Source code for yt.funcs

"""
Useful functions.  If non-original, see function for citation.



"""
from __future__ import print_function

#-----------------------------------------------------------------------------
# Copyright (c) 2013, yt Development Team.
#
# Distributed under the terms of the Modified BSD License.
#
# The full license is in the file COPYING.txt, distributed with this software.
#-----------------------------------------------------------------------------

from yt.extern.six import string_types
import time, types, signal, inspect, traceback, sys, pdb, os, re
import contextlib
import warnings, struct, subprocess
import numpy as np
from distutils.version import LooseVersion
from math import floor, ceil
from numbers import Number as numeric_type

from yt.extern.six.moves import builtins, urllib
from yt.utilities.exceptions import *
from yt.utilities.logger import ytLogger as mylog
import yt.extern.progressbar as pb
import yt.utilities.rpdb as rpdb
from yt.units.yt_array import YTArray, YTQuantity
from collections import defaultdict
from functools import wraps

# Some functions for handling sequences and other types

[docs]def iterable(obj): """ Grabbed from Python Cookbook / matploblib.cbook. Returns true/false for *obj* iterable. """ try: len(obj) except: return False return True
[docs]def ensure_list(obj): """ This function ensures that *obj* is a list. Typically used to convert a string to a list, for instance ensuring the *fields* as an argument is a list. """ if obj is None: return [obj] if not isinstance(obj, list): return [obj] return obj
def ensure_numpy_array(obj): """ This function ensures that *obj* is a numpy array. Typically used to convert scalar, list or tuple argument passed to functions using Cython. """ if isinstance(obj, np.ndarray): if obj.shape == (): return np.array([obj]) # We cast to ndarray to catch ndarray subclasses return np.array(obj) elif isinstance(obj, (list, tuple)): return np.asarray(obj) else: return np.asarray([obj]) def ensure_tuple(obj): """ This function ensures that *obj* is a tuple. Typically used to convert scalar, list, or array arguments specified by a user in a context where we assume a tuple internally """ if isinstance(obj, tuple): return obj elif isinstance(obj, (list, np.ndarray)): return tuple(obj) else: return (obj,) def read_struct(f, fmt): """ This reads a struct, and only that struct, from an open file. """ s = f.read(struct.calcsize(fmt)) return struct.unpack(fmt, s)
[docs]def just_one(obj): # If we have an iterable, sometimes we only want one item if hasattr(obj,'flat'): if isinstance(obj, YTArray): return YTQuantity(obj.flat[0], obj.units, registry=obj.units.registry) return obj.flat[0] elif iterable(obj): return obj[0] return obj
def compare_dicts(dict1, dict2): if not set(dict1) <= set(dict2): return False for key in dict1.keys(): if dict1[key] is not None and dict2[key] is not None: if isinstance(dict1[key], dict): if compare_dicts(dict1[key], dict2[key]): continue else: return False try: comparison = (dict1[key] == dict2[key]).all() except AttributeError: comparison = (dict1[key] == dict2[key]) if not comparison: return False return True # Taken from # http://www.goldb.org/goldblog/2008/02/06/PythonConvertSecsIntoHumanReadableTimeStringHHMMSS.aspx
[docs]def humanize_time(secs): """ Takes *secs* and returns a nicely formatted string """ mins, secs = divmod(secs, 60) hours, mins = divmod(mins, 60) return '%02d:%02d:%02d' % (hours, mins, secs) # # Some function wrappers that come in handy once in a while # # we use the resource module to get the memory page size
try: import resource except ImportError: pass def get_memory_usage(subtract_share = False): """ Returning resident size in megabytes """ pid = os.getpid() try: pagesize = resource.getpagesize() except NameError: return -1024 status_file = "/proc/%s/statm" % (pid) if not os.path.isfile(status_file): return -1024 line = open(status_file).read() size, resident, share, text, library, data, dt = [int(i) for i in line.split()] if subtract_share: resident -= share return resident * pagesize / (1024 * 1024) # return in megs
