Informativity of IRAM 30m EMIR observations

import os
import sys
import yaml

sys.path.insert(0, os.path.join(".."))

import matplotlib.pyplot as plt

from infobs.informativity import Infobs
from infobs.instruments import IRAM30mEMIR
from infobs.graphics import Settings
from infobs.sampling import Constant, LogUniform

Settings.rotational_only = True

transitions = True  # Set to False to hide line transitions

Instrument

linewidth = 10  # km/s

instrument = IRAM30mEMIR(linewidth)

Simulator

n_samples = 10_000

samplers = {
    "Av": LogUniform(1, 25),
    "G0": LogUniform(1e1, 1e3),
    "Pth": LogUniform(1e5, 5e6),
    "angle": Constant(0.0),
}

with open(os.path.join("informativity", "regimes.yaml"), "r") as f:
    regimes = yaml.safe_load(f)

print("Physical regimes")
for name, desc in regimes.items():
    print(f"{name}: {desc}")

Physical regimes
all: {'Av': [None, None]}
diffuse: {'Av': [None, 3]}
translucent: {'Av': [3, 6]}
filamentary: {'Av': [6, 12]}
dense: {'Av': [12, None]}

Studied lines

instrument.available_molecules()

['c_c3h2',
 'so',
 'c2h',
 'hcn',
 'hcop',
 'hnc',
 'cs',
 '13c_18o',
 'c_18o',
 '13c_o',
 'cn',
 'co',
 'h2o',
 'h2_18o',
 'n']

lines = instrument.available_lines(["co", "13c_o", "c_18o"])
lines.sort()
lines

['13c_o_j1__j0',
 '13c_o_j2__j1',
 '13c_o_j3__j2',
 'c_18o_j1__j0',
 'c_18o_j2__j1',
 'c_18o_j3__j2',
 'co_v0_j1__v0_j0',
 'co_v0_j2__v0_j1',
 'co_v0_j3__v0_j2',
 'co_v1_j1__v1_j0',
 'co_v1_j2__v1_j1',
 'co_v1_j3__v1_j2']

lines = [line for line in lines if "v1" not in line]
lines

['13c_o_j1__j0',
 '13c_o_j2__j1',
 '13c_o_j3__j2',
 'c_18o_j1__j0',
 'c_18o_j2__j1',
 'c_18o_j3__j2',
 'co_v0_j1__v0_j0',
 'co_v0_j2__v0_j1',
 'co_v0_j3__v0_j2']

Definition of instances of Infobs

Reference use case

infobs = Infobs(instrument, n_samples, samplers, obstime=30)

infobs.set_path(os.path.join("informativity", "data"))

infobs.set_regimes(regimes)

Deep integration

infobs_integ = Infobs(instrument, n_samples, samplers, obstime=300)

infobs_integ.set_path(os.path.join("informativity", "data-integration"))

infobs_integ.set_regimes(regimes)

Excellent observing conditions

infobs_ipwv = Infobs(IRAM30mEMIR(linewidth, ipwv=0), n_samples, samplers, obstime=30)

infobs_ipwv.set_path(os.path.join("informativity", "data-ipwv"))

infobs_ipwv.set_regimes(regimes)

Informativity of CO isotopologues lines

Simple informativity plot

plt.figure(figsize=(1.6 * 6.4, 1.6 * 0.4 * 4.8), dpi=150)

infobs.plot_info_bars(
    lines,
    "Av",
    restriction="all",
    errorbars=True,
    nfirst=None,
    transitions=True,
)

plt.show()

Influence of integration time

plt.figure(figsize=(1.6 * 6.4, 1.6 * 0.4 * 4.8), dpi=150)

infobs_integ.plot_info_bars_comparison(
    infobs,
    lines,
    "Av",
    restriction="all",
    errorbars=True,
    labels=["Deeper integration", "Reference"],
    transitions=True,
)

plt.show()

Influence of observing conditions

plt.figure(figsize=(1.6 * 6.4, 1.6 * 0.4 * 4.8), dpi=150)

infobs_ipwv.plot_info_bars_comparison(
    infobs,
    lines,
    "Av",
    restriction="all",
    errorbars=True,
    labels=["Excellent conditions", "Normal conditions"],
    transitions=True,
)

plt.show()

Influence of combining lines

import itertools as itt

plt.figure(figsize=(1.6 * 6.4, 1.6 * 0.4 * 4.8), dpi=150)

infobs.plot_info_bars(
    itt.combinations(lines, 2),
    "Av",
    restriction="all",
    errorbars=True,
    nfirst=15,  # Display only the 20 most informative lines
    transitions=True,
    bottom_val=2.0,
)

plt.show()

plt.figure(figsize=(1.6 * 6.4, 1.6 * 0.4 * 4.8), dpi=150)

infobs.plot_info_matrix(lines, "Av", restriction="all", transitions=True)

plt.show()

Influence of physical regime

regimes = ["diffuse", "translucent", "filamentary", "dense"]
for regime in regimes:

    plt.figure(figsize=(1.6 * 6.4, 1.6 * 0.4 * 4.8), dpi=150)

    infobs.plot_info_bars(
        lines,
        "Av",
        restriction=regime,
        errorbars=True,
        nfirst=None,
        transitions=True,
    )
    plt.title(f"{regime.capitalize()} gas", fontsize=20, pad=10)

    plt.show()

Note: The Infobs.plot_info_bars method should be seen as a simple, easy-to-use helpers for displaying mutual information values and the associated error bars. For more specific scripts, feel free to implement your own function by taking Infobs.plot_info_bars as a template.

