"""Some miscellaneous utility functions.""" # Authors: The MNE-Python contributors. # License: BSD-3-Clause # Copyright the MNE-Python contributors. import fnmatch import gc import hashlib import inspect import os import subprocess import sys import traceback import weakref from contextlib import ExitStack, contextmanager from importlib.resources import files from math import log from queue import Empty, Queue from string import Formatter from textwrap import dedent from threading import Thread import numpy as np from decorator import FunctionMaker from ._logging import logger, verbose, warn from .check import _check_option, _validate_type def _identity_function(x): return x # TODO: no longer needed when py3.9 is minimum supported version def _empty_hash(kind="md5"): func = getattr(hashlib, kind) if "usedforsecurity" in inspect.signature(func).parameters: return func(usedforsecurity=False) else: return func() def _pl(x, non_pl="", pl="s"): """Determine if plural should be used.""" len_x = x if isinstance(x, int | np.generic) else len(x) return non_pl if len_x == 1 else pl def _explain_exception(start=-1, stop=None, prefix="> "): """Explain an exception.""" # start=-1 means "only the most recent caller" etype, value, tb = sys.exc_info() string = traceback.format_list(traceback.extract_tb(tb)[start:stop]) string = "".join(string).split("\n") + traceback.format_exception_only(etype, value) string = ":\n" + prefix + ("\n" + prefix).join(string) return string def _sort_keys(x): """Sort and return keys of dict.""" keys = list(x.keys()) # note: not thread-safe idx = np.argsort([str(k) for k in keys]) keys = [keys[ii] for ii in idx] return keys class _DefaultEventParser: """Parse none standard events.""" def __init__(self): self.event_ids = dict() def __call__(self, description, offset=1): if description not in self.event_ids: self.event_ids[description] = offset + len(self.event_ids) return self.event_ids[description] class _FormatDict(dict): """Help pformat() work properly.""" def __missing__(self, key): return "{" + key + "}" def pformat(temp, **fmt): """Format a template string partially. Examples -------- >>> pformat("{a}_{b}", a='x') 'x_{b}' """ formatter = Formatter() mapping = _FormatDict(fmt) return formatter.vformat(temp, (), mapping) def _enqueue_output(out, queue): for line in iter(out.readline, b""): queue.put(line) @verbose def run_subprocess(command, return_code=False, verbose=None, *args, **kwargs): """Run command using subprocess.Popen. Run command and wait for command to complete. If the return code was zero then return, otherwise raise CalledProcessError. By default, this will also add stdout= and stderr=subproces.PIPE to the call to Popen to suppress printing to the terminal. Parameters ---------- command : list of str | str Command to run as subprocess (see subprocess.Popen documentation). return_code : bool If True, return the return code instead of raising an error if it's non-zero. .. versionadded:: 0.20 %(verbose)s *args, **kwargs : arguments Additional arguments to pass to subprocess.Popen. Returns ------- stdout : str Stdout returned by the process. stderr : str Stderr returned by the process. code : int The return code, only returned if ``return_code == True``. """ all_out = "" all_err = "" # non-blocking adapted from https://stackoverflow.com/questions/375427/non-blocking-read-on-a-subprocess-pipe-in-python#4896288 # noqa: E501 out_q = Queue() err_q = Queue() control_stdout = "stdout" not in kwargs control_stderr = "stderr" not in kwargs with running_subprocess(command, *args, **kwargs) as p: if control_stdout: out_t = Thread(target=_enqueue_output, args=(p.stdout, out_q)) out_t.daemon = True out_t.start() if control_stderr: err_t = Thread(target=_enqueue_output, args=(p.stderr, err_q)) err_t.daemon = True err_t.start() while True: do_break = p.poll() is not None # read all current lines without blocking while True: # process stdout try: out = out_q.get(timeout=0.01) except Empty: break else: out = out.decode("utf-8") log_out = out.removesuffix("\n") logger.info(log_out) all_out += out while True: # process stderr try: err = err_q.get(timeout=0.01) except Empty: break else: err = err.decode("utf-8") err_out = err.removesuffix("\n") # Leave this as logger.warning rather