def post(info, sample=None):
logger_setup(info.get(_debug), info.get(_debug_file))
log = logging.getLogger(__name__.split(".")[-1])
# MARKED FOR DEPRECATION IN v3.0
# BEHAVIOUR TO BE REPLACED BY ERROR:
check_deprecated_modules_path(info)
# END OF DEPRECATION BLOCK
try:
info_post = info[_post]
except KeyError:
raise LoggedError(log, "No 'post' block given. Nothing to do!")
if get_mpi_rank():
log.warning(
"Post-processing is not yet MPI-aware. Doing nothing for rank > 1 processes.")
return
if info.get(_resume):
log.warning("Resuming not implemented for post-processing. Re-starting.")
# 1. Load existing sample
output_in = get_output(output_prefix=info.get(_output_prefix))
if output_in:
try:
info_in = output_in.reload_updated_info()
except FileNotFoundError:
raise LoggedError(log, "Error loading input model: "
"could not find input info at %s",
output_in.file_updated)
else:
info_in = deepcopy_where_possible(info)
dummy_model_in = DummyModel(info_in[_params], info_in[kinds.likelihood],
info_in.get(_prior, None))
if output_in:
if not output_in.find_collections():
raise LoggedError(log, "No samples found for the input model with prefix %s",
os.path.join(output_in.folder, output_in.prefix))
collection_in = output_in.load_collections(
dummy_model_in, skip=info_post.get("skip", 0), thin=info_post.get("thin", 1),
concatenate=True)
elif sample:
if isinstance(sample, Collection):
sample = [sample]
collection_in = deepcopy(sample[0])
for s in sample[1:]:
try:
collection_in.append(s)
except:
raise LoggedError(log, "Failed to load some of the input samples.")
else:
raise LoggedError(log,
"Not output from where to load from or input collections given.")
log.info("Will process %d samples.", len(collection_in))
if len(collection_in) <= 1:
raise LoggedError(
log, "Not enough samples for post-processing. Try using a larger sample, "
"or skipping or thinning less.")
# 2. Compare old and new info: determine what to do
add = info_post.get(_post_add, {}) or {}
remove = info_post.get(_post_remove, {})
# Add a dummy 'one' likelihood, to absorb unused parameters
if not add.get(kinds.likelihood):
add[kinds.likelihood] = {}
add[kinds.likelihood]["one"] = None
# Expand the "add" info
add = update_info(add)
# 2.1 Adding/removing derived parameters and changes in priors of sampled parameters
out = {_params: deepcopy_where_possible(info_in[_params])}
for p in remove.get(_params, {}):
pinfo = info_in[_params].get(p)
if pinfo is None or not is_derived_param(pinfo):
raise LoggedError(
log,
"You tried to remove parameter '%s', which is not a derived parameter. "
"Only derived parameters can be removed during post-processing.", p)
out[_params].pop(p)
# Force recomputation of aggregated chi2
for p in list(out[_params]):
if p.startswith(_get_chi2_name("")):
out[_params].pop(p)
mlprior_names_add = []
for p, pinfo in add.get(_params, {}).items():
pinfo_in = info_in[_params].get(p)
if is_sampled_param(pinfo):
if not is_sampled_param(pinfo_in):
# No added sampled parameters (de-marginalisation not implemented)
if pinfo_in is None:
raise LoggedError(
log, "You added a new sampled parameter %r (maybe accidentally "
"by adding a new likelihood that depends on it). "
"Adding new sampled parameters is not possible. Try fixing "
"it to some value.", p)
else:
raise LoggedError(
log,
"You tried to change the prior of parameter '%s', "
"but it was not a sampled parameter. "
"To change that prior, you need to define as an external one.", p)
if mlprior_names_add[:1] != _prior_1d_name:
mlprior_names_add = ([_minuslogprior + _separator + _prior_1d_name]
+ mlprior_names_add)
elif is_derived_param(pinfo):
if p in out[_params]:
raise LoggedError(
log, "You tried to add derived parameter '%s', which is already "
"present. To force its recomputation, 'remove' it too.", p)
elif is_fixed_param(pinfo):
# Only one possibility left "fixed" parameter that was not present before:
# input of new likelihood, or just an argument for dynamical derived (dropped)
if ((p in info_in[_params] and
pinfo[partag.value] != (pinfo_in or {}).get(partag.value, None))):
raise LoggedError(
log,
"You tried to add a fixed parameter '%s: %r' that was already present"
" but had a different value or was not fixed. This is not allowed. "
"The old info of the parameter was '%s: %r'",
p, dict(pinfo), p, dict(pinfo_in))
else:
raise LoggedError(log, "This should not happen. Contact the developers.")
out[_params][p] = pinfo
# For the likelihood only, turn the rest of *derived* parameters into constants,
# so that the likelihoods do not try to compute them)
# But be careful to exclude *input* params that have a "derived: True" value
# (which in "updated info" turns into "derived: 'lambda [x]: [x]'")
out_params_like = deepcopy_where_possible(out[_params])
for p, pinfo in out_params_like.items():
if ((is_derived_param(pinfo) and not (partag.value in pinfo)
and p not in add.get(_params, {}))):
out_params_like[p] = {partag.value: np.nan, partag.drop: True}
# 2.2 Manage adding/removing priors and likelihoods
warn_remove = False
for level in [_prior, kinds.likelihood]:
out[level] = getattr(dummy_model_in, level)
if level == _prior:
out[level].remove(_prior_1d_name)
for pdf in info_post.get(_post_remove, {}).get(level, []) or []:
try:
out[level].remove(pdf)
warn_remove = True
except ValueError:
raise LoggedError(
log, "Trying to remove %s '%s', but it is not present. "
"Existing ones: %r", level, pdf, out[level])
if warn_remove:
log.warning("You are removing a prior or likelihood pdf. "
"Notice that if the resulting posterior is much wider "
"than the original one, or displaced enough, "
"it is probably safer to explore it directly.")
