Examples

Basic Cosmology Calculations

This example demonstrates how to use the w0waCDMCosmology class to compute various cosmological quantities and visualize them.

Setting up the Cosmology

First, let's create a cosmology instance with standard ΛCDM parameters (Planck 2018):

import jax
import jax.numpy as jnp
import matplotlib.pyplot as plt
import jaxace

# Create a cosmology instance with Planck 2018 parameters
cosmo = jaxace.w0waCDMCosmology(
    ln10As=3.044,      # ln(10^10 A_s)
    ns=0.9649,         # Scalar spectral index
    h=0.6736,          # Hubble parameter
    omega_b=0.02237,   # Baryon density
    omega_c=0.1200,    # CDM density
    m_nu=0.06,         # Sum of neutrino masses in eV
    w0=-1.0,           # Dark energy equation of state
    wa=0.0             # Dark energy evolution parameter
)

Computing Distances

Now let's compute the comoving distance for a range of redshifts:

# Create redshift array from 0 to 3
z = jnp.linspace(0.01, 3.0, 100)

# Compute comoving distance
r_comoving = cosmo.r_z(z)

# Also compute angular diameter and luminosity distances
dA = cosmo.dA_z(z)
dL = cosmo.dL_z(z)

Computing Growth Functions

Let's compute the linear growth factor and growth rate:

# Compute growth factor D(z) and growth rate f(z)
D = cosmo.D_z(z)
f = cosmo.f_z(z)

# We can also get both at once for efficiency
D_both, f_both = cosmo.D_f_z(z)

Visualizing the Results

Now let's create plots to visualize these cosmological quantities:

# Create a figure with subplots
fig, axes = plt.subplots(2, 2, figsize=(12, 10))

# Plot 1: Distance measures
ax = axes[0, 0]
ax.plot(z, r_comoving, label='Comoving distance $r(z)$', linewidth=2)
ax.plot(z, dA, label='Angular diameter distance $d_A(z)$', linewidth=2)
ax.plot(z, dL, label='Luminosity distance $d_L(z)$', linewidth=2)
ax.set_xlabel('Redshift $z$')
ax.set_ylabel('Distance [Mpc]')
ax.set_title('Cosmological Distances')
ax.legend()
ax.grid(True, alpha=0.3)

# Plot 2: Hubble parameter
ax = axes[0, 1]
E_z = jaxace.E_z(z, cosmo.omega_b + cosmo.omega_c, cosmo.h,
                 mν=cosmo.m_nu, w0=cosmo.w0, wa=cosmo.wa)
H_z = 100 * cosmo.h * E_z  # H(z) in km/s/Mpc
ax.plot(z, H_z, color='darkblue', linewidth=2)
ax.set_xlabel('Redshift $z$')
ax.set_ylabel('$H(z)$ [km/s/Mpc]')
ax.set_title('Hubble Parameter Evolution')
ax.grid(True, alpha=0.3)

# Plot 3: Growth factor
ax = axes[1, 0]
ax.plot(z, D, color='darkgreen', linewidth=2)
ax.set_xlabel('Redshift $z$')
ax.set_ylabel('$D(z)$')
ax.set_title('Linear Growth Factor')
ax.grid(True, alpha=0.3)
ax.invert_xaxis()  # Convention: D grows as z decreases

# Plot 4: Growth rate
ax = axes[1, 1]
ax.plot(z, f, color='darkred', linewidth=2)
ax.set_xlabel('Redshift $z$')
ax.set_ylabel('$f(z) = d\\ln D / d\\ln a$')
ax.set_title('Growth Rate')
ax.grid(True, alpha=0.3)
ax.set_ylim([0.4, 1.0])

plt.tight_layout()
plt.show()

Using JAX Features

One of the advantages of jaxace is that all functions are JAX-compatible, enabling automatic differentiation and JIT compilation:

# JIT compile for faster repeated calculations
@jax.jit
def compute_distances_fast(z_array, omega_c, omega_b, h):
    """JIT-compiled function for fast distance calculations."""
    cosmo = jaxace.w0waCDMCosmology(
        ln10As=3.044, ns=0.9649, h=h,
        omega_b=omega_b, omega_c=omega_c,
        m_nu=0.06, w0=-1.0, wa=0.0
    )
    return cosmo.r_z(z_array)

