"""Wrappers for the single tasks, multi processor handling."""
# Built ins
import logging
import os
from shutil import rmtree
from collections.abc import Sequence
# External libs
import multiprocessing as mp
import numpy as np
# Locals
import oggm
from oggm import cfg, tasks, utils
from oggm.core import centerlines, flowline
from oggm.exceptions import InvalidParamsError
# MPI
try:
import oggm.mpi as ogmpi
_have_ogmpi = True
except ImportError:
_have_ogmpi = False
# Module logger
log = logging.getLogger(__name__)
# Multiprocessing Pool
_mp_pool = None
def _init_pool_globals(_cfg_contents, global_lock):
cfg.unpack_config(_cfg_contents)
utils.lock = global_lock
def init_mp_pool(reset=False):
"""Necessary because at import time, cfg might be uninitialized"""
global _mp_pool
if _mp_pool and not reset:
return _mp_pool
cfg.CONFIG_MODIFIED = False
if _mp_pool and reset:
_mp_pool.terminate()
_mp_pool = None
cfg_contents = cfg.pack_config()
global_lock = mp.Manager().Lock()
mpp = cfg.PARAMS['mp_processes']
if mpp == -1:
try:
mpp = int(os.environ['SLURM_JOB_CPUS_PER_NODE'])
log.workflow('Multiprocessing: using slurm allocated '
'processors (N={})'.format(mpp))
except KeyError:
mpp = mp.cpu_count()
log.workflow('Multiprocessing: using all available '
'processors (N={})'.format(mpp))
else:
log.workflow('Multiprocessing: using the requested number of '
'processors (N={})'.format(mpp))
_mp_pool = mp.Pool(mpp, initializer=_init_pool_globals,
initargs=(cfg_contents, global_lock))
return _mp_pool
def _merge_dicts(*dicts):
r = {}
for d in dicts:
r.update(d)
return r
class _pickle_copier(object):
"""Pickleable alternative to functools.partial,
Which is not pickleable in python2 and thus doesn't work
with Multiprocessing."""
def __init__(self, func, kwargs):
self.call_func = func
self.out_kwargs = kwargs
def __call__(self, arg):
if self.call_func:
gdir = arg
call_func = self.call_func
else:
call_func, gdir = arg
if isinstance(gdir, Sequence) and not isinstance(gdir, str):
gdir, gdir_kwargs = gdir
gdir_kwargs = _merge_dicts(self.out_kwargs, gdir_kwargs)
return call_func(gdir, **gdir_kwargs)
else:
return call_func(gdir, **self.out_kwargs)
def reset_multiprocessing():
"""Reset multiprocessing state
Call this if you changed configuration parameters mid-run and need them to
be re-propagated to child processes.
"""
global _mp_pool
if _mp_pool:
_mp_pool.terminate()
_mp_pool = None
cfg.CONFIG_MODIFIED = False
[docs]def execute_entity_task(task, gdirs, **kwargs):
"""Execute a task on gdirs.
If you asked for multiprocessing, it will do it.
If ``task`` has more arguments than `gdir` they have to be keyword
arguments.
Parameters
----------
task : function
the entity task to apply
gdirs : list of :py:class:`oggm.GlacierDirectory` objects
the glacier directories to process
"""
# If not iterable it's ok
try:
len(gdirs)
except TypeError:
gdirs = [gdirs]
if len(gdirs) == 0:
return
log.workflow('Execute entity task %s on %d glaciers',
task.__name__, len(gdirs))
if task.__dict__.get('global_task', False):
return task(gdirs, **kwargs)
pc = _pickle_copier(task, kwargs)
if _have_ogmpi:
if ogmpi.OGGM_MPI_COMM is not None:
return ogmpi.mpi_master_spin_tasks(pc, gdirs)
if cfg.PARAMS['use_multiprocessing']:
mppool = init_mp_pool(cfg.CONFIG_MODIFIED)
out = mppool.map(pc, gdirs, chunksize=1)
else:
out = [pc(gdir) for gdir in gdirs]
return out
def execute_parallel_tasks(gdir, tasks):
"""Execute a list of task on a single gdir (experimental!).
This is useful when running a non-sequential list of task on a gdir,
mostly for e.g. different experiments with different output files.
Parameters
----------
gdir : :py:class:`oggm.GlacierDirectory`
the directory to process.
tasks : list
the the list of entity tasks to apply.
Optionally, each list element can be a tuple, with the first element
being the task, and the second element a dict that
will be passed to the task function as ``**kwargs``.