if _prior in add:
mlprior_names_add += [_minuslogprior + _separator + name for name in add[_prior]]
out[_prior] += list(add[_prior])
prior_recompute_1d = (
mlprior_names_add[:1] == [_minuslogprior + _separator + _prior_1d_name])
# Don't initialise the theory code if not adding/recomputing theory,
# theory-derived params or likelihoods
recompute_theory = info_in.get(kinds.theory) and not (
list(add[kinds.likelihood]) == ["one"] and
not any(is_derived_param(pinfo) for pinfo in add.get(_params, {}).values()))
if recompute_theory:
# Inherit from the original chain (needs input|output_params, renames, etc
add_theory = add.get(kinds.theory)
if add_theory:
info_theory_out = {}
if len(add_theory) > 1:
log.warning('Importance sampling with more than one theory is '
'not really tested')
add_theory = add_theory.copy()
for theory, theory_info in info_in[kinds.theory].items():
theory_copy = deepcopy_where_possible(theory_info)
if theory in add_theory:
info_theory_out[theory] = \
recursive_update(theory_copy, add_theory.pop(theory))
else:
info_theory_out[theory] = theory_copy
info_theory_out.update(add_theory)
else:
info_theory_out = deepcopy_where_possible(info_in[kinds.theory])
else:
info_theory_out = None
chi2_names_add = [
_get_chi2_name(name) for name in add[kinds.likelihood] if name != "one"]
out[kinds.likelihood] += [l for l in add[kinds.likelihood] if l != "one"]
if recompute_theory:
log.warning("You are recomputing the theory, but in the current version this does"
" not force recomputation of any likelihood or derived parameter, "
"unless explicitly removed+added.")
for level in [_prior, kinds.likelihood]:
for i, x_i in enumerate(out[level]):
if x_i in list(out[level])[i + 1:]:
raise LoggedError(
log, "You have added %s '%s', which was already present. If you "
"want to force its recomputation, you must also 'remove' it.",
level, x_i)
# 3. Create output collection
if _post_suffix not in info_post:
raise LoggedError(log, "You need to provide a '%s' for your chains.",
_post_suffix)
# Use default prefix if it exists. If it does not, produce no output by default.
# {post: {output: None}} suppresses output, and if it's a string, updates it.
out_prefix = info_post.get(_output_prefix, info.get(_output_prefix))
if out_prefix not in [None, False]:
out_prefix += _separator_files + _post + _separator_files + info_post[
_post_suffix]
output_out = get_output(output_prefix=out_prefix, force=info.get(_force))
if output_out and not output_out.force and output_out.find_collections():
raise LoggedError(log, "Found existing post-processing output with prefix %r. "
"Delete it manually or re-run with `force: True` "
"(or `-f`, `--force` from the shell).", out_prefix)
elif output_out and output_out.force:
output_out.delete_infos()
for regexp in output_out.find_collections():
output_out.delete_with_regexp(re.compile(regexp))
info_out = deepcopy_where_possible(info)
info_out[_post] = info_post
# Updated with input info and extended (updated) add info
info_out.update(info_in)
info_out[_post][_post_add] = add
dummy_model_out = DummyModel(out[_params], out[kinds.likelihood],
info_prior=out[_prior])
if recompute_theory:
# TODO: May need updating for more than one, or maybe can be removed
theory = list(info_theory_out)[0]
if _input_params not in info_theory_out[theory]:
raise LoggedError(
log,
"You appear to be post-processing a chain generated with an older "
"version of Cobaya. For post-processing to work, please edit the "
"'[root].updated.yaml' file of the original chain to add, inside the "
"theory code block, the list of its input parameters. E.g.\n----\n"
"theory:\n %s:\n input_params: [param1, param2, ...]\n"
"----\nIf you get strange errors later, it is likely that you did not "
"specify the correct set of theory parameters.\n"
"The full set of input parameters are %s.",
theory, list(dummy_model_out.parameterization.input_params()))
# TODO: check allow_renames=False?
# TODO: May well be simplifications here, this is v close to pre-refactor logic
# Have not gone through or understood all the parameterization stuff
model_add = Model(out_params_like, add[kinds.likelihood], info_prior=add.get(_prior),
info_theory=info_theory_out, packages_path=info.get(_packages_path),
allow_renames=False, post=True,
prior_parameterization=dummy_model_out.parameterization)
# Remove auxiliary "one" before dumping -- 'add' *is* info_out[_post][_post_add]
add[kinds.likelihood].pop("one")
collection_out = Collection(dummy_model_out, output_out, name="1")
output_out.check_and_dump_info(None, info_out, check_compatible=False)
# Prepare recomputation of aggregated chi2
# (they need to be recomputed by hand, because its autocomputation won't pick up
# old likelihoods for a given type)
all_types = {
like: str_to_list(add[kinds.likelihood].get(
like, info_in[kinds.likelihood].get(like)).get("type", []) or [])
for like in out[kinds.likelihood]}
types = set(chain(*list(all_types.values())))
inv_types = {t: [like for like, like_types in all_types.items() if t in like_types]
for t in types}
# 4. Main loop!
log.info("Running post-processing...")
