I am using ax.pcolor() to create a plot in polar coordinates, and I would like to add a radial axis-like line at a fixed angle (e.g., 250 degrees).

Specifically, I would like the axis-like line to:

• point along the radial direction;

• have an arrowhead at the end;

• have radial ticks with a consistent visual length;

• be gray;

• optionally extend slightly beyond the current radial limit.

I can draw the arrow using:

axs[0].annotate('', xy = (_phi, rs[-1]), xytext = (_phi, rs[0]),
                arrowprops = dict(arrowstyle = '->', color = 'gray', lw = 1.5))
_annotation.arrow_patch.set_clip_on(False)

But `annotate` does not connect starting and end points exactly as instructed (it include a shift automatically). Also if I extend the endpoint beyond the current radial limit, the line disappears.

I have also tried `FancyArrowPatch` from `matplotlib.patches`, but it ruins the entire figure:

p_0 = ax.transData.transform((_phi, rs[0]))
p_1 = ax.transData.transform((_phi, rs[-1]))
arrow = FancyArrowPatch(p_0, p_1, arrowstyle = '->', mutation_scale = 12, color='gray', lw = 1.2)
axs[0].add_patch(arrow)

What would be the recommended matplotlib approach for creating such a radial axis with an arrow, while optionally allowing it to extend beyond the polar axes? Thanks in advance!

My MWE and its outcome is attached:

#!/usr/bin/env python
import copy
import numpy as np
import os, sys

import matplotlib
matplotlib.use('Agg')
matplotlib.rcParams['text.latex.preamble'] = r'\usepackage{amsmath}'
matplotlib.rc('text', usetex = True)
from matplotlib import pyplot as py
from matplotlib.patches import FancyArrowPatch

def plot():
    ## setup canvas
    n_columns, n_rows = 1, 1
    figure = py.figure(figsize = (n_columns * 7.5, n_rows * 5.0))
    axs = [py.subplot(n_rows, n_columns, _ + 1, projection = 'polar') for _ in range(1)]

    ## get r and phi values
    rs = np.linspace(0.0, 3.0, 50)
    phis = np.linspace(0.0, 2.0 * np.pi, 100)

    ## make plot
    r_grid, phi_grid = np.meshgrid(rs, phis)
    y_grid = []
    for phi in phis:
        _y_grid = [np.cos(phi) * qt for qt in rs]
        y_grid.append(np.array(_y_grid))
    axs[0].pcolor(phi_grid, r_grid, y_grid, cmap = 'bwr')

    ## draw radial axis
    _phi = np.deg2rad(250.0)
    # _annotation = axs[0].annotate('', xy = (_phi, 1.2 * rs[-1]), xytext = (_phi, rs[0]),
    #                               arrowprops = dict(arrowstyle = '->', color = 'gray', lw = 1.5))
    _annotation = axs[0].annotate('', xy = (_phi, rs[-1]), xytext = (_phi, rs[0]),
                                  arrowprops = dict(arrowstyle = '->', color = 'gray', lw = 1.5))
    _annotation.arrow_patch.set_clip_on(False)
    for _r in [1.0, 2.0, 3.0]:
        d_phi = np.deg2rad(3.0) / _r
        axs[0].plot([_phi - d_phi, _phi + d_phi], [_r, _r], color = 'gray', lw = 0.8, zorder = 5)

    axs[0].grid(False)
    axs[0].set_xticklabels([])
    axs[0].set_yticklabels([])

    ## save figure
    name  = './test.pdf'
    py.subplots_adjust(wspace = 0.0)
    py.savefig(name, bbox_inches = 'tight')
    py.close()
    return name

if __name__ == '__main__':
    plot()