"""
pc = _pickle_copier(None, {})
_tasks = []
for task in tasks:
kwargs = {}
if isinstance(task, Sequence):
task, kwargs = task
_tasks.append((task, (gdir, kwargs)))
if _have_ogmpi:
if ogmpi.OGGM_MPI_COMM is not None:
ogmpi.mpi_master_spin_tasks(pc, _tasks)
return
if cfg.PARAMS['use_multiprocessing']:
mppool = init_mp_pool(cfg.CONFIG_MODIFIED)
mppool.map(pc, _tasks, chunksize=1)
else:
for task in _tasks:
task()
def gdir_from_prepro(entity, from_prepro_level=None,
prepro_border=None, prepro_rgi_version=None,
check_demo_glacier=False):
if prepro_border is None:
prepro_border = int(cfg.PARAMS['border'])
if prepro_rgi_version is None:
prepro_rgi_version = cfg.PARAMS['rgi_version']
try:
rid = entity.RGIId
except AttributeError:
rid = entity
demo_url = False
if check_demo_glacier:
demo_id = utils.demo_glacier_id(rid)
if demo_id is not None:
rid = demo_id
entity = demo_id
demo_url = True
tar_base = utils.get_prepro_gdir(prepro_rgi_version, rid, prepro_border,
from_prepro_level, demo_url=demo_url)
from_tar = os.path.join(tar_base.replace('.tar', ''), rid + '.tar.gz')
return oggm.GlacierDirectory(entity, from_tar=from_tar)
[docs]def init_glacier_regions(rgidf=None, *, reset=False, force=False,
from_prepro_level=None, prepro_border=None,
prepro_rgi_version=None,
from_tar=False, delete_tar=False,
use_demo_glaciers=None):
"""Initializes the list of Glacier Directories for this run.
This is the very first task to do (always). If the directories are already
available in the working directory, use them. If not, create new ones.
Parameters
----------
rgidf : GeoDataFrame or list of ids, optional for pre-computed runs
the RGI glacier outlines. If unavailable, OGGM will parse the
information from the glacier directories found in the working
directory. It is required for new runs.
reset : bool
delete the existing glacier directories if found.
force : bool
setting `reset=True` will trigger a yes/no question to the user. Set
`force=True` to avoid this.
from_prepro_level : int
get the gdir data from the official pre-processed pool. See the
documentation for more information
prepro_border : int
for `from_prepro_level` only: if you want to override the default
behavior which is to use `cfg.PARAMS['border']`
prepro_rgi_version : str
for `from_prepro_level` only: if you want to override the default
behavior which is to use `cfg.PARAMS['rgi_version']`
use_demo_glaciers : bool
whether to check the demo glaciers for download (faster than the
standard prepro downloads). The default is to decide whether or
not to check based on simple crietria such as glacier list size.
from_tar : bool, default=False
extract the gdir data from a tar file. If set to `True`,
will check for a tar file at the expected location in `base_dir`.
delete_tar : bool, default=False
delete the original tar file after extraction.
delete_tar : bool, default=False
delete the original tar file after extraction.
Returns
-------
gdirs : list of :py:class:`oggm.GlacierDirectory` objects
the initialised glacier directories
"""
if reset and not force:
reset = utils.query_yes_no('Delete all glacier directories?')
if prepro_border is None:
prepro_border = int(cfg.PARAMS['border'])
if from_prepro_level and prepro_border not in [10, 80, 160, 250]:
if 'test' not in utils._downloads.GDIR_URL:
raise InvalidParamsError("prepro_border or cfg.PARAMS['border'] "
"should be one of: 10, 80, 160, 250.")
# if reset delete also the log directory
if reset:
fpath = os.path.join(cfg.PATHS['working_dir'], 'log')
if os.path.exists(fpath):
rmtree(fpath)
gdirs = []
new_gdirs = []
if rgidf is None:
if reset:
raise ValueError('Cannot use reset without setting rgidf')
log.workflow('init_glacier_regions by parsing available folders '
'(can be slow).')
# The dirs should be there already
gl_dir = os.path.join(cfg.PATHS['working_dir'], 'per_glacier')
for root, _, files in os.walk(gl_dir):
if files and ('dem.tif' in files):
gdirs.append(oggm.GlacierDirectory(os.path.basename(root)))
else:
# Check if dataframe or list of strs
try:
entities = []
for _, entity in rgidf.iterrows():
entities.append(entity)
except AttributeError:
entities = utils.tolist(rgidf)
# Check demo
if use_demo_glaciers is None:
use_demo_glaciers = len(entities) < 100
if from_prepro_level is not None:
log.workflow('init_glacier_regions from prepro level {} on '
'{} glaciers.'.format(from_prepro_level,
len(entities)))
gdirs = execute_entity_task(gdir_from_prepro, entities,
from_prepro_level=from_prepro_level,
prepro_border=prepro_border,
prepro_rgi_version=prepro_rgi_version,
check_demo_glacier=use_demo_glaciers)
else:
# TODO: if necessary this could use multiprocessing as well
for entity in entities:
gdir = oggm.GlacierDirectory(entity, reset=reset,
from_tar=from_tar,
delete_tar=delete_tar)
if not os.path.exists(gdir.get_filepath('dem')):
new_gdirs.append((gdir, dict(entity=entity)))
gdirs.append(gdir)
# We can set the intersects file automatically here
if (cfg.PARAMS['use_intersects'] and new_gdirs and
(len(cfg.PARAMS['intersects_gdf']) == 0)):
rgi_ids = np.unique(np.sort([t[0].rgi_id for t in new_gdirs]))
rgi_version = new_gdirs[0][0].rgi_version
fp = utils.get_rgi_intersects_entities(rgi_ids, version=rgi_version)
cfg.set_intersects_db(fp)
# If not initialized, run the task in parallel
execute_entity_task(tasks.define_glacier_region, new_gdirs)
return gdirs
[docs]def gis_prepro_tasks(gdirs):
"""Shortcut function: run all flowline preprocessing tasks.