last_percent = 0
for i, point in collection_in.data.iterrows():
log.debug("Point: %r", point)
sampled = [point[param] for param in
dummy_model_in.parameterization.sampled_params()]
derived = {param: point.get(param, None)
for param in dummy_model_out.parameterization.derived_params()}
inputs = {param: point.get(
param, dummy_model_in.parameterization.constant_params().get(
param, dummy_model_out.parameterization.constant_params().get(
param, None)))
for param in dummy_model_out.parameterization.input_params()}
# Solve inputs that depend on a function and were not saved
# (we don't use the Parameterization_to_input method in case there are references
# to functions that cannot be loaded at the moment)
for p, value in inputs.items():
if value is None:
func = dummy_model_out.parameterization._input_funcs[p]
args = dummy_model_out.parameterization._input_args[p]
inputs[p] = func(*[point.get(arg) for arg in args])
# Add/remove priors
priors_add = model_add.prior.logps(sampled)
if not prior_recompute_1d:
priors_add = priors_add[1:]
logpriors_add = dict(zip(mlprior_names_add, priors_add))
logpriors_new = [logpriors_add.get(name, - point.get(name, 0))
for name in collection_out.minuslogprior_names]
if log.getEffectiveLevel() <= logging.DEBUG:
log.debug(
"New set of priors: %r", dict(zip(dummy_model_out.prior, logpriors_new)))
if -np.inf in logpriors_new:
continue
# Add/remove likelihoods
output_like = []
if add[kinds.likelihood]:
# Notice "one" (last in likelihood_add) is ignored: not in chi2_names
loglikes_add, output_like = model_add.logps(inputs, return_derived=True)
loglikes_add = dict(zip(chi2_names_add, loglikes_add))
output_like = dict(zip(model_add.output_params, output_like))
else:
loglikes_add = dict()
loglikes_new = [loglikes_add.get(name, -0.5 * point.get(name, 0))
for name in collection_out.chi2_names]
if log.getEffectiveLevel() <= logging.DEBUG:
log.debug(
"New set of likelihoods: %r",
dict(zip(dummy_model_out.likelihood, loglikes_new)))
if output_like:
log.debug("New set of likelihood-derived parameters: %r", output_like)
if -np.inf in loglikes_new:
continue
# Add/remove derived parameters and change priors of sampled parameters
for p in add[_params]:
if p in dummy_model_out.parameterization._directly_output:
derived[p] = output_like[p]
elif p in dummy_model_out.parameterization._derived_funcs:
func = dummy_model_out.parameterization._derived_funcs[p]
args = dummy_model_out.parameterization._derived_args[p]
derived[p] = func(
*[point.get(arg, output_like.get(arg, None)) for arg in args])
# We need to recompute the aggregated chi2 by hand
for type_, likes in inv_types.items():
derived[_get_chi2_name(type_)] = sum(
[-2 * lvalue for lname, lvalue
in zip(collection_out.chi2_names, loglikes_new)
if _undo_chi2_name(lname) in likes])
if log.getEffectiveLevel() <= logging.DEBUG:
log.debug("New derived parameters: %r",
dict([(p, derived[p])
for p in dummy_model_out.parameterization.derived_params()
if p in add[_params]]))
# Save to the collection (keep old weight for now)
collection_out.add(
sampled, derived=derived.values(), weight=point.get(_weight),
logpriors=logpriors_new, loglikes=loglikes_new)
# Display progress
percent = np.round(i / len(collection_in) * 100)
if percent != last_percent and not percent % 5:
last_percent = percent
progress_bar(log, percent, " (%d/%d)" % (i, len(collection_in)))
if not collection_out.data.last_valid_index():
raise LoggedError(
log, "No elements in the final sample. Possible causes: "
"added a prior or likelihood valued zero over the full sampled domain, "
"or the computation of the theory failed everywhere, etc.")
# Reweight -- account for large dynamic range!
# Prefer to rescale +inf to finite, and ignore final points with -inf.
# Remove -inf's (0-weight), and correct indices
difflogmax = max(collection_in[_minuslogpost] - collection_out[_minuslogpost])
collection_out.data[_weight] *= np.exp(
collection_in[_minuslogpost] - collection_out[_minuslogpost] - difflogmax)
collection_out.data = (
collection_out.data[collection_out.data.weight > 0].reset_index(drop=True))
collection_out._n = collection_out.data.last_valid_index() + 1
# Write!
collection_out.out_update()
log.info("Finished! Final number of samples: %d", len(collection_out))
return info_out, {"sample": collection_out}
def run(
info_or_yaml_or_file: Union[InputDict, str, os.PathLike],
packages_path: Optional[str] = None,
output: Union[str, LiteralFalse, None] = None,
debug: Union[bool, int, None] = None,
stop_at_error: Optional[bool] = None,
resume: bool = False,
force: bool = False,
no_mpi: bool = False,
test: bool = False,
override: Optional[InputDict] = None,
) -> Union[InfoSamplerTuple, PostTuple]:
"""
Run from an input dictionary, file name or yaml string, with optional arguments
to override settings in the input as needed.
:param info_or_yaml_or_file: input options dictionary, yaml file, or yaml text
:param packages_path: path where external packages were installed
:param output: path name prefix for output files, or False for no file output
:param debug: true for verbose debug output, or a specific logging level
:param stop_at_error: stop if an error is raised
:param resume: continue an existing run
:param force: overwrite existing output if it exists
:param no_mpi: run without MPI
:param test: only test initialization rather than actually running
:param override: option dictionary to merge into the input one, overriding settings
(but with lower precedence than the explicit keyword arguments)
:return: (updated_info, sampler) tuple of options dictionary and Sampler instance,
or (updated_info, results) if using "post" post-processing
"""
# This function reproduces the model-->output-->sampler pipeline one would follow
# when instantiating by hand, but alters the order to performs checks and dump info
# as early as possible, e.g. to check if resuming possible or `force` needed.
if no_mpi or test:
mpi.set_mpi_disabled()
with mpi.ProcessState("run"):
info: InputDict = load_info_overrides(info_or_yaml_or_file, debug,
stop_at_error, packages_path,
override)
if test:
info["test"] = True
# If any of resume|force given as cmd args, ignore those in the input file
if resume or force:
if resume and force:
raise ValueError("'rename' and 'force' are exclusive options")
info["resume"] = bool(resume)
info["force"] = bool(force)
if info.get("post"):
if isinstance(output, str) or output is False:
info["post"]["output"] = output or None
return post(info)
if isinstance(output, str) or output is False:
info["output"] = output or None
logger_setup(info.get("debug"), info.get("debug_file"))
logger_run = get_logger(run.__name__)
# MARKED FOR DEPRECATION IN v3.0
# BEHAVIOUR TO BE REPLACED BY ERROR:
check_deprecated_modules_path(info)
# END OF DEPRECATION BLOCK
# 1. Prepare output driver, if requested by defining an output_prefix
# GetDist needs to know the original sampler, so don't overwrite if minimizer
try:
which_sampler = list(info["sampler"])[0]
except (KeyError, TypeError):
raise LoggedError(
logger_run,
"You need to specify a sampler using the 'sampler' key "
"as e.g. `sampler: {mcmc: None}.`")
infix = "minimize" if which_sampler == "minimize" else None
with get_output(prefix=info.get("output"),
resume=info.get("resume"),
force=info.get("force"),
infix=infix) as out:
# 2. Update the input info with the defaults for each component
updated_info = update_info(info)
if is_debug(logger_run):
# Dump only if not doing output
# (otherwise, the user can check the .updated file)
if not out and mpi.is_main_process():
logger_run.info(
"Input info updated with defaults (dumped to YAML):\n%s",
yaml_dump(sort_cosmetic(updated_info)))
# 3. If output requested, check compatibility if existing one, and dump.
# 3.1 First: model only
out.check_and_dump_info(info,
updated_info,
cache_old=True,
ignore_blocks=["sampler"])
# 3.2 Then sampler -- 1st get the last sampler mentioned in the updated.yaml
# TODO: ideally, using Minimizer would *append* to the sampler block.
# Some code already in place, but not possible at the moment.
try:
last_sampler = list(updated_info["sampler"])[-1]
last_sampler_info = {
last_sampler: updated_info["sampler"][last_sampler]
}
except (KeyError, TypeError):
raise LoggedError(logger_run, "No sampler requested.")