# Compute gradient with respect to cosmological parameters
grad_fn = jax.grad(lambda omega_c: compute_distances_fast(
    jnp.array([1.0]), omega_c, 0.02237, 0.6736
).sum())

# Calculate derivative of comoving distance at z=1 with respect to omega_c
d_r_d_omega_c = grad_fn(0.12)
print(f"∂r(z=1)/∂Ω_c = {d_r_d_omega_c:.2f} Mpc")

Computing Jacobians of Growth Functions

A powerful feature of jaxace is the ability to compute Jacobians (derivatives) of cosmological quantities with respect to all input parameters simultaneously:

import jax
import jax.numpy as jnp
import matplotlib.pyplot as plt
import jaxace

# Define a function that computes growth factor for given parameters
def growth_factor_function(params, z):
    """Compute growth factor D(z) for given cosmological parameters."""
    omega_c, omega_b, h, m_nu, w0, wa = params
    cosmo = jaxace.w0waCDMCosmology(
        ln10As=3.044,  # Keep fixed for this example
        ns=0.9649,      # Keep fixed for this example
        h=h,
        omega_b=omega_b,
        omega_c=omega_c,
        m_nu=m_nu,
        w0=w0,
        wa=wa
    )
    return cosmo.D_z(z)

# Define fiducial parameters (Planck 2018)
fiducial_params = jnp.array([
    0.1200,   # omega_c
    0.02237,  # omega_b
    0.6736,   # h
    0.06,     # m_nu (eV)
    -1.0,     # w0
    0.0       # wa
])

# Redshift array
z = jnp.linspace(0.01, 3.0, 50)

# Compute Jacobian matrix: dD/dθ for all parameters at all redshifts
jacobian_fn = jax.jacobian(growth_factor_function, argnums=0)
jacobian = jacobian_fn(fiducial_params, z)

# jacobian shape is (n_z, n_params)
print(f"Jacobian shape: {jacobian.shape}")
print(f"Parameters: [omega_c, omega_b, h, m_nu, w0, wa]")

# Plot the derivatives
fig, axes = plt.subplots(2, 3, figsize=(15, 10))
param_names = ['$\\Omega_c$', '$\\Omega_b$', '$h$', '$m_\\nu$ [eV]', '$w_0$', '$w_a$']
param_labels = ['omega_c', 'omega_b', 'h', 'm_nu', 'w0', 'wa']

for i, (ax, name, label) in enumerate(zip(axes.flat, param_names, param_labels)):
    ax.plot(z, jacobian[:, i], linewidth=2.5, color=plt.cm.viridis(i/5))
    ax.set_xlabel('Redshift $z$')
    ax.set_ylabel(f'$\\partial D/\\partial${name}')
    ax.set_title(f'Growth Factor Sensitivity to {name}')
    ax.grid(True, alpha=0.3)
    ax.axhline(y=0, color='k', linestyle='--', alpha=0.3)

    # Add text with max sensitivity
    max_idx = jnp.argmax(jnp.abs(jacobian[:, i]))
    max_z = z[max_idx]
    max_val = jacobian[max_idx, i]
    ax.text(0.95, 0.95, f'Max at z={max_z:.2f}\\n∂D/∂{label}={max_val:.3f}',
            transform=ax.transAxes, ha='right', va='top',
            bbox=dict(boxstyle='round', facecolor='wheat', alpha=0.5))

plt.suptitle('Jacobian of Growth Factor D(z) w.r.t. Cosmological Parameters',
             fontsize=16, fontweight='bold')
plt.tight_layout()
plt.show()

Computing Jacobians for Multiple Quantities

We can also compute Jacobians for both growth factor D(z) and growth rate f(z) simultaneously:

# Define a function that returns both D and f
def growth_functions(params, z):
    """Compute both D(z) and f(z) for given parameters."""
    omega_c, omega_b, h, m_nu, w0, wa = params
    cosmo = jaxace.w0waCDMCosmology(
        ln10As=3.044, ns=0.9649,
        h=h, omega_b=omega_b, omega_c=omega_c,
        m_nu=m_nu, w0=w0, wa=wa
    )
    D, f = cosmo.D_f_z(z)
    return jnp.stack([D, f])  # Stack to get shape (2, n_z)

