Introduced area-based contours

[Paras20222]

PR Description

This PR introduces area-based contour calculations to enhance visualization accuracy. The changes address issue #7420 and provide a more intuitive method for determining contour levels. This improvement enables better data interpretation in X-ray and radio imaging applications.

[nabobalis]

Thank you for the PR @Paras20222. I will do a quick review now.

[nabobalis]

I am not sure the word containment is correct here.

[nabobalis]

I think the imports are missing?

Also the print is not needed here. The doctest will print out the return if you do not assign it to a variable.

[nabobalis]

We can't change the default behavior of this method.

I think for now, adding a "method" keyword (to the draw contours method) that allows you to select this new method would be better.

[nabobalis]

This will also need to be tested both as a normal unit test but also as a figure test. See https://docs.sunpy.org/en/latest/dev_guide/contents/tests.html for more details.

[nabobalis]

Hi @Paras20222, will you have some time soon to address the outstanding comments?

[AquarlisPrime]

Hello, i have made few changes to the gist.
`import numpy as np
import matplotlib.pyplot as plt

def area_contour_image(img: np.ndarray, quantiles: list[float], ax=None, x=None, y=None, cmap='Reds', **contour_kwds):

# Normalizing img data
normalized_img = (img - np.min(img)) / np.ptp(img)

# Flatten img data
flattened_img = normalized_img.ravel()

# Sort pixels: small areas first, large areas later
sorted_values = np.sort(flattened_img)[::-1]

# Computing cumi sum
cumulative_sum = np.cumsum(sorted_values) / np.sum(sorted_values)

# Indices corresponding to specified quantiles
indices = np.searchsorted(cumulative_sum, quantiles)

# Threshold values for desired percentile and for drawing contours
thresholds = np.sort(sorted_values[indices])

if ax is None:
    fig, ax = plt.subplots()
if x is None or y is None:
    return ax.contour(normalized_img, levels=thresholds, cmap=cmap, **contour_kwds)
else:
    return ax.contour(x, y, normalized_img, levels=thresholds, cmap=cmap, **contour_kwds)

def test():
x = np.linspace(-10, 10, num=1000)
y = x.copy()
xx, yy = np.meshgrid(x, y)

# Similar to Gaussian func
sigx, sigy = 2, 5
z = 5 * np.exp(-(xx**2 / sigx**2 + yy**2 / sigy**2))

# Plot
fig, ax = plt.subplots(figsize=(6, 6))
ax.pcolormesh(xx, yy, z, cmap='inferno')

# Generating contours
contours = area_contour_image(z, [0.05, 0.3, 0.5, 0.8], ax=ax, x=x, y=y, cmap='Reds', linewidths=1.75)

# Plot customization
ax.set(xlabel='x', ylabel='y', title='Area Contours')
ax.grid(True)
plt.colorbar(contours, ax=ax)
plt.show()

if name == 'main':
test()`

[nabobalis]

Hello, i have made few changes to the gist. `import numpy as np import matplotlib.pyplot as plt

Modifying the gist is not the most useful, we can't make suggestions nor does it move this pull request forward.

I would suggest that you take over this pull request, you will need to fetch the original comments from this pull request onto your sunpy fork, then submit a new pull request with any changes you think need to be made as well addressing the comments that i left here?

Does that sound ok? We can also help with any of the steps if you get stuck.

[AquarlisPrime]

Sure, I would be happy to work upon it. I am new to open source contributions, so I am thankful for your words of guidance.