sampler_name, sampler_class = get_sampler_name_and_class(
last_sampler_info)
check_sampler_info((out.reload_updated_info(use_cache=True)
or {}).get("sampler"),
updated_info["sampler"],
is_resuming=out.is_resuming())
# Dump again, now including sampler info
out.check_and_dump_info(info, updated_info, check_compatible=False)
# Check if resumable run
sampler_class.check_force_resume(
out, info=updated_info["sampler"][sampler_name])
# 4. Initialize the posterior and the sampler
with Model(updated_info["params"],
updated_info["likelihood"],
updated_info.get("prior"),
updated_info.get("theory"),
packages_path=info.get("packages_path"),
timing=updated_info.get("timing"),
allow_renames=False,
stop_at_error=info.get("stop_at_error",
False)) as model:
# Re-dump the updated info, now containing parameter routes and version
updated_info = recursive_update(updated_info, model.info())
out.check_and_dump_info(None,
updated_info,
check_compatible=False)
sampler = sampler_class(
updated_info["sampler"][sampler_name],
model,
out,
name=sampler_name,
packages_path=info.get("packages_path"))
# Re-dump updated info, now also containing updates from the sampler
updated_info["sampler"][sampler_name] = \
recursive_update(updated_info["sampler"][sampler_name],
sampler.info())
out.check_and_dump_info(None,
updated_info,
check_compatible=False)
mpi.sync_processes()
if info.get("test", False):
logger_run.info(
"Test initialization successful! "
"You can probably run now without `--%s`.", "test")
return InfoSamplerTuple(updated_info, sampler)
# Run the sampler
sampler.run()
return InfoSamplerTuple(updated_info, sampler)
def run_script():
warn_deprecation()
import os
import argparse
parser = argparse.ArgumentParser(description="Cobaya's run script.")
parser.add_argument("input_file",
nargs=1,
action="store",
metavar="input_file.yaml",
help="An input file to run.")
parser.add_argument("-" + _packages_path_arg[0],
"--" + _packages_path_arg_posix,
action="store",
nargs=1,
metavar="/packages/path",
default=[None],
help="Path where external packages were installed.")
# MARKED FOR DEPRECATION IN v3.0
modules = "modules"
parser.add_argument("-" + modules[0],
"--" + modules,
action="store",
nargs=1,
required=False,
metavar="/packages/path",
default=[None],
help="To be deprecated! "
"Alias for %s, which should be used instead." %
_packages_path_arg_posix)
# END OF DEPRECATION BLOCK -- CONTINUES BELOW!
parser.add_argument("-" + _output_prefix[0],
"--" + _output_prefix,
action="store",
nargs=1,
metavar="/some/path",
default=[None],
help="Path and prefix for the text output.")
parser.add_argument("-" + _debug[0],
"--" + _debug,
action="store_true",
help="Produce verbose debug output.")
continuation = parser.add_mutually_exclusive_group(required=False)
continuation.add_argument(
"-" + _resume[0],
"--" + _resume,
action="store_true",
help="Resume an existing chain if it has similar info "
"(fails otherwise).")
continuation.add_argument("-" + _force[0],
"--" + _force,
action="store_true",
help="Overwrites previous output, if it exists "
"(use with care!)")
parser.add_argument("--%s" % _test_run,
action="store_true",
help="Initialize model and sampler, and exit.")
parser.add_argument("--version", action="version", version=__version__)
parser.add_argument("--no-mpi",
action='store_true',
help="disable MPI when mpi4py installed but MPI does "
"not actually work")
arguments = parser.parse_args()
if arguments.no_mpi or getattr(arguments, _test_run, False):
set_mpi_disabled()
if any((os.path.splitext(f)[0] in ("input", "updated"))
for f in arguments.input_file):
raise ValueError("'input' and 'updated' are reserved file names. "
"Please, use a different one.")
load_input = import_MPI(".input", "load_input")
given_input = arguments.input_file[0]
if any(given_input.lower().endswith(ext) for ext in _yaml_extensions):
info = load_input(given_input)
output_prefix_cmd = getattr(arguments, _output_prefix)[0]
output_prefix_input = info.get(_output_prefix)
info[_output_prefix] = output_prefix_cmd or output_prefix_input
else:
# Passed an existing output_prefix? Try to find the corresponding *.updated.yaml
updated_file = get_info_path(*split_prefix(given_input),
kind="updated")
try:
info = load_input(updated_file)
except IOError:
raise ValueError(
"Not a valid input file, or non-existent run to resume")
# We need to update the output_prefix to resume the run *where it is*
info[_output_prefix] = given_input
# If input given this way, we obviously want to resume!
info[_resume] = True
# solve packages installation path cmd > env > input
# MARKED FOR DEPRECATION IN v3.0
if getattr(arguments, modules) != [None]:
logger_setup()
logger = logging.getLogger(__name__.split(".")[-1])
logger.warning(
"*DEPRECATION*: -m/--modules will be deprecated in favor of "
"-%s/--%s in the next version. Please, use that one instead.",
_packages_path_arg[0], _packages_path_arg_posix)
# BEHAVIOUR TO BE REPLACED BY ERROR:
if getattr(arguments, _packages_path_arg) == [None]:
setattr(arguments, _packages_path_arg, getattr(arguments, modules))
# BEHAVIOUR TO BE REPLACED BY ERROR:
check_deprecated_modules_path(info)
# END OF DEPRECATION BLOCK
info[_packages_path] = \
getattr(arguments, _packages_path_arg)[0] or info.get(_packages_path)
info[_debug] = getattr(arguments, _debug) or info.get(
_debug, _debug_default)
info[_test_run] = getattr(arguments, _test_run, False)
# If any of resume|force given as cmd args, ignore those in the input file
resume_arg, force_arg = [
getattr(arguments, arg) for arg in [_resume, _force]
]
if any([resume_arg, force_arg]):
info[_resume], info[_force] = resume_arg, force_arg
if _post in info:
post(info)
else:
run(info)
def run(info):
