如何使用 APLpy 或 python 2.7 在 FITS 图上绘制线段？

Question

I want to draw a line segment joining two points on a FITS figure.

(x,y) co-ordinates of these points are (200,250) & (300,400).

I am using APLpy for this.

My code is:

import matplotlib.pyplot as plt
import aplpy
import numpy as np

fig = aplpy.FITSFigure('test.fits')
fig.show_grayscale()

a=np.ndarray(shape=(2,2))
a[0][0]=200
a[0][1]=250
a[1][0]=300
a[1][1]=400

fig.show_lines(a)

plt.show() 

I am using "fig.show_lines()" function of APLpy described on following web-page: http://aplpy.readthedocs.org/en/latest/quick_reference.html#shapes

It says 'use lists of numpy arrays' as argument to show_lines().

But I got following error message:

Traceback (most recent call last):
File "draw.py", line 16, in <module>
fig.show_lines(a)
File "<string>", line 2, in show_lines
File "/home/swapnil/anaconda/lib/python2.7/site-packages/aplpy/decorators.py", line  25, in _auto_refresh
return f(*args, **kwargs)
File "/home/swapnil/anaconda/lib/python2.7/site-packages/aplpy/aplpy.py", line 1275, in show_lines
xp, yp = wcs_util.world2pix(self._wcs, line[0, :], line[1, :])
IndexError: too many indices

Any help will be appreciated.

Thanks.

Answer 1

I understand that it should be a list of 2xN numpy arrays:

x = np.array([[200], [300]])
y = np.array([[250], [400]])

fig.show_lines([x, y])

HTH,

Germán.

Answer 2

You need to do something like this:

iline = np.array([[x1, x2],[y1,y2]])
fig.show_lines([iline], color = 'r')

Where x1 , x2 , y1 , y2 are in the correct units (for me this was degrees)

Answer 3

I ran into a similar issue recently and the answers did not seem to work. Have you tried plotting the lines using

    matplotlib.pyplot.plot ( usually called as plt.plot)?

That did the trick for me.

I opened a figure object using:

    fig = aplpy.FITSFigure('fitsfile.fits')

and then simply did:

    plt.plot([x1, x2], [y1, y2], color = 'r')

and it worked perfectly well. I did not even have to use the RA and Dec and directly used the pixel coordinates obtained from using astropy's all_world2pix(). I'm pretty sure aplpy has an equivalent to that but I never tried it.