Implementing render functions

[dfm]

After this PR, the following code:

import matplotlib.pyplot as plt
import jax.numpy as jnp
from jaxoplanet.experimental.starry.render import render

y = jnp.array(
    [1.00,  0.22,  0.19,  0.11,  0.11,  0.07,  -0.11, 0.00,  -0.05,
     0.12,  0.16,  -0.05, 0.06,  0.12,  0.05,  -0.10, 0.04,  -0.02,
     0.01,  0.10,  0.08,  0.15,  0.13,  -0.11, -0.07, -0.14, 0.06,
     -0.19, -0.02, 0.07,  -0.02, 0.07,  -0.01, -0.07, 0.04,  0.00]
)
img = render(5, 500, 0.0, 0.0, 0.0, y)
plt.imshow(img, origin="lower")
plt.savefig("render.png", dpi=100, bbox_inches="tight");

produces this image:

which we can compare to the starry docs:

[shashankdholakia]

Sorry for lurking--been using this render function and I noticed a possible typo in left_project that makes this function fail for rotation angles different from 0:

jaxoplanet/src/jaxoplanet/experimental/starry/render.py

Lines 19 to 34 in 361397c

	def left_project(deg, M, theta, inc, obl):
	# Note that here we are using the fact that R . M = (M^T . R^T)^T
	MT = jnp.transpose(M)
	# Rotate to the polar frame
	MT = dot_rotation_matrix(deg, 1.0, 0.0, 0.0, -0.5 * jnp.pi)(MT)
	MT = dot_rotation_matrix(deg, None, None, 1.0, -theta)(MT)
	# Rotate to the sky frame
	MT = dot_rotation_matrix(deg, 1.0, 0.0, 0.0, 0.5 * jnp.pi)(MT)
	MT = dot_rotation_matrix(deg, None, None, 1.0, -obl)(MT)
	MT = dot_rotation_matrix(
	deg, -jnp.cos(obl), -jnp.sin(obl), 0.0, 0.5 * jnp.pi - inc
	)(MT)
	return M.T

Last line should return MT, not M.T?

[dfm]

That seems like it would be better, doesn't it! Would you be willing to open a small PR to that branch?

[dfm]

Superseded by #111