# This function reproduces the model-->output-->sampler pipeline one would follow
# when instantiating by hand, but alters the order to performs checks and dump info
# as early as possible, e.g. to check if resuming possible or `force` needed.
assert isinstance(info, Mapping), (
"The first argument must be a dictionary with the info needed for the run. "
"If you were trying to pass the name of an input file instead, "
"load it first with 'cobaya.input.load_input', "
"or, if you were passing a yaml string, load it with 'cobaya.yaml.yaml_load'."
)
logger_setup(info.get(_debug), info.get(_debug_file))
logger_run = logging.getLogger(__name__.split(".")[-1])
# MARKED FOR DEPRECATION IN v3.0
# BEHAVIOUR TO BE REPLACED BY ERROR:
check_deprecated_modules_path(info)
# END OF DEPRECATION BLOCK
# 1. Prepare output driver, if requested by defining an output_prefix
output = get_output(output_prefix=info.get(_output_prefix),
resume=info.get(_resume),
force=info.get(_force))
# 2. Update the input info with the defaults for each component
updated_info = update_info(info)
if logging.root.getEffectiveLevel() <= logging.DEBUG:
# Dump only if not doing output (otherwise, the user can check the .updated file)
if not output and is_main_process():
logger_run.info(
"Input info updated with defaults (dumped to YAML):\n%s",
yaml_dump(sort_cosmetic(updated_info)))
# 3. If output requested, check compatibility if existing one, and dump.
# 3.1 First: model only
output.check_and_dump_info(info,
updated_info,
cache_old=True,
ignore_blocks=[kinds.sampler])
# 3.2 Then sampler -- 1st get the last sampler mentioned in the updated.yaml
# TODO: ideally, using Minimizer would *append* to the sampler block.
# Some code already in place, but not possible at the moment.
try:
last_sampler = list(updated_info[kinds.sampler])[-1]
last_sampler_info = {
last_sampler: updated_info[kinds.sampler][last_sampler]
}
except (KeyError, TypeError):
raise LoggedError(logger_run, "No sampler requested.")
sampler_name, sampler_class = get_sampler_name_and_class(last_sampler_info)
check_sampler_info((output.reload_updated_info(use_cache=True)
or {}).get(kinds.sampler),
updated_info[kinds.sampler],
is_resuming=output.is_resuming())
# Dump again, now including sampler info
output.check_and_dump_info(info, updated_info, check_compatible=False)
# Check if resumable run
sampler_class.check_force_resume(
output, info=updated_info[kinds.sampler][sampler_name])
# 4. Initialize the posterior and the sampler
with Model(updated_info[_params], updated_info[kinds.likelihood],
updated_info.get(_prior), updated_info.get(kinds.theory),
packages_path=info.get(_packages_path), timing=updated_info.get(_timing),
allow_renames=False, stop_at_error=info.get("stop_at_error", False)) \
as model:
# Re-dump the updated info, now containing parameter routes and version info
updated_info = recursive_update(updated_info, model.info())
output.check_and_dump_info(None, updated_info, check_compatible=False)
sampler = sampler_class(updated_info[kinds.sampler][sampler_name],
model,
output,
packages_path=info.get(_packages_path))
# Re-dump updated info, now also containing updates from the sampler
updated_info[kinds.sampler][sampler.get_name()] = \
recursive_update(
updated_info[kinds.sampler][sampler.get_name()], sampler.info())
# TODO -- maybe also re-dump model info, now possibly with measured speeds
# (waiting until the camb.transfers issue is solved)
output.check_and_dump_info(None, updated_info, check_compatible=False)
if info.get(_test_run, False):
logger_run.info(
"Test initialization successful! "
"You can probably run now without `--%s`.", _test_run)
return updated_info, sampler
# Run the sampler
sampler.run()
return updated_info, sampler
def post(
info_or_yaml_or_file: Union[InputDict, str, os.PathLike],
sample: Union[SampleCollection, List[SampleCollection], None] = None
) -> PostTuple:
info = load_input_dict(info_or_yaml_or_file)
logger_setup(info.get("debug"), info.get("debug_file"))
log = get_logger(__name__)
# MARKED FOR DEPRECATION IN v3.0
# BEHAVIOUR TO BE REPLACED BY ERROR:
check_deprecated_modules_path(info)
# END OF DEPRECATION BLOCK
info_post: PostDict = info.get("post") or {}
if not info_post:
raise LoggedError(log, "No 'post' block given. Nothing to do!")
if mpi.is_main_process() and info.get("resume"):
log.warning(
"Resuming not implemented for post-processing. Re-starting.")
