Bug: pie(), boxplot(), stem(), surf(), and contour() crash on size-1 inputs due to unconditional np.squeeze()

Bug Description

Several plotting methods crash with AssertionError when given size-1 inputs (e.g. a single-slice pie chart or a single-row surface plot). The root cause is an unconditional np.squeeze() call that collapses valid dimensions before the ndim assertion runs.

Steps to Reproduce

import visdom
import numpy as np

viz = visdom.Visdom()

# 1. Single-slice pie chart (e.g. "100% complete")
viz.pie(X=[100])
# AssertionError: X should be one-dimensional

# 2. Single-point boxplot
viz.boxplot(X=[10.5])
# AssertionError: X should be one or two-dimensional

# 3. Single-point stem plot
viz.stem(X=[5.0])
# AssertionError: X should be one or two-dimensional

# 4. Single-row surface plot (1×5 grid)
viz.surf(X=np.ones((1, 5)))
# AssertionError: X should be two-dimensional

# 5. Single-column contour plot (5×1 grid)
viz.contour(X=np.ones((5, 1)))
# AssertionError: X should be two-dimensional

Multi-element inputs work fine for all methods (e.g. pie([30, 70]), surf(np.ones((5, 5)))).

Expected Behavior

• pie(X=[100]) should render a single-slice pie chart.
• boxplot(X=[10.5]) should render a single-observation boxplot.
• stem(X=[5.0]) should render a single stem.
• surf(X=np.ones((1, 5))) and contour(X=np.ones((5, 1))) should render the 2D matrix as-is (valid Plotly z data).

Root Cause

In py/visdom/__init__.py, the affected methods call X = np.squeeze(X) unconditionally before asserting dimensionality:

By contrast, violin() (L4498) uses the guarded pattern and handles size-1 inputs correctly:

X = np.asarray(X)
if X.ndim > 2:
    X = np.squeeze(X)

Similarly, heatmap() (L3219) does not call np.squeeze() at all, so heatmap(np.ones((1, 5))) works fine — even though it accepts the same 2D matrix format as surf() and contour().

Suggested Fix

Replace unconditional np.squeeze(X) with the guarded pattern already used by violin():

For pie() (L3946):

X = np.asarray(X)
if X.ndim > 1:
    X = np.squeeze(X)
if X.ndim == 0:
    X = X.reshape(1)

For boxplot() (L3621) and stem() (L3857):

X = np.asarray(X)
if X.ndim > 2:
    X = np.squeeze(X)

For _surface() (L3672):

X = np.asarray(X)
if X.ndim > 2:
    X = np.squeeze(X)

Environment

• Visdom version: 0.2.4 (current dev branch)
• Python: 3.12+
• NumPy: any version (core np.squeeze behavior)
• OS: macOS / Linux / Windows (platform-independent)