csiborgtools/scripts/field_sample.py
2023-12-13 20:42:58 +00:00

225 lines
8.5 KiB
Python

# Copyright (C) 2023 Richard Stiskalek
# This program is free software; you can redistribute it and/or modify it
# under the terms of the GNU General Public License as published by the
# Free Software Foundation; either version 3 of the License, or (at your
# option) any later version.
#
# This program is distributed in the hope that it will be useful, but
# WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General
# Public License for more details.
#
# You should have received a copy of the GNU General Public License along
# with this program; if not, write to the Free Software Foundation, Inc.,
# 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
"""
Sample a CSiBORG field at galaxy positions and save the result to disk.
"""
from argparse import ArgumentParser
from distutils.util import strtobool
from os.path import join
import numpy
from astropy.cosmology import FlatLambdaCDM
from h5py import File
from mpi4py import MPI
from taskmaster import work_delegation
from tqdm import tqdm
import csiborgtools
from utils import get_nsims
MPC2BOX = 1 / 677.7
def steps(cls, survey_name):
"""Make a list of selection criteria to apply to a survey."""
if survey_name == "SDSS":
return [
# (lambda x: cls[x], ("IN_DR7_LSS",)),
# (lambda x: cls[x] < 17.6, ("ELPETRO_APPMAG_r", )),
(lambda x: cls[x] < 155.5, ("DIST", ))
]
else:
raise NotImplementedError(f"Survey `{survey_name}` not implemented.")
def open_galaxy_positions(survey_name, comm):
"""
Load the survey galaxy positions and indices, broadcasting them to all
ranks.
"""
rank, size = comm.Get_rank(), comm.Get_size()
if rank == 0:
if survey_name == "SDSS":
survey = csiborgtools.SDSS()()
pos = numpy.vstack([survey["DIST_UNCORRECTED"],
survey["RA"],
survey["DEC"]],
).T
pos = pos.astype(numpy.float32)
indxs = survey["INDEX"]
if survey_name == "SDSSxALFALFA":
survey = csiborgtools.SDSSxALFALFA()()
pos = numpy.vstack([survey["DIST_UNCORRECTED"],
survey["RA_1"],
survey["DEC_1"]],
).T
pos = pos.astype(numpy.float32)
indxs = survey["INDEX"]
elif survey_name == "GW170817":
samples = File("/mnt/extraspace/rstiskalek/GWLSS/H1L1V1-EXTRACT_POSTERIOR_GW170817-1187008600-400.hdf", 'r')["samples"] # noqa
cosmo = FlatLambdaCDM(H0=100, Om0=0.3175)
pos = numpy.vstack([
cosmo.comoving_distance(samples["redshift"][:]).value,
samples["ra"][:] * 180 / numpy.pi,
samples["dec"][:] * 180 / numpy.pi],
).T
indxs = numpy.arange(pos.shape[0])
else:
raise NotImplementedError(f"Survey `{survey_name}` not "
"implemented.")
else:
pos = None
indxs = None
comm.Barrier()
if size > 1:
pos = comm.bcast(pos, root=0)
indxs = comm.bcast(indxs, root=0)
return pos, indxs
def evaluate_field(field, pos, nrand, smooth_scales=None, seed=42,
verbose=True):
"""
Evaluate the field at the given sky positions. Additionally, evaluate the
field at `nrand` random positions.
"""
if smooth_scales is None:
smooth_scales = [0.]