if not info.get("output") and info_post.get("output") \
and not info.get("params"):
raise LoggedError(
log, "The input dictionary must have be a full option "
"dictionary, or have an existing 'output' root to load "
"previous settings from ('output' to read from is in the "
"main block not under 'post'). ")
# 1. Load existing sample
output_in = get_output(prefix=info.get("output"))
if output_in:
info_in = output_in.load_updated_info() or update_info(info)
else:
info_in = update_info(info)
params_in: ExpandedParamsDict = info_in["params"] # type: ignore
dummy_model_in = DummyModel(params_in, info_in.get("likelihood", {}),
info_in.get("prior"))
in_collections = []
thin = info_post.get("thin", 1)
skip = info_post.get("skip", 0)
if info.get('thin') is not None or info.get(
'skip') is not None: # type: ignore
raise LoggedError(
log, "'thin' and 'skip' should be "
"parameters of the 'post' block")
if sample:
# If MPI, assume for each MPI process post is passed in the list of
# collections that should be processed by that process
# (e.g. single chain output from sampler)
if isinstance(sample, SampleCollection):
in_collections = [sample]
else:
in_collections = sample
for i, collection in enumerate(in_collections):
if skip:
if 0 < skip < 1:
skip = int(round(skip * len(collection)))
collection = collection.filtered_copy(slice(skip, None))
if thin != 1:
collection = collection.thin_samples(thin)
in_collections[i] = collection
elif output_in:
files = output_in.find_collections()
numbered = files
if not numbered:
# look for un-numbered output files
files = output_in.find_collections(name=False)
if files:
if mpi.size() > len(files):
raise LoggedError(
log, "Number of MPI processes (%s) is larger than "
"the number of sample files (%s)", mpi.size(), len(files))
for num in range(mpi.rank(), len(files), mpi.size()):
in_collections += [
SampleCollection(dummy_model_in,
output_in,
onload_thin=thin,
onload_skip=skip,
load=True,
file_name=files[num],
name=str(num + 1) if numbered else "")
]
else:
raise LoggedError(
log, "No samples found for the input model with prefix %s",
os.path.join(output_in.folder, output_in.prefix))
else:
raise LoggedError(
log, "No output from where to load from, "
"nor input collections given.")
if any(len(c) <= 1 for c in in_collections):
raise LoggedError(
log,
"Not enough samples for post-processing. Try using a larger sample, "
"or skipping or thinning less.")
mpi.sync_processes()
log.info("Will process %d sample points.",
sum(len(c) for c in in_collections))
# 2. Compare old and new info: determine what to do
add = info_post.get("add") or {}
if "remove" in add:
raise LoggedError(log,
"remove block should be under 'post', not 'add'")
remove = info_post.get("remove") or {}
# Add a dummy 'one' likelihood, to absorb unused parameters
if not add.get("likelihood"):
add["likelihood"] = {}
add["likelihood"]["one"] = None
# Expand the "add" info, but don't add new default sampled parameters
orig_params = set(add.get("params") or [])
add = update_info(add, add_aggr_chi2=False)
add_params: ExpandedParamsDict = add["params"] # type: ignore
for p in set(add_params) - orig_params:
if p in params_in:
add_params.pop(p)
# 2.1 Adding/removing derived parameters and changes in priors of sampled parameters
out_combined_params = deepcopy_where_possible(params_in)
remove_params = list(str_to_list(remove.get("params")) or [])
for p in remove_params:
pinfo = params_in.get(p)
if pinfo is None or not is_derived_param(pinfo):
raise LoggedError(
log,
"You tried to remove parameter '%s', which is not a derived parameter. "
"Only derived parameters can be removed during post-processing.",
p)
out_combined_params.pop(p)
# Force recomputation of aggregated chi2
for p in list(out_combined_params):
if p.startswith(get_chi2_name("")):
out_combined_params.pop(p)
prior_recompute_1d = False
for p, pinfo in add_params.items():
pinfo_in = params_in.get(p)
if is_sampled_param(pinfo):
if not is_sampled_param(pinfo_in):
# No added sampled parameters (de-marginalisation not implemented)
if pinfo_in is None:
raise LoggedError(
log,
"You added a new sampled parameter %r (maybe accidentally "
"by adding a new likelihood that depends on it). "
"Adding new sampled parameters is not possible. Try fixing "
"it to some value.", p)
else:
raise LoggedError(
log,
"You tried to change the prior of parameter '%s', "
"but it was not a sampled parameter. "
"To change that prior, you need to define as an external one.",
p)
# recompute prior if potentially changed sampled parameter priors
prior_recompute_1d = True
elif is_derived_param(pinfo):
if p in out_combined_params:
raise LoggedError(
log,
"You tried to add derived parameter '%s', which is already "
"present. To force its recomputation, 'remove' it too.", p)
elif is_fixed_or_function_param(pinfo):
# Only one possibility left "fixed" parameter that was not present before:
# input of new likelihood, or just an argument for dynamical derived (dropped)
if pinfo_in and p in params_in and pinfo["value"] != pinfo_in.get(
"value"):
raise LoggedError(
log,
"You tried to add a fixed parameter '%s: %r' that was already present"
" but had a different value or was not fixed. This is not allowed. "
"The old info of the parameter was '%s: %r'", p,
dict(pinfo), p, dict(pinfo_in))
elif not pinfo_in: # OK as long as we have known value for it
raise LoggedError(log, "Parameter %s no known value. ", p)
out_combined_params[p] = pinfo
out_combined: InputDict = {"params": out_combined_params} # type: ignore
# Turn the rest of *derived* parameters into constants,
# so that the likelihoods do not try to recompute them
# But be careful to exclude *input* params that have a "derived: True" value
# (which in "updated info" turns into "derived: 'lambda [x]: [x]'")
# Don't assign to derived parameters to theories, only likelihoods, so they can be
# recomputed if needed. If the theory does not need to be computed, it doesn't matter
# if it is already assigned parameters in the usual way; likelihoods can get
# the required derived parameters from the stored sample derived parameter inputs.
out_params_with_computed = deepcopy_where_possible(out_combined_params)
dropped_theory = set()
for p, pinfo in out_params_with_computed.items():
if (is_derived_param(pinfo) and "value" not in pinfo
and p not in add_params):
out_params_with_computed[p] = {"value": np.nan}
dropped_theory.add(p)
# 2.2 Manage adding/removing priors and likelihoods
warn_remove = False
kind: ModelBlock
for kind in ("prior", "likelihood", "theory"):
out_combined[kind] = deepcopy_where_possible(info_in.get(kind)) or {}
for remove_item in str_to_list(remove.get(kind)) or []:
try:
out_combined[kind].pop(remove_item, None)
if remove_item not in (add.get(kind)
or []) and kind != "theory":
warn_remove = True
except ValueError:
raise LoggedError(
log, "Trying to remove %s '%s', but it is not present. "
"Existing ones: %r", kind, remove_item,
list(out_combined[kind]))
if kind != "theory" and kind in add:
dups = set(add.get(kind) or []).intersection(
out_combined[kind]) - {"one"}
if dups:
raise LoggedError(
log,
"You have added %s '%s', which was already present. If you "
"want to force its recomputation, you must also 'remove' it.",
kind, dups)
out_combined[kind].update(add[kind])
if warn_remove and mpi.is_main_process():
log.warning("You are removing a prior or likelihood pdf. "
"Notice that if the resulting posterior is much wider "
"than the original one, or displaced enough, "
"it is probably safer to explore it directly.")
mlprior_names_add = minuslogprior_names(add.get("prior") or [])
chi2_names_add = [
get_chi2_name(name) for name in add["likelihood"] if name != "one"
]
out_combined["likelihood"].pop("one", None)
add_theory = add.get("theory")
if add_theory:
if len(add["likelihood"]) == 1 and not any(
is_derived_param(pinfo) for pinfo in add_params.values()):
log.warning(
"You are adding a theory, but this does not force recomputation "
"of any likelihood or derived parameters unless explicitly "
"removed+added.")