[docs]def time_execution(func): r""" Decorator for seeing how long a given function takes, depending on whether or not the global 'yt.timefunctions' config parameter is set. """ @wraps(func) def wrapper(*arg, **kw): t1 = time.time() res = func(*arg, **kw) t2 = time.time() mylog.debug('%s took %0.3f s', func.__name__, (t2-t1)) return res from yt.config import ytcfg if ytcfg.getboolean("yt","timefunctions") == True: return wrapper else: return func
[docs]def rootonly(func): """ This is a decorator that, when used, will only call the function on the root processor and then broadcast the results of the function to all other processors. This can be used like so: .. code-block:: python @rootonly def some_root_only_function(...): """ from yt.config import ytcfg @wraps(func) def check_parallel_rank(*args, **kwargs): if ytcfg.getint("yt","__topcomm_parallel_rank") > 0: return return func(*args, **kwargs) return check_parallel_rank
def rootloginfo(*args): from yt.config import ytcfg if ytcfg.getint("yt", "__topcomm_parallel_rank") > 0: return mylog.info(*args)
[docs]def deprecate(func): """ This decorator issues a deprecation warning. This can be used like so: .. code-block:: python @deprecate def some_really_old_function(...): """ @wraps(func) def run_func(*args, **kwargs): warnings.warn("%s has been deprecated and may be removed without notice!" \ % func.__name__, DeprecationWarning, stacklevel=2) func(*args, **kwargs) return run_func
[docs]def pdb_run(func): """ This decorator inserts a pdb session on top of the call-stack into a function. This can be used like so: .. code-block:: python @pdb_run def some_function_to_debug(...): """ @wraps(func) def wrapper(*args, **kw): pdb.runcall(func, *args, **kw) return wrapper
__header = """ == Welcome to the embedded IPython Shell == You are currently inside the function: %(fname)s Defined in: %(filename)s:%(lineno)s """ def get_ipython_api_version(): import IPython if LooseVersion(IPython.__version__) <= LooseVersion('0.10'): api_version = '0.10' elif LooseVersion(IPython.__version__) <= LooseVersion('1.0'): api_version = '0.11' else: api_version = '1.0' return api_version
[docs]def insert_ipython(num_up=1): """ Placed inside a function, this will insert an IPython interpreter at that current location. This will enabled detailed inspection of the current exeuction environment, as well as (optional) modification of that environment. *num_up* refers to how many frames of the stack get stripped off, and defaults to 1 so that this function itself is stripped off. """ api_version = get_ipython_api_version() frame = inspect.stack()[num_up] loc = frame[0].f_locals.copy() glo = frame[0].f_globals dd = dict(fname = frame[3], filename = frame[1], lineno = frame[2]) if api_version == '0.10': ipshell = IPython.Shell.IPShellEmbed() ipshell(header = __header % dd, local_ns = loc, global_ns = glo) else: from IPython.config.loader import Config cfg = Config() cfg.InteractiveShellEmbed.local_ns = loc cfg.InteractiveShellEmbed.global_ns = glo IPython.embed(config=cfg, banner2 = __header % dd) if api_version == '0.11': from IPython.frontend.terminal.embed import InteractiveShellEmbed else: from IPython.terminal.embed import InteractiveShellEmbed ipshell = InteractiveShellEmbed(config=cfg) del ipshell # # Our progress bar types and how to get one #
class DummyProgressBar(object): # This progressbar gets handed if we don't # want ANY output def __init__(self, *args, **kwargs): return def update(self, *args, **kwargs): return def finish(self, *args, **kwargs): return class ParallelProgressBar(object): # This is just a simple progress bar # that prints on start/stop def __init__(self, title, maxval): self.title = title mylog.info("Starting '%s'", title) def update(self, *args, **kwargs): return def finish(self): mylog.info("Finishing '%s'", self.title) class GUIProgressBar(object): def __init__(self, title, maxval): import wx self.maxval = maxval self.last = 0 self._pbar = wx.ProgressDialog("Working...", title, maximum=maxval, style=wx.PD_REMAINING_TIME|wx.PD_ELAPSED_TIME|wx.PD_APP_MODAL) def update(self, val): # An update is only meaningful if it's on the order of 1/100 or greater if ceil(100*self.last / self.maxval) + 1 == \ floor(100*val / self.maxval) or val == self.maxval: self._pbar.Update(val) self.last = val def finish(self): self._pbar.Destroy()