than warn(...) to # mirror the logger.info above for stdout. This function # is basically just a version of subprocess.call, and # shouldn't emit Python warnings due to stderr outputs # (the calling function can check for stderr output and # emit a warning if it wants). logger.warning(err_out) all_err += err if do_break: break output = (all_out, all_err) if return_code: output = output + (p.returncode,) elif p.returncode: stdout = all_out if control_stdout else None stderr = all_err if control_stderr else None raise subprocess.CalledProcessError( p.returncode, command, output=stdout, stderr=stderr ) return output @contextmanager def running_subprocess(command, after="wait", verbose=None, *args, **kwargs): """Context manager to do something with a command running via Popen. Parameters ---------- command : list of str | str Command to run as subprocess (see :class:`python:subprocess.Popen`). after : str Can be: - "wait" to use :meth:`~python:subprocess.Popen.wait` - "communicate" to use :meth:`~python.subprocess.Popen.communicate` - "terminate" to use :meth:`~python:subprocess.Popen.terminate` - "kill" to use :meth:`~python:subprocess.Popen.kill` %(verbose)s *args, **kwargs : arguments Additional arguments to pass to subprocess.Popen. Returns ------- p : instance of Popen The process. """ _validate_type(after, str, "after") _check_option("after", after, ["wait", "terminate", "kill", "communicate"]) contexts = list() for stdxxx in ("stderr", "stdout"): if stdxxx not in kwargs: kwargs[stdxxx] = subprocess.PIPE contexts.append(stdxxx) # Check the PATH environment variable. If run_subprocess() is to be called # frequently this should be refactored so as to only check the path once. env = kwargs.get("env", os.environ) if any(p.startswith("~") for p in env["PATH"].split(os.pathsep)): warn( "Your PATH environment variable contains at least one path " 'starting with a tilde ("~") character. Such paths are not ' "interpreted correctly from within Python. It is recommended " 'that you use "$HOME" instead of "~".' ) if isinstance(command, str): command_str = command else: command = [str(s) for s in command] command_str = " ".join(s for s in command) logger.info(f"Running subprocess: {command_str}") try: p = subprocess.Popen(command, *args, **kwargs) except Exception: if isinstance(command, str): command_name = command.split()[0] else: command_name = command[0] logger.error(f"Command not found: {command_name}") raise try: with ExitStack() as stack: for context in contexts: stack.enter_context(getattr(p, context)) yield p finally: getattr(p, after)() p.wait() def _clean_names(names, remove_whitespace=False, before_dash=True): """Remove white-space on topo matching. This function handles different naming conventions for old VS new VectorView systems (`remove_whitespace`) and removes system specific parts in CTF channel names (`before_dash`). Usage ----- # for new VectorView (only inside layout) ch_names = _clean_names(epochs.ch_names, remove_whitespace=True) # for CTF ch_names = _clean_names(epochs.ch_names, before_dash=True) """ cleaned = [] for name in names: if " " in name and remove_whitespace: name = name.replace(" ", "") if "-" in name and before_dash: name = name.split("-")[0] if name.endswith("_v"): name = name[:-2] cleaned.append(name) if len(set(cleaned)) != len(names): # this was probably not a VectorView or CTF dataset, and we now broke the # dataset by creating duplicates, so let's use the original channel names. return names return cleaned def _get_argvalues(): """Return all arguments (except self) and values of read_raw_xxx.""" # call stack # read_raw_xxx -> -> BaseRaw.__init__ -> _get_argvalues # This is equivalent to `frame = inspect.stack(0)[4][0]` but faster frame = inspect.currentframe() try: for _ in range(3): frame = frame.f_back fname = frame.f_code.co_filename if not fnmatch.fnmatch(fname, "*/mne/io/*"): return None args, _, _, values = inspect.getargvalues(frame) finally: del frame params = dict() for arg in args: params[arg] = values[arg] params.pop("self", None) return params def sizeof_fmt(num): """Turn number of bytes into human-readable str. Parameters ---------- num : int The number of bytes. Returns ------- size : str The size in human-readable format. """ units = ["bytes", "KiB", "MiB", "GiB", "TiB", "PiB"] decimals = [0, 0, 1, 2, 2, 2] if num > 1: exponent = min(int(log(num, 1024)), len(units) - 1) quotient = float(num) / 1024**exponent unit = units[exponent] num_decimals = decimals[exponent] format_string = f"{{0:.