# Inherit from the original chain (input|output_params, renames, etc)
added_theory = add_theory.copy()
for theory, theory_info in out_combined["theory"].items():
if theory in list(added_theory):
out_combined["theory"][theory] = \
recursive_update(theory_info, added_theory.pop(theory))
out_combined["theory"].update(added_theory)
# Prepare recomputation of aggregated chi2
# (they need to be recomputed by hand, because auto-computation won't pick up
# old likelihoods for a given type)
all_types = {
like: str_to_list(opts.get("type") or [])
for like, opts in out_combined["likelihood"].items()
}
types = set(chain(*all_types.values()))
inv_types = {
t: [like for like, like_types in all_types.items() if t in like_types]
for t in sorted(types)
}
add_aggregated_chi2_params(out_combined_params, types)
# 3. Create output collection
# Use default prefix if it exists. If it does not, produce no output by default.
# {post: {output: None}} suppresses output, and if it's a string, updates it.
out_prefix = info_post.get("output", info.get("output"))
if out_prefix:
suffix = info_post.get("suffix")
if not suffix:
raise LoggedError(
log, "You need to provide a '%s' for your output chains.",
"suffix")
out_prefix += separator_files + "post" + separator_files + suffix
output_out = get_output(prefix=out_prefix, force=info.get("force"))
output_out.set_lock()
if output_out and not output_out.force and output_out.find_collections():
raise LoggedError(
log, "Found existing post-processing output with prefix %r. "
"Delete it manually or re-run with `force: True` "
"(or `-f`, `--force` from the shell).", out_prefix)
elif output_out and output_out.force and mpi.is_main_process():
output_out.delete_infos()
for _file in output_out.find_collections():
output_out.delete_file_or_folder(_file)
info_out = deepcopy_where_possible(info)
info_post = info_post.copy()
info_out["post"] = info_post
# Updated with input info and extended (updated) add info
info_out.update(info_in) # type: ignore
info_post["add"] = add
dummy_model_out = DummyModel(out_combined_params,
out_combined["likelihood"],
info_prior=out_combined["prior"])
out_func_parameterization = Parameterization(out_params_with_computed)
# TODO: check allow_renames=False?
model_add = Model(out_params_with_computed,
add["likelihood"],
info_prior=add.get("prior"),
info_theory=out_combined["theory"],
packages_path=(info_post.get("packages_path")
or info.get("packages_path")),
allow_renames=False,
post=True,
stop_at_error=info.get('stop_at_error', False),
skip_unused_theories=True,
dropped_theory_params=dropped_theory)
# Remove auxiliary "one" before dumping -- 'add' *is* info_out["post"]["add"]
add["likelihood"].pop("one")
out_collections = [
SampleCollection(dummy_model_out,
output_out,
name=c.name,
cache_size=OutputOptions.default_post_cache_size)
for c in in_collections
]
# TODO: should maybe add skip/thin to out_combined, so can tell post-processed?
output_out.check_and_dump_info(info_out,
out_combined,
check_compatible=False)
collection_in = in_collections[0]
collection_out = out_collections[0]
last_percent = None
known_constants = dummy_model_out.parameterization.constant_params()
known_constants.update(dummy_model_in.parameterization.constant_params())
missing_params = dummy_model_in.parameterization.sampled_params().keys(
) - set(collection_in.columns)
if missing_params:
raise LoggedError(
log, "Input samples do not contain expected sampled parameter "
"values: %s", missing_params)
missing_priors = set(
name for name in collection_out.minuslogprior_names
if name not in mlprior_names_add and name not in collection_in.columns)
if _minuslogprior_1d_name in missing_priors:
prior_recompute_1d = True
if prior_recompute_1d:
missing_priors.discard(_minuslogprior_1d_name)
mlprior_names_add.insert(0, _minuslogprior_1d_name)
prior_regenerate: Optional[Prior]
if missing_priors and "prior" in info_in:
# in case there are input priors that are not stored in input samples
# e.g. when postprocessing GetDist/CosmoMC-format chains
in_names = minuslogprior_names(info_in["prior"])
info_prior = {
piname: inf
for (piname,
inf), in_name in zip(info_in["prior"].items(), in_names)
if in_name in missing_priors
}
regenerated_prior_names = minuslogprior_names(info_prior)
missing_priors.difference_update(regenerated_prior_names)
prior_regenerate = Prior(dummy_model_in.parameterization, info_prior)
else:
prior_regenerate = None
regenerated_prior_names = None
if missing_priors:
raise LoggedError(log, "Missing priors: %s", missing_priors)
mpi.sync_processes()
output_in.check_lock()
# 4. Main loop! Loop over input samples and adjust as required.
if mpi.is_main_process():
log.info("Running post-processing...")
