Quick Start Guide

This guide will help you get started with Goku-ELG quickly.

Basic Usage

Computing Galaxy Clustering

Here’s a simple example to compute the galaxy-galaxy correlation function:

import numpy as np
from gal_goku import gal

# Initialize the galaxy clustering module
galaxy_model = gal.GalBase(logging_level='INFO')

# Define cosmological parameters
# [Omega_m, Omega_b, h, A_s, n_s, w0, wa, N_ur, alpha_s, m_nu]
cosmo_params = [0.3, 0.05, 0.7, 2.1e-9, 0.96, -1.0, 0.0, 3.046, 0.0, 0.06]

# Get the initial linear power spectrum
k, P_lin = galaxy_model.get_init_linear_power(
    cosmo_params,
    redshifts=[2.0]
)

print(f"Computed power spectrum for {len(k)} k-modes")

Using the Halo Mass Function Emulator

from gal_goku import emus

# Initialize the HMF emulator
hmf_emu = emus.Hmf(
    data_dir='path/to/data',
    y_log=True,
    fid='L2',
    logging_level='INFO'
)

# Train the emulator (if not already trained)
# This step may take a few minutes
predictions, truth, mass_bins = hmf_emu.train_pred_all_sims()

print(f"HMF predictions shape: {predictions.shape}")
print(f"Mass bins: {mass_bins}")

Making Predictions

Once the emulator is trained, you can make predictions for new cosmological parameters:

import numpy as np

# Define test cosmological parameters
# Shape: (n_samples, n_parameters)
X_test = np.array([
    [0.31, 0.049, 0.68, 2.0e-9, 0.965, -1.0, 0.0, 3.046, 0.0, 0.06],
    [0.28, 0.047, 0.72, 2.2e-9, 0.955, -0.9, 0.0, 3.046, 0.0, 0.08],
])

# Get predictions and uncertainties
predictions, variances = hmf_emu.predict(X_test)

print(f"Predictions: {predictions}")
print(f"Variances: {variances}")

Working with Halo-Halo Correlations

from gal_goku_sims import xi

# Load halo-halo correlation function data
# This assumes you have simulation data available
halo_xi = xi.HaloXi(data_dir='path/to/halo/data')

# Process and analyze correlations
# (Specific methods depend on your data structure)

Complete Workflow Example

Here’s a complete workflow from cosmology to galaxy clustering:

import numpy as np
from gal_goku import gal, emus
import matplotlib.pyplot as plt

# Step 1: Set up cosmological parameters
cosmo_params = {
    'omega_m': 0.3,
    'omega_b': 0.05,
    'h': 0.7,
    'A_s': 2.1e-9,
    'n_s': 0.96,
    'w0': -1.0,
    'wa': 0.0,
    'N_ur': 3.046,
    'alpha_s': 0.0,
    'm_nu': 0.06
}

# Convert to array format
cosmo_array = [cosmo_params[key] for key in [
    'omega_m', 'omega_b', 'h', 'A_s', 'n_s',
    'w0', 'wa', 'N_ur', 'alpha_s', 'm_nu'
]]

# Step 2: Initialize galaxy model
gal_model = gal.GalBase(logging_level='INFO')

# Step 3: Get linear power spectrum
k, P_lin = gal_model.get_init_linear_power(
    cosmo_array,
    redshifts=[2.0]
)

# Step 4: Plot the results
plt.figure(figsize=(10, 6))
plt.loglog(k, P_lin)
plt.xlabel('k [h/Mpc]')
plt.ylabel('P(k) [(Mpc/h)^3]')
plt.title('Linear Matter Power Spectrum')
plt.grid(True, alpha=0.3)
plt.savefig('power_spectrum.png')
print("Power spectrum saved to power_spectrum.png")

Understanding Parameters

Cosmological Parameters

The emulator uses 10 cosmological parameters:

Parameter	Symbol	Description
omega_m	\(\Omega_m\)	Total matter density
omega_b	\(\Omega_b\)	Baryon density
h	\(h\)	Hubble parameter (H0 = 100h km/s/Mpc)
A_s	\(A_s\)	Scalar amplitude (×10⁻⁹)
n_s	\(n_s\)	Scalar spectral index
w0	\(w_0\)	Dark energy EoS (present)
wa	\(w_a\)	Dark energy EoS (evolution)
N_ur	\(N_{ur}\)	Effective number of relativistic species
alpha_s	\(\alpha_s\)	Running of spectral index
m_nu	\(m_\nu\)	Sum of neutrino masses [eV]

Typical Ranges

Here are typical parameter ranges for ΛCDM-like cosmologies:

parameter_ranges = {
    'omega_m': (0.24, 0.40),
    'omega_b': (0.04, 0.06),
    'h': (0.60, 0.80),
    'A_s': (1.8e-9, 2.4e-9),
    'n_s': (0.92, 1.00),
    'w0': (-1.2, -0.8),
    'wa': (-0.5, 0.5),
    'N_ur': (2.5, 3.5),
    'alpha_s': (-0.02, 0.02),
    'm_nu': (0.0, 0.15)
}

Best Practices

Start Simple: Begin with the default/fiducial cosmology before exploring parameter space
Validate: Always check emulator predictions against known results
Stay in Range: Keep parameters within the trained range for reliable predictions
Check Uncertainties: Use the variance estimates to assess prediction confidence
Logging: Enable logging to track computation progress and debug issues

Common Patterns

Leave-One-Out Cross-Validation

from gal_goku import emus

hmf_emu = emus.Hmf(data_dir='path/to/data')

# Perform leave-one-out cross-validation
hmf_emu.loo_train_pred(savefile='loo_results.h5')

Batch Predictions

# Generate many test cosmologies
n_samples = 100
X_test = np.random.uniform(low=[0.24, 0.04, 0.60, 1.8e-9, 0.92,
                                 -1.2, -0.5, 2.5, -0.02, 0.0],
                            high=[0.40, 0.06, 0.80, 2.4e-9, 1.00,
                                  -0.8, 0.5, 3.5, 0.02, 0.15],
                            size=(n_samples, 10))

# Make predictions for all samples
predictions, variances = hmf_emu.predict(X_test)

Next Steps

Explore Tutorials for detailed examples
Check the API Reference for complete API documentation
See example notebooks in the emu/notebooks/ directory