{num_decimals}f}} {{1}}" return format_string.format(quotient, unit) if num == 0: return "0 bytes" if num == 1: return "1 byte" def _file_like(obj): # An alternative would be:: # # isinstance(obj, (TextIOBase, BufferedIOBase, RawIOBase, IOBase)) # # but this might be more robust to file-like objects not properly # inheriting from these classes: return all(callable(getattr(obj, name, None)) for name in ("read", "seek")) def _fullname(obj): klass = obj.__class__ module = klass.__module__ if module == "builtins": return klass.__qualname__ return module + "." + klass.__qualname__ def _assert_no_instances(cls, when=""): __tracebackhide__ = True n = 0 ref = list() gc.collect() objs = gc.get_objects() for obj in objs: try: check = isinstance(obj, cls) except Exception: # such as a weakref check = False if check: if cls.__name__ == "Brain": ref.append(f'Brain._cleaned = {getattr(obj, "_cleaned", None)}') rr = gc.get_referrers(obj) count = 0 for r in rr: if ( r is not objs and r is not globals() and r is not locals() and not inspect.isframe(r) ): if isinstance(r, list | dict | tuple): rep = f"len={len(r)}" r_ = gc.get_referrers(r) types = (_fullname(x) for x in r_) types = "/".join(sorted(set(x for x in types if x is not None))) rep += f", {len(r_)} referrers: {types}" del r_ else: rep = repr(r)[:100].replace("\n", " ") # If it's a __closure__, get more information if rep.startswith(" 0 del obj del objs gc.collect() assert n == 0, f"\n{n} {cls.__name__} @ {when}:\n" + "\n".join(ref) def _resource_path(submodule, filename): """Return a full system path to a package resource (AKA a file). Parameters ---------- submodule : str An import-style module or submodule name (e.g., "mne.datasets.testing"). filename : str The file whose full path you want. Returns ------- path : str The full system path to the requested file. """ return files(submodule).joinpath(filename) def repr_html(f): """Decorate _repr_html_ methods. If a _repr_html_ method is decorated with this decorator, the repr in a notebook will show HTML or plain text depending on the config value MNE_REPR_HTML (by default "true", which will render HTML). Parameters ---------- f : function The function to decorate. Returns ------- wrapper : function The decorated function. """ from ..utils import get_config def wrapper(*args, **kwargs): if get_config("MNE_REPR_HTML", "true").lower() == "false": import html r = "
" + html.escape(repr(args[0])) + "
" return r.replace("\n", "
") else: return f(*args, **kwargs) return wrapper def _auto_weakref(function): """Create weakrefs to self (or other free vars in __closure__) then evaluate. When a nested function is defined within an instance method, and the function makes use of ``self``, it creates a reference cycle that the Python garbage collector is not smart enough to resolve, so the parent object is never GC'd. (The reference to ``self`` becomes part of the ``__closure__`` of the nested function). This decorator allows the nested function to access ``self`` without increasing the reference counter on ``self``, which will prevent the memory leak. If the referent is not found (usually because already GC'd) it will short-circuit the decorated function and return ``None``. """ names = function.__code__.co_freevars assert len(names) == len(function.__closure__) __weakref_values__ = dict() evaldict = dict(__weakref_values__=__weakref_values__) for name, value in zip(names, function.__closure__): __weakref_values__[name] = weakref.ref(value.cell_contents) body = dedent(inspect.getsource(function)) body = body.splitlines() for li, line in enumerate(body): if line.startswith(" "): body = body[li:] break old_body = "\n".join(body) body = """\ def %(name)s(%(signature)s): """ for name in names: body += f""" {name} = __weakref_values__[{repr(name)}]() if {name} is None: return """ body = body + old_body fm = FunctionMaker(function) fun = fm.make(body, evaldict, addsource=True) fun.__globals__.update(function.__globals__) assert fun.__closure__ is None, fun.__closure__ return fun