Add show and fig underscore magic to magpylib.show(...) and style.legend.show property

[Alexboiboi]

Related Issues

• Fixes showlegend in plotly #650
• Fixes Custom Source Bug #1 #654
• Fixes CustomSource Bug #2 #655

Notes

fig and show kwargs magic

In the same vein as the style magic:

• kwargs with fig prefix are forwarded to the figure constructor (Figure for mpl and plotly, Plotter for pyvista)
• kwargs with show prefix are forwarded to the show call (for all backends)
• if fig or show are prefixed with a backend name, the kwargs will only be forwarded to the specific backend.

Here an example below with increasing order of precedence:

import magpylib as magpy

c = magpy.magnet.Cuboid((0, 0, 1000), (1, 1, 1))
c.show(
    fig=dict(layout_title_text="fig"),
    fig_layout_title_text="fig_layout",
    plotly_fig=dict(layout_title_text="plotly_fig"),
    plotly_fig_layout_title_text="plotly_fig_layout",
    backend="plotly",
)

legend style

For the legend I introduced it at the object level as style.legend.show. The advantage is, that you one can turn off specific traces at will.

magpy.show(..., style_legend_show=False) # hides all legends

[codecov]

Codecov Report

Merging #657 (9021fd0) into main (2169f6d) will increase coverage by 0.03%.
The diff coverage is 99.56%.

@@            Coverage Diff             @@
##             main     #657      +/-   ##
==========================================
+ Coverage   99.22%   99.26%   +0.03%     
==========================================
  Files          55       55              
  Lines        7000     7073      +73     
==========================================
+ Hits         6946     7021      +75     
+ Misses         54       52       -2     

Files Changed	Coverage Δ
magpylib/_src/defaults/defaults_values.py	100.00% <ø> (ø)
magpylib/_src/style.py	99.73% <94.44%> (-0.13%)	⬇️
magpylib/_src/display/backend_matplotlib.py	100.00% <100.00%> (ø)
magpylib/_src/display/backend_plotly.py	100.00% <100.00%> (ø)
magpylib/_src/display/backend_pyvista.py	83.33% <100.00%> (+1.78%)	⬆️
magpylib/_src/display/display.py	100.00% <100.00%> (ø)
magpylib/_src/display/traces_core.py	100.00% <100.00%> (ø)
magpylib/_src/display/traces_generic.py	100.00% <100.00%> (ø)
magpylib/_src/display/traces_utility.py	100.00% <100.00%> (+0.24%)	⬆️
magpylib/_src/obj_classes/class_Collection.py	100.00% <100.00%> (ø)
... and 11 more

[Alexboiboi]

One example to test the legends

import magpylib as magpy
import numpy as np

objs = []
f=0.5
N = 3
xs = f * np.array([-1, -1, 1, 1, -1, -1, -1, -1, -1, 1, 1, 1, 1, 1, 1, -1])
ys = f * np.array([-1, 1, 1, -1, -1, -1, 1, 1, 1, 1, 1, 1, -1, -1, -1, -1])
zs = f * np.array([-1, -1, -1, -1, -1, 1, 1, -1, 1, 1, -1, 1, 1, -1, 1, 1])
trace_mpl = dict(
    backend="matplotlib",
    constructor="plot",
    args=(xs, ys, zs),
    kwargs={"marker": "", "lw": 1},
)
trace_plotly = dict(
    backend="plotly",
    constructor="scatter3d",
    kwargs={"x": xs, "y": ys, "z": zs, "mode": "lines"},
)
trace_generic = dict(
    backend="generic",
    constructor="scatter3d",
    kwargs={"x": xs, "y": ys, "z": zs, "mode": "lines"},
)
objs.append(magpy.misc.CustomSource(style_label="Matplotlib extra trace"))
objs[-1].style.model3d.add_trace(trace_mpl)

objs.append(magpy.magnet.Cuboid((0,0,1000), [1]*3, style_label="Plotly extra trace"))
objs[-1].style.model3d.add_trace(trace_plotly)

objs.append(magpy.magnet.Cuboid((0,0,1000), [1]*3, style_label="Generic extra trace"))
objs[-1].style.model3d.add_trace(trace_generic)

for i, obj in enumerate(objs):
    obj.rotate_from_angax([10 * i for i in range(N)], "y", start=0, anchor=(0, 0, 10))
    obj.move((0,2*i,0))
for backend in ("matplotlib", "plotly"):
    magpy.show(
        *objs,
        backend=backend,
        style_path_frames=2,
        #style_legend_show=False,
        #style_model3d_showdefault=False,
    )

[Alexboiboi]

@OrtnerMichael ready for review

[OrtnerMichael]

hi @Alexboiboi

---

style_legend_show=True/False works nicely !!! HUGE fan of this !!!

Question: do we have a format option for legend display style ? like we have it in describe ?

---

show commands I don't understand where kwargs are sent. there is also no example.

---

fig commands:

1. Works beautifully with Plotly

2. Not very helpful with Matplotlib because all the styling happens at Axes level. At Figure level there is very little to configure.

magpy.show(s1,s2,sens,
    backend='matplotlib',
    fig_facecolor="pink",
)

We could offer fig sensing to Figure and ax sending to Axes

3. Also not helpful with Pyvista because the Plotter constructor does not accept kwargs - all styles are set via methods/properties. Maybe we can add the styles via the methods ?

magpy.show(s1,s2,sens,
    backend='pyvista',
    fig_background_color="pink",
)

should be done like this

pl = pv.Plotter()
# add Magpylib stuff
pl.background_color = "pink"
pl.show()

---

We really need a set of excellent examples and use-cases. I already outlined them in #650. Lets bring them here in the next comment.

[OrtnerMichael]

List of use-cases - KEEP ADDING

• change background color
• change figure title
• remove axes
• set camera angle

[OrtnerMichael]

passt