# Compute Jacobian for both quantities
jacobian_both_fn = jax.jacobian(growth_functions, argnums=0)
jacobian_both = jacobian_both_fn(fiducial_params, z)

# jacobian_both shape is (2, n_z, n_params)
# where index 0 is D(z) and index 1 is f(z)
jacobian_D = jacobian_both[0]  # Shape: (n_z, n_params)
jacobian_f = jacobian_both[1]  # Shape: (n_z, n_params)

# Plot comparison of sensitivities for D and f
fig, axes = plt.subplots(2, 3, figsize=(15, 10))

for i, (ax, name, label) in enumerate(zip(axes.flat, param_names, param_labels)):
    ax.plot(z, jacobian_D[:, i], label='$\\partial D/\\partial$' + name,
            linewidth=2.5, color='blue')
    ax.plot(z, jacobian_f[:, i], label='$\\partial f/\\partial$' + name,
            linewidth=2.5, color='red', linestyle='--')
    ax.set_xlabel('Redshift $z$')
    ax.set_ylabel('Derivative')
    ax.set_title(f'Sensitivity to {name}')
    ax.legend(loc='best')
    ax.grid(True, alpha=0.3)
    ax.axhline(y=0, color='k', linestyle=':', alpha=0.3)

plt.suptitle('Comparison of D(z) and f(z) Sensitivities',
             fontsize=16, fontweight='bold')
plt.tight_layout()
plt.show()

Comparing Different Cosmologies

Let's compare how different dark energy models affect the growth of structure:

# Define different cosmologies
cosmologies = {
    'ΛCDM': {'w0': -1.0, 'wa': 0.0},
    'wCDM': {'w0': -0.9, 'wa': 0.0},
    'w0waCDM': {'w0': -0.95, 'wa': 0.3},
}

# Compute growth factors for each cosmology
fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(12, 5))

for name, params in cosmologies.items():
    cosmo_test = jaxace.w0waCDMCosmology(
        ln10As=3.044, ns=0.9649, h=0.6736,
        omega_b=0.02237, omega_c=0.1200,
        m_nu=0.06, **params
    )

    D_test = cosmo_test.D_z(z)
    f_test = cosmo_test.f_z(z)

    ax1.plot(z, D_test, label=name, linewidth=2)
    ax2.plot(z, f_test, label=name, linewidth=2)

ax1.set_xlabel('Redshift $z$')
ax1.set_ylabel('$D(z)$')
ax1.set_title('Growth Factor Comparison')
ax1.legend()
ax1.grid(True, alpha=0.3)
ax1.invert_xaxis()

ax2.set_xlabel('Redshift $z$')
ax2.set_ylabel('$f(z)$')
ax2.set_title('Growth Rate Comparison')
ax2.legend()
ax2.grid(True, alpha=0.3)
ax2.set_ylim([0.4, 1.0])

plt.tight_layout()
plt.show()

Performance Tips

JIT Compilation

For optimal performance, especially when computing many values:

# JIT compile the entire calculation pipeline
@jax.jit
def compute_all_quantities(z_array, cosmo):
    """Compute all cosmological quantities at once."""
    r = cosmo.r_z(z_array)
    D, f = cosmo.D_f_z(z_array)
    return r, D, f

# First call will compile
z_test = jnp.linspace(0.01, 3.0, 1000)
r, D, f = compute_all_quantities(z_test, cosmo)

# Subsequent calls will be much faster
%timeit compute_all_quantities(z_test, cosmo)

Vectorization

All functions support vectorized inputs for efficient batch calculations:

# Compute for multiple redshifts at once (vectorized)
z_grid = jnp.logspace(-2, 0.5, 50)  # z from 0.01 to ~3.16
r_grid = cosmo.r_z(z_grid)

# This is much faster than a loop:
# for zi in z_grid:
#     ri = cosmo.r_z(zi)  # Slower!

Next Steps

• Explore the API Reference for detailed function documentation
• Check the neutrino mass effects by varying m_nu
• Experiment with different dark energy models using w0 and wa
• Use JAX's automatic differentiation to compute derivatives of cosmological quantities