nsample = pos.shape[0]
nsmooth = len(smooth_scales)
val = numpy.full((nsample, nsmooth), numpy.nan, dtype=field.dtype)
if nrand > 0:
rand_val = numpy.full((nsample, nsmooth, nrand), numpy.nan,
dtype=field.dtype)
else:
rand_val = None
for i, scale in enumerate(tqdm(smooth_scales, desc="Smoothing",
disable=not verbose)):
if scale > 0:
field_smoothed = csiborgtools.field.smoothen_field(
field, scale * MPC2BOX, boxsize=1, make_copy=True)
else:
field_smoothed = numpy.copy(field)
val[:, i] = csiborgtools.field.evaluate_sky(
field_smoothed, pos=pos, mpc2box=MPC2BOX)
if nrand == 0:
continue
for j in range(nrand):
gen = numpy.random.default_rng(seed + j)
pos_rand = numpy.vstack([
gen.permutation(pos[:, 0]),
gen.uniform(0, 360, nsample),
90 - numpy.rad2deg(numpy.arccos(gen.uniform(-1, 1, nsample))),
]).T
rand_val[:, i, j] = csiborgtools.field.evaluate_sky(
field_smoothed, pos=pos_rand, mpc2box=MPC2BOX)
return val, rand_val, smooth_scales
def match_to_no_selection(val, rand_val, parser_args):
if parser_args.survey == "SDSSxALFALFA":
survey = csiborgtools.SDSSxALFALFA()(apply_selection=False)
else:
raise NotImplementedError(
f"Survey `{parser_args.survey}` not implemented for matching to no selection.") # noqa
if val is not None:
val = csiborgtools.read.match_array_to_no_masking(val, survey)
if rand_val is not None:
rand_val = csiborgtools.read.match_array_to_no_masking(rand_val,
survey)
return val, rand_val
def main(nsim, parser_args, pos, indxs, paths, verbose):
"""Load the field, interpolate it and save it to disk."""
fpath_field = paths.field(parser_args.kind, parser_args.MAS,
parser_args.grid, nsim, parser_args.in_rsp)
field = numpy.load(fpath_field)
val, rand_val, smooth_scales = evaluate_field(
field, pos, nrand=parser_args.nrand,
smooth_scales=parser_args.smooth_scales, verbose=verbose)
if parser_args.survey == "GW170817":
kind = parser_args.kind
kind = kind + "_rsp" if parser_args.in_rsp else kind
fout = join(
"/mnt/extraspace/rstiskalek/GWLSS/",
f"{kind}_{parser_args.MAS}_{parser_args.grid}_{nsim}_H1L1V1-EXTRACT_POSTERIOR_GW170817-1187008600-400.npz") # noqa
else:
fout = paths.field_interpolated(parser_args.survey, parser_args.kind,
parser_args.MAS, parser_args.grid,
nsim, parser_args.in_rsp)
# The survey above had some cuts, however for compatibility we want
# the same shape as the `uncut` survey
val, rand_val = match_to_no_selection(val, rand_val, parser_args)
if verbose:
print(f"Saving to ... `{fout}`.")
numpy.savez(fout, val=val, rand_val=rand_val, indxs=indxs,
smooth_scales=smooth_scales)
if __name__ == "__main__":
parser = ArgumentParser()
parser.add_argument("--nsims", type=int, nargs="+", default=None,
help="IC realisations. If `-1` processes all.")
parser.add_argument("--survey", type=str, required=True,
choices=["SDSS", "SDSSxALFALFA", "GW170817"],
help="Galaxy survey")
parser.add_argument("--smooth_scales", type=float, nargs="+", default=None,
help="Smoothing scales in Mpc / h.")
parser.add_argument("--kind", type=str,
choices=["density", "rspdensity", "velocity", "radvel",
"potential"],
help="What field to interpolate.")
parser.add_argument("--MAS", type=str,
choices=["NGP", "CIC", "TSC", "PCS"],
help="Mass assignment scheme.")
parser.add_argument("--grid", type=int, help="Grid resolution.")
parser.add_argument("--in_rsp", type=lambda x: bool(strtobool(x)),
help="Field in RSP?")
parser.add_argument("--nrand", type=int, required=True,
help="Number of rand. positions to evaluate the field")
parser.add_argument("--simname", type=str, default="csiborg1",
choices=["csiborg1"], help="Simulation name")
args = parser.parse_args()
paths = csiborgtools.read.Paths(**csiborgtools.paths_glamdring)
nsims = get_nsims(args, paths)
pos, indxs = open_galaxy_positions(args.survey, MPI.COMM_WORLD)
def _main(nsim):
main(nsim, args, pos, indxs, paths,
verbose=MPI.COMM_WORLD.Get_size() == 1)
work_delegation(_main, nsims, MPI.COMM_WORLD)