right",
+        "bottom", "left".
+
+    Returns
+    -------
+    rect
+        Coordinates of the rectangle perimeter.
+
+    """
+
+    if num[0] <= 1 or num[1] <= 1:
+        raise ValueError("Both values in `num` must be greater than 1.")
+
+    inc = set(include) if include else {"top", "right", "bottom", "left"}
+
+    c0_range = np.linspace(start=start[0], stop=stop[0], num=num[0], **kwargs)
+    c1_range_all = np.linspace(start=start[1], stop=stop[1], num=num[1], **kwargs)
+
+    c1_range = c1_range_all
+    if "bottom" in inc:
+        c1_range = c1_range[1:]
+    if "top" in inc:
+        c1_range = c1_range[:-1]
+
+    c0_range_c1_start = np.vstack([c0_range, np.repeat(start[1], num[0])])
+    c0_range_c1_stop = np.vstack([c0_range, np.repeat(c1_range_all[-1], num[0])])
+
+    c1_range_c0_start = np.vstack([np.repeat(start[0], len(c1_range)), c1_range])
+    c1_range_c0_stop = np.vstack([np.repeat(c0_range[-1], len(c1_range)), c1_range])
+
+    stacked = []
+    if "top" in inc:
+        stacked.append(c0_range_c1_stop)
+    if "right" in inc:
+        stacked.append(c1_range_c0_stop)
+    if "bottom" in inc:
+        stacked.append(c0_range_c1_start)
+    if "left" in inc:
+        stacked.append(c1_range_c0_start)
+
+    return np.hstack(stacked)
+
+
+def dict_values_process_flat(
+    d: Mapping[T, T2 | list[T2]], callable: Callable[[list[T2]], list[T3]]
+) -> Mapping[T, T3 | list[T3]]:
+    """
+    Return a copy of a dict, where the values are processed by a callable that is to
+    be called only once, and where the values may be single items or lists of items.
+
+    Examples
+    --------
+    d = {'a': 0, 'b': [1, 2], 'c': 5}
+    >>> dict_values_process_flat(d, callable=lambda x: [i + 1 for i in x])
+    {'a': 1, 'b': [2, 3], 'c': 6}
+
+    """
+    flat: list[T2] = []  # values of `d`, flattened
+    is_multi: list[tuple[bool, int]] = (
+        []
+    )  # whether a list, and the number of items to process
+    for i in d.values():
+        if isinstance(i, list):
+            flat.extend(cast("list[T2]", i))
+            is_multi.append((True, len(i)))
+        else:
+            flat.append(cast("T2", i))
+            is_multi.append((False, 1))
+
+    processed = callable(flat)
+
+    out: dict[T, T3 | list[T3]] = {}
+    for idx_i, (m, k) in enumerate(zip(is_multi, d)):
+        start_idx = sum(i[1] for i in is_multi[:idx_i])
+        end_idx = start_idx + m[1]
+        proc_idx_k = processed[start_idx:end_idx]
+        if not m[0]:
+            out[k] = proc_idx_k[0]
+        else:
+            out[k] = proc_idx_k
+
+    return out
+
+
+def nth_key(dct: Iterable[T], n: int) -> T:
+    """
+    Given a dict in some order, get the n'th key of that dict.
+    """
+    it = iter(dct)
+    next(islice(it, n, n), None)
+    return next(it)
+
+
+def nth_value(dct: dict[Any, T], n: int) -> T:
+    """
+    Given a dict in some order, get the n'th value of that dict.
+    """
+    return dct[nth_key(dct, n)]
+
+
+def normalise_timestamp(timestamp: datetime) -> datetime:
+    """
+    Force a timestamp to have UTC as its timezone,
+    then convert to use the local timezone.
+    """
+    return timestamp.replace(tzinfo=timezone.utc).astimezone()
+
+
+def parse_timestamp(timestamp: str | datetime, ts_fmt: str) -> datetime:
+    """
+    Standard timestamp parsing.
+    Ensures that timestamps are internally all UTC.