difflogmax: Optional[float] = None
to_do = sum(len(c) for c in in_collections)
weights = []
done = 0
last_dump_time = time.time()
for collection_in, collection_out in zip(in_collections, out_collections):
importance_weights = []
def set_difflogmax():
nonlocal difflogmax
difflog = (collection_in[
OutPar.minuslogpost].to_numpy()[:len(collection_out)] -
collection_out[OutPar.minuslogpost].to_numpy())
difflogmax = np.max(difflog)
if abs(difflogmax) < 1:
difflogmax = 0 # keep simple when e.g. very similar
log.debug("difflogmax: %g", difflogmax)
if mpi.more_than_one_process():
difflogmax = max(mpi.allgather(difflogmax))
if mpi.is_main_process():
log.debug("Set difflogmax: %g", difflogmax)
_weights = np.exp(difflog - difflogmax)
importance_weights.extend(_weights)
collection_out.reweight(_weights)
for i, point in collection_in.data.iterrows():
all_params = point.to_dict()
for p in remove_params:
all_params.pop(p, None)
log.debug("Point: %r", point)
sampled = np.array([
all_params[param]
for param in dummy_model_in.parameterization.sampled_params()
])
all_params = out_func_parameterization.to_input(all_params).copy()
# Add/remove priors
if prior_recompute_1d:
priors_add = [model_add.prior.logps_internal(sampled)]
if priors_add[0] == -np.inf:
continue
else:
priors_add = []
if model_add.prior.external:
priors_add.extend(model_add.prior.logps_external(all_params))
logpriors_add = dict(zip(mlprior_names_add, priors_add))
logpriors_new = [
logpriors_add.get(name, -point.get(name, 0))
for name in collection_out.minuslogprior_names
]
if prior_regenerate:
regenerated = dict(
zip(regenerated_prior_names,
prior_regenerate.logps_external(all_params)))
for _i, name in enumerate(collection_out.minuslogprior_names):
if name in regenerated_prior_names:
logpriors_new[_i] = regenerated[name]
if is_debug(log):
log.debug("New set of priors: %r",
dict(zip(dummy_model_out.prior, logpriors_new)))
if -np.inf in logpriors_new:
continue
# Add/remove likelihoods and/or (re-)calculate derived parameters
loglikes_add, output_derived = model_add.logps(all_params)
loglikes_add = dict(zip(chi2_names_add, loglikes_add))
output_derived = dict(zip(model_add.output_params, output_derived))
loglikes_new = [
loglikes_add.get(name, -0.5 * point.get(name, 0))
for name in collection_out.chi2_names
]
if is_debug(log):
log.debug("New set of likelihoods: %r",
dict(zip(dummy_model_out.likelihood, loglikes_new)))
if output_derived:
log.debug("New set of derived parameters: %r",
output_derived)
if -np.inf in loglikes_new:
continue
all_params.update(output_derived)
all_params.update(out_func_parameterization.to_derived(all_params))
derived = {
param: all_params.get(param)
for param in dummy_model_out.parameterization.derived_params()
}
# We need to recompute the aggregated chi2 by hand
for type_, likes in inv_types.items():
derived[get_chi2_name(type_)] = sum(
-2 * lvalue for lname, lvalue in zip(
collection_out.chi2_names, loglikes_new)
if undo_chi2_name(lname) in likes)
if is_debug(log):
log.debug(
"New derived parameters: %r", {
p: derived[p]
for p in dummy_model_out.parameterization.
derived_params() if p in add["params"]
})
# Save to the collection (keep old weight for now)
weight = point.get(OutPar.weight)
mpi.check_errors()
if difflogmax is None and i > OutputOptions.reweight_after and \
time.time() - last_dump_time > OutputOptions.output_inteveral_s / 2:
set_difflogmax()
collection_out.out_update()
if difflogmax is not None:
logpost_new = sum(logpriors_new) + sum(loglikes_new)
importance_weight = np.exp(logpost_new +
point.get(OutPar.minuslogpost) -
difflogmax)
weight = weight * importance_weight
importance_weights.append(importance_weight)
if time.time(
) - last_dump_time > OutputOptions.output_inteveral_s:
collection_out.out_update()
last_dump_time = time.time()
if weight > 0:
collection_out.add(sampled,
derived=derived.values(),
weight=weight,
logpriors=logpriors_new,
loglikes=loglikes_new)
# Display progress
percent = int(np.round((i + done) / to_do * 100))
if percent != last_percent and not percent % 5:
last_percent = percent
progress_bar(log, percent, " (%d/%d)" % (i + done, to_do))
if difflogmax is None:
set_difflogmax()
if not collection_out.data.last_valid_index():
raise LoggedError(
log, "No elements in the final sample. Possible causes: "
"added a prior or likelihood valued zero over the full sampled "
"domain, or the computation of the theory failed everywhere, etc."
)
collection_out.out_update()
weights.append(np.array(importance_weights))
done += len(collection_in)
assert difflogmax is not None
points = 0
tot_weight = 0
min_weight = np.inf
max_weight = -np.inf
max_output_weight = -np.inf
sum_w2 = 0
points_removed = 0
for collection_in, collection_out, importance_weights in zip(
in_collections, out_collections, weights):
output_weights = collection_out[OutPar.weight]
points += len(collection_out)
tot_weight += np.sum(output_weights)
points_removed += len(importance_weights) - len(output_weights)
min_weight = min(min_weight, np.min(importance_weights))
max_weight = max(max_weight, np.max(importance_weights))
max_output_weight = max(max_output_weight, np.max(output_weights))
sum_w2 += np.dot(output_weights, output_weights)
(tot_weights, min_weights, max_weights, max_output_weights, sum_w2s,
points_s, points_removed_s) = mpi.zip_gather([
tot_weight, min_weight, max_weight, max_output_weight, sum_w2, points,
points_removed
])
if mpi.is_main_process():
output_out.clear_lock()
log.info("Finished! Final number of distinct sample points: %s",
sum(points_s))
log.info("Importance weight range: %.4g -- %.4g", min(min_weights),
max(max_weights))
if sum(points_removed_s):
log.info("Points deleted due to zero weight: %s",
sum(points_removed_s))
log.info(
"Effective number of single samples if independent (sum w)/max(w): %s",
int(sum(tot_weights) / max(max_output_weights)))
log.info(
"Effective number of weighted samples if independent (sum w)^2/sum(w^2): "
"%s", int(sum(tot_weights)**2 / sum(sum_w2s)))
products: PostResultDict = {
"sample": value_or_list(out_collections),
"stats": {
'min_importance_weight': (min(min_weights) / max(max_weights)),
'points_removed': sum(points_removed_s),
'tot_weight': sum(tot_weights),
'max_weight': max(max_output_weights),
'sum_w2': sum(sum_w2s),
'points': sum(points_s)
},
"logpost_weight_offset": difflogmax,
"weights": value_or_list(weights)
}
return PostTuple(info=out_combined, products=products)