Parameters
----------
gdirs : list of :py:class:`oggm.GlacierDirectory` objects
the glacier directories to process
"""
task_list = [
tasks.glacier_masks,
tasks.compute_centerlines,
tasks.initialize_flowlines,
tasks.compute_downstream_line,
tasks.compute_downstream_bedshape,
tasks.catchment_area,
tasks.catchment_intersections,
tasks.catchment_width_geom,
tasks.catchment_width_correction
]
for task in task_list:
execute_entity_task(task, gdirs)
[docs]def climate_tasks(gdirs):
"""Shortcut function: run all climate related tasks.
Parameters
----------
gdirs : list of :py:class:`oggm.GlacierDirectory` objects
the glacier directories to process
"""
# If not iterable it's ok
try:
len(gdirs)
except TypeError:
gdirs = [gdirs]
# Which climate should we use?
if cfg.PARAMS['baseline_climate'] == 'CRU':
_process_task = tasks.process_cru_data
elif cfg.PARAMS['baseline_climate'] == 'CUSTOM':
_process_task = tasks.process_custom_climate_data
elif cfg.PARAMS['baseline_climate'] == 'HISTALP':
_process_task = tasks.process_histalp_data
else:
raise ValueError('baseline_climate parameter not understood')
execute_entity_task(_process_task, gdirs)
# Then, calibration?
if cfg.PARAMS['run_mb_calibration']:
tasks.compute_ref_t_stars(gdirs)
# Mustar and the apparent mass-balance
execute_entity_task(tasks.local_t_star, gdirs)
execute_entity_task(tasks.mu_star_calibration, gdirs)
[docs]def inversion_tasks(gdirs):
"""Shortcut function: run all ice thickness inversion tasks.
Parameters
----------
gdirs : list of :py:class:`oggm.GlacierDirectory` objects
the glacier directories to process
"""
# Init
execute_entity_task(tasks.prepare_for_inversion, gdirs)
# Inversion for all glaciers
execute_entity_task(tasks.mass_conservation_inversion, gdirs)
# Filter
execute_entity_task(tasks.filter_inversion_output, gdirs)
def merge_glacier_tasks(gdirs, main_rgi_ids, glcdf=None,
filename='climate_monthly', input_filesuffix=''):
"""Shortcut function: run all tasks to merge tributaries to a main glacier
TODO - Automatic search for tributary glaciers
TODO - Every tributary should only be used once
Parameters
----------
gdirs : list of :py:class:`oggm.GlacierDirectory`
all glaciers, main and tributary. Preprocessed and initialised
main_rgi_ids: list of str
RGI IDs of the main glaciers of interest
glcdf: geopandas.GeoDataFrame
which contains the main glaciers, will be downloaded if None
filename: str
Baseline climate file
input_filesuffix: str
Filesuffix to the climate file
Returns
-------
merged_gdirs: list of merged GlacierDirectories
"""
# make sure rgi_ids are iteratable
main_rgi_ids = utils.tolist(main_rgi_ids)
# split main glaciers from candidates
gdirs_main = [gd for gd in gdirs if gd.rgi_id in main_rgi_ids]
# find true tributary glaciers
tributaries = execute_entity_task(
centerlines.intersect_downstream_lines,
gdirs_main, candidates=gdirs)
# make one dictionary and a list of all gdirs for further preprocessing
tribs_dict = {}
gdirs_tribs = []
for trb in tributaries:
tribs_dict.update(trb)
for gd in trb.values():
gdirs_tribs += gd
# check if all tributaries are only used once
rgiids = [gd.rgi_id for gd in gdirs_tribs]
if not len(np.unique(rgiids)) == len(rgiids):
raise RuntimeError('Every tributary glacier should only be used once!')
# create merged glacier directories
gdirs_merged = execute_entity_task(
utils.initialize_merged_gdir, gdirs_main, tribs=tribs_dict,
glcdf=glcdf, filename=filename, input_filesuffix=input_filesuffix)
# Merge the Tributary glacier flowlines to the main glacier one
execute_entity_task(flowline.merge_tributary_flowlines,
gdirs_merged, tribs=tribs_dict,
filename=filename, input_filesuffix=input_filesuffix)
return gdirs_merged