+    """
+    return normalise_timestamp(
+        timestamp
+        if isinstance(timestamp, datetime)
+        else datetime.strptime(timestamp, ts_fmt)
+    )
+
+
+def current_timestamp() -> datetime:
+    """
+    Get a UTC timestamp for the current time
+    """
+    return datetime.now(timezone.utc)
+
+
+def timedelta_format(td: timedelta) -> str:
+    """
+    Convert time delta to string in standard form.
+    """
+    days, seconds = td.days, td.seconds
+    hours = seconds // (60 * 60)
+    seconds -= hours * (60 * 60)
+    minutes = seconds // 60
+    seconds -= minutes * 60
+    return f"{days}-{hours:02}:{minutes:02}:{seconds:02}"
+
+
+_TD_RE = re.compile(r"(\d+)-(\d+):(\d+):(\d+)")
+
+
+def timedelta_parse(td_str: str) -> timedelta:
+    """
+    Parse a string in standard form as a time delta.
+    """
+    if not (m := _TD_RE.fullmatch(td_str)):
+        raise ValueError("not a supported timedelta form")
+    days, hours, mins, secs = map(int, m.groups())
+    return timedelta(days=days, hours=hours, minutes=mins, seconds=secs)
+
+
+def open_text_resource(package: ModuleType | str, resource: str) -> IO[str]:
+    """
+    Open a file in a package.
+    """
+    try:
+        return resources.files(package).joinpath(resource).open("r")
+    except AttributeError:
+        # < python 3.9; `resource.open_text` deprecated since 3.11
+        return resources.open_text(package, resource)
+
+
+def get_file_context(
+    package: ModuleType | str, src: str | None = None
+) -> AbstractContextManager[Path]:
+    """
+    Find a file or directory in a package.
+    """
+    try:
+        files = resources.files(package)
+        return resources.as_file(files.joinpath(src) if src else files)
+        # raises ModuleNotFoundError
+    except AttributeError:
+        # < python 3.9
+        return resources.path(package, src or "")
+
+
+@contextlib.contextmanager
+def redirect_std_to_file(
+    file,
+    mode: Literal["w", "a"] = "a",
+    ignore: Callable[[BaseException], Literal[True] | int] | None = None,
+) -> Iterator[None]:
+    """Temporarily redirect both stdout and stderr to a file, and if an exception is
+    raised, catch it, print the traceback to that file, and exit.
+
+    File creation is deferred until an actual write is required.
+
+    Parameters
+    ----------
+    ignore
+        Callable to test if a given exception should be ignored. If an exception is
+        not ignored, its traceback will be printed to `file` and the program will
+        exit with exit code 1. The callable should accept one parameter, the
+        exception, and should return True if that exception should be ignored, or
+        an integer representing the exit code to exit the program with if that
+        exception should not be ignored.  By default, no exceptions are ignored.
+
+    """
+    ignore = ignore or (lambda _: 1)
+    with DeferredFileWriter(file, mode=mode) as fp:
+        with contextlib.redirect_stdout(fp):
+            with contextlib.redirect_stderr(fp):
+                try:
+                    yield
+                except BaseException as exc:
+                    ignore_ret = ignore(exc)
+                    if ignore_ret is not True:
+                        traceback.print_exc()
+                        sys.exit(ignore_ret)
+
+
+async def to_thread(func, /, *args, **kwargs):
+    """Copied from https://github.com/python/cpython/blob/4b4227b907a262446b9d276c274feda2590a4e6e/Lib/asyncio/threads.py
+    to support Python 3.8, which does not have `asyncio.to_thread`.
+
+    Asynchronously run function *func* in a separate thread.
+
+    Any *args and **kwargs supplied for this function are directly passed
+    to *func*. Also, the current :class:`contextvars.Context` is propagated,
+    allowing context variables from the main thread to be accessed in the
+    separate thread.
+
+    Return a coroutine that can be awaited to get the eventual result of *func*.
+    """
+    loop = events.get_running_loop()
+    ctx = contextvars.copy_context()
+    func_call = functools.partial(ctx.run, func, *args, **kwargs)
+    return await loop.run_in_executor(None, func_call)