[docs]def get_pbar(title, maxval): """ This returns a progressbar of the most appropriate type, given a *title* and a *maxval*. """ maxval = max(maxval, 1) from yt.config import ytcfg if ytcfg.getboolean("yt", "suppressStreamLogging") or \ "__IPYTHON__" in dir(builtins) or \ ytcfg.getboolean("yt", "__withintesting"): return DummyProgressBar() elif ytcfg.getboolean("yt", "__withinreason"): from yt.gui.reason.extdirect_repl import ExtProgressBar return ExtProgressBar(title, maxval) elif ytcfg.getboolean("yt", "__parallel"): return ParallelProgressBar(title, maxval) widgets = [ title, pb.Percentage(), ' ', pb.Bar(marker=pb.RotatingMarker()), ' ', pb.ETA(), ' '] pbar = pb.ProgressBar(widgets=widgets, maxval=maxval).start() return pbar
[docs]def only_on_root(func, *args, **kwargs): """ This function accepts a *func*, a set of *args* and *kwargs* and then only on the root processor calls the function. All other processors get "None" handed back. """ from yt.config import ytcfg if kwargs.pop("global_rootonly", False): cfg_option = "__global_parallel_rank" else: cfg_option = "__topcomm_parallel_rank" if not ytcfg.getboolean("yt","__parallel"): return func(*args,**kwargs) if ytcfg.getint("yt", cfg_option) > 0: return return func(*args, **kwargs)
[docs]def is_root(): """ This function returns True if it is on the root processor of the topcomm and False otherwise. """ from yt.config import ytcfg cfg_option = "__topcomm_parallel_rank" if not ytcfg.getboolean("yt","__parallel"): return True if ytcfg.getint("yt", cfg_option) > 0: return False return True # # Our signal and traceback handling functions #
def signal_print_traceback(signo, frame): print(traceback.print_stack(frame)) def signal_problem(signo, frame): raise RuntimeError() def signal_ipython(signo, frame): insert_ipython(2)
[docs]def paste_traceback(exc_type, exc, tb): """ This is a traceback handler that knows how to paste to the pastebin. Should only be used in sys.excepthook. """ sys.__excepthook__(exc_type, exc, tb) from yt.extern.six.moves import StringIO, xmlrpc_client p = xmlrpc_client.ServerProxy( "http://paste.yt-project.org/xmlrpc/", allow_none=True) s = StringIO() traceback.print_exception(exc_type, exc, tb, file=s) s = s.getvalue() ret = p.pastes.newPaste('pytb', s, None, '', '', True) print() print("Traceback pasted to http://paste.yt-project.org/show/%s" % (ret)) print()
def paste_traceback_detailed(exc_type, exc, tb): """ This is a traceback handler that knows how to paste to the pastebin. Should only be used in sys.excepthook. """ import cgitb from yt.extern.six.moves import StringIO, xmlrpc_client s = StringIO() handler = cgitb.Hook(format="text", file = s) handler(exc_type, exc, tb) s = s.getvalue() print(s) p = xmlrpc_client.ServerProxy( "http://paste.yt-project.org/xmlrpc/", allow_none=True) ret = p.pastes.newPaste('text', s, None, '', '', True) print() print("Traceback pasted to http://paste.yt-project.org/show/%s" % (ret)) print() _ss = "fURbBUUBE0cLXgETJnZgJRMXVhVGUQpQAUBuehQMUhJWRFFRAV1ERAtBXw1dAxMLXT4zXBFfABNN\nC0ZEXw1YUURHCxMXVlFERwxWCQw=\n" def _rdbeta(key): import itertools, base64 enc_s = base64.decodestring(_ss) dec_s = ''.join([ chr(ord(a) ^ ord(b)) for a, b in zip(enc_s, itertools.cycle(key)) ]) print(dec_s) # # Some exceptions # class NoCUDAException(Exception): pass class YTEmptyClass(object): pass def update_hg(path, skip_rebuild = False): if sys.version_info >= (3,0,0): print("python-hglib does not currently work with Python 3,") print("so this function is currently disabled.") return -1 try: import hglib except ImportError: print("Updating requires python-hglib to be installed.") print("Try: pip install python-hglib") return -1 f = open(os.path.join(path, "yt_updater.log"), "a") repo = hglib.open(path) repo.pull() ident = repo.identify() if "+" in ident: print("Can't rebuild modules by myself.") print("You will have to do this yourself. Here's a sample commands:") print("") print(" $ cd %s" % (path)) print(" $ hg up") print(" $ %s setup.py develop" % (sys.executable)) return 1 print("Updating the repository") f.write("Updating the repository\n\n") repo.update(check=True) f.write("Updated from %s to %s\n\n" % (ident, repo.identify())) if skip_rebuild: return f.write("Rebuilding modules\n\n") p = subprocess.Popen([sys.executable, "setup.py", "build_ext", "-i"], cwd=path, stdout = subprocess.PIPE, stderr = subprocess.STDOUT) stdout, stderr = p.communicate() f.write(stdout) f.write("\n\n") if p.returncode: print("BROKEN: See %s" % (os.path.join(path, "yt_updater.log"))) sys.exit(1) f.write("Successful!