Information maps

n_samples = 200_000

samplers = {
    "Av": LogUniform(1, 40),
    "G0": LogUniform(1, 1e5),
    "Pth": Constant(1e5),
    "angle": Constant(0.0),
}

global_kwargs = {
    "parameters": "Av",
    "xaxis_param": "Av",
    "yaxis_param": "G0",
}

computation_kwargs = {"points": 50, "progress_bar": True}

plot_kwargs = {"cmap": "jet"}

Reference use case

infobs_maps = Infobs(instrument, n_samples, samplers, obstime=30)

infobs_maps.set_path(os.path.join("informativity", "maps"))

infobs_maps.set_regimes(regimes)

Deep integration

infobs_maps_integ = Infobs(instrument, n_samples, samplers, obstime=300)

infobs_maps_integ.set_path(os.path.join("informativity", "maps-integration"))

infobs_maps_integ.set_regimes(regimes)

Information maps for single lines

# Pre-compute maps to be able to get the pixel-wise maximum for colormaps

for line in ["co_v0_j1__v0_j0", "13c_o_j1__j0", "c_18o_j1__j0"]:

    infobs_maps.compute_map(line, **global_kwargs, **computation_kwargs)

# Get maximum mutual information

vmax = infobs_maps.get_info_maps_max(
    ["co_v0_j1__v0_j0", "13c_o_j1__j0", "c_18o_j1__j0"], **global_kwargs
)

fig, axs = plt.subplots(1, 3, figsize=(3 * 6.4, 4.8), dpi=150)
axs = axs.flatten()

# 12C0

line = "co_v0_j1__v0_j0"
plt.sca(axs[0])

infobs_maps.plot_info_map(line, **global_kwargs, **plot_kwargs, vmax=vmax)
plt.title(f"${infobs_maps.plotter.line_formatter(line)}$", fontsize=20, pad=10)

# 13C0

line = "13c_o_j1__j0"
plt.sca(axs[1])

infobs_maps.plot_info_map(line, **global_kwargs, **plot_kwargs, vmax=vmax)
plt.title(f"${infobs_maps.plotter.line_formatter(line)}$", fontsize=20, pad=10)

# C180

line = "c_18o_j1__j0"
plt.sca(axs[2])

infobs_maps.plot_info_map(line, **global_kwargs, **plot_kwargs, vmax=vmax)
plt.title(f"${infobs_maps.plotter.line_formatter(line)}$", fontsize=20, pad=10)

plt.tight_layout()
plt.show()

Influence of integration time

line = "c_18o_j1__j0"

# Pre-compute maps to be able to get the pixel-wise maximum for colormaps

infobs_maps.compute_map(line, **global_kwargs, **computation_kwargs)

infobs_integ.compute_map(line, **global_kwargs, **computation_kwargs)

# Get maximum mutual information

vmax = max(
    infobs_maps.get_info_maps_max(
        [line],
        **global_kwargs,
    ),
    infobs_maps_integ.get_info_maps_max(
        [line],
        **global_kwargs,
    ),
)

fig, axs = plt.subplots(1, 2, figsize=(2 * 6.4, 4.8), dpi=150)
axs = axs.flatten()

# Reference

plt.sca(axs[0])

infobs_maps.plot_info_map(line, **global_kwargs, **plot_kwargs, vmax=vmax)
plt.title(f"${infobs_maps.plotter.line_formatter(line)}$", fontsize=20, pad=10)

# 10 times longer integration

plt.sca(axs[1])

infobs_maps_integ.plot_info_map(line, **global_kwargs, **plot_kwargs, vmax=vmax)
plt.title(f"${infobs_maps_integ.plotter.line_formatter(line)}$", fontsize=20, pad=10)

plt.tight_layout()
plt.show()

Influence of combining lines

lines = ["co_v0_j1__v0_j0", "13c_o_j1__j0", "c_18o_j1__j0"]

infobs_maps.compute_map(lines, **global_kwargs, **computation_kwargs)

fig = plt.figure(dpi=100)

infobs_maps.plot_info_map(lines, **global_kwargs, **plot_kwargs, vmax=vmax)
plt.title(f"${infobs_maps.plotter.lines_comb_formatter(lines)}$", fontsize=20, pad=10)

plt.show()

Note: The Infobs.plot_info_map method should be seen as a simple, easy-to-use helpers for displaying mutual information maps. For more specific scripts, feel free to implement your own function by taking Infobs.plot_info_map as a template.