\n") print("Updated successfully.") def get_hg_version(path): if sys.version_info >= (3,0,0): print("python-hglib does not currently work with Python 3,") print("so this function is currently disabled.") return -1 try: import hglib except ImportError: print("Updating and precise version information requires ") print("python-hglib to be installed.") print("Try: pip install python-hglib") return -1 repo = hglib.open(path) return repo.identify() def get_yt_version(): try: from yt.__hg_version__ import hg_version return hg_version except ImportError: pass import pkg_resources yt_provider = pkg_resources.get_provider("yt") path = os.path.dirname(yt_provider.module_path) version = get_hg_version(path)[:12] return version def get_version_stack(): import numpy, matplotlib, h5py version_info = {} version_info['yt'] = get_yt_version() version_info['numpy'] = numpy.version.version version_info['matplotlib'] = matplotlib.__version__ version_info['h5py'] = h5py.version.version return version_info def get_script_contents(): top_frame = inspect.stack()[-1] finfo = inspect.getframeinfo(top_frame[0]) if finfo[2] != "<module>": return None if not os.path.exists(finfo[0]): return None try: contents = open(finfo[0]).read() except: contents = None return contents def download_file(url, filename): class MyURLopener(urllib.request.FancyURLopener): def http_error_default(self, url, fp, errcode, errmsg, headers): raise RuntimeError("Attempt to download file from %s failed with error %s: %s." % \ (url, errcode, errmsg)) fn, h = MyURLopener().retrieve(url, filename) return fn # This code snippet is modified from Georg Brandl def bb_apicall(endpoint, data, use_pass = True): uri = 'https://api.bitbucket.org/1.0/%s/' % endpoint # since bitbucket doesn't return the required WWW-Authenticate header when # making a request without Authorization, we cannot use the standard urllib2 # auth handlers; we have to add the requisite header from the start if data is not None: data = urllib.parse.urlencode(data) req = urllib.request.Request(uri, data) if use_pass: username = raw_input("Bitbucket Username? ") password = getpass.getpass() upw = '%s:%s' % (username, password) req.add_header('Authorization', 'Basic %s' % base64.b64encode(upw).strip()) return urllib.request.urlopen(req).read() def get_yt_supp(): supp_path = os.path.join(os.environ["YT_DEST"], "src", "yt-supplemental") # Now we check that the supplemental repository is checked out. if not os.path.isdir(supp_path): print() print("*** The yt-supplemental repository is not checked ***") print("*** out. I can do this for you, but because this ***") print("*** is a delicate act, I require you to respond ***") print("*** to the prompt with the word 'yes'. ***") print() response = raw_input("Do you want me to try to check it out? ") if response != "yes": print() print("Okay, I understand. You can check it out yourself.") print("This command will do it:") print() print("$ hg clone http://bitbucket.org/yt_analysis/yt-supplemental/ ", end=' ') print("%s" % (supp_path)) print() sys.exit(1) rv = commands.clone(uu, "http://bitbucket.org/yt_analysis/yt-supplemental/", supp_path) if rv: print("Something has gone wrong. Quitting.") sys.exit(1) # Now we think we have our supplemental repository. return supp_path def fix_length(length, ds=None): assert ds is not None if ds is not None: registry = ds.unit_registry else: registry = None if isinstance(length, YTArray): if registry is not None: length.units.registry = registry return length.in_units("code_length") if isinstance(length, numeric_type): return YTArray(length, 'code_length', registry=registry) length_valid_tuple = isinstance(length, (list, tuple)) and len(length) == 2 unit_is_string = isinstance(length[1], str) if length_valid_tuple and unit_is_string: return YTArray(*length, registry=registry) else: raise RuntimeError("Length %s is invalid" % str(length)) @contextlib.contextmanager def parallel_profile(prefix): r"""A context manager for profiling parallel code execution using cProfile This is a simple context manager that automatically profiles the execution of a snippet of code. Parameters ---------- prefix : string A string name to prefix outputs with. Examples -------- >>> with parallel_profile('my_profile'): ... yt.PhasePlot(ds.all_data(), 'density', 'temperature', 'cell_mass') """ import cProfile from yt.config import ytcfg fn = "%s_%04i_%04i.cprof" % (prefix, ytcfg.getint("yt", "__topcomm_parallel_size"), ytcfg.getint("yt", "__topcomm_parallel_rank")) p = cProfile.Profile() p.enable() yield fn p.disable() p.dump_stats(fn) def get_num_threads(): from .config import ytcfg nt = ytcfg.getint("yt","numthreads") if nt < 0: return os.environ.get("OMP_NUM_THREADS", 0) return nt def fix_axis(axis, ds): return ds.coordinates.axis_id.get(axis, axis) def get_image_suffix(name): suffix = os.path.splitext(name)[1] return suffix if suffix in ['.png', '.eps', '.ps', '.pdf'] else '' def ensure_dir_exists(path): r"""Create all directories in path recursively in a parallel safe manner""" my_dir = os.path.dirname(path) if not my_dir: return if not os.path.exists(my_dir): only_on_root(os.makedirs, my_dir) def ensure_dir(path): r"""Parallel safe directory maker.""" if not os.path.exists(path): only_on_root(os.makedirs, path) return path def validate_width_tuple(width): if not iterable(width) or len(width) != 2: raise YTInvalidWidthError("width (%s) is not a two element tuple" % width) if not isinstance(width[0], numeric_type) and isinstance(width[1], string_types): msg = "width (%s) is invalid. " % str(width) msg += "Valid widths look like this: (12, 'au')" raise YTInvalidWidthError(msg) def camelcase_to_underscore(name): s1 = re.sub('(.)([A-Z][a-z]+)', r'\1_\2', name) return re.sub('([a-z0-9])([A-Z])', r'\1_\2', s1).lower() def set_intersection(some_list): if len(some_list) == 0: return set([]) # This accepts a list of iterables, which we get the intersection of. s = set(some_list[0]) for l in some_list[1:]: s.intersection_update(l) return s @contextlib.contextmanager def memory_checker(interval = 15, dest = None): r"""This is a context manager that monitors memory usage. Parameters ---------- interval : int The number of seconds between printing the current memory usage in gigabytes of the current Python interpreter. Examples -------- >>> with memory_checker(10): ... arr = np.zeros(1024*1024*1024, dtype="float64") ... time.sleep(15) ... del arr """ import threading if dest is None: dest = sys.stdout class MemoryChecker(threading.Thread): def __init__(self, event, interval): self.event = event self.interval = interval threading.Thread.__init__(self) def run(self): while not self.event.wait(self.interval): print("MEMORY: %0.3e gb" % (get_memory_usage()/1024.), file=dest) e = threading.Event() mem_check = MemoryChecker(e, interval) mem_check.start() try: yield finally: e.set() def deprecated_class(cls): @wraps(cls) def _func(*args, **kwargs): # Note we use SyntaxWarning because by default, DeprecationWarning is # not shown. warnings.warn( "This usage is deprecated. Please use %s instead." % cls.__name__, SyntaxWarning, stacklevel=2) return cls(*args, **kwargs) return _func def enable_plugins(): import yt from yt.fields.my_plugin_fields import my_plugins_fields from yt.config import ytcfg my_plugin_name = ytcfg.get("yt","pluginfilename") # We assume that it is with respect to the $HOME/.yt directory if os.path.isfile(my_plugin_name): _fn = my_plugin_name else: _fn = os.path.expanduser("~/.yt/%s" % my_plugin_name) if os.path.isfile(_fn): mylog.info("Loading plugins from %s", _fn) execdict = yt.__dict__.copy() execdict['add_field'] = my_plugins_fields.add_field with open(_fn) as f: code = compile(f.read(), _fn, 'exec') exec(code, execdict) def fix_unitary(u): if u == '1': return 'unitary' else: return u