Histograms and hover

histplot takes counts and bin edges, like mplhep.histplot, and draws one or several histograms at once:

php.histplot(fig, H, bins)                                     # step outline, edges down to zero
php.histplot(fig, [H1, H2], bins, label=["Z", "tt"], stack=True, histtype="fill")
php.histplot(fig, data, bins, yerr=True, histtype="errorbar", color="black", label="Data")
php.histplot(fig, H, bins, density=True)                       # normalised to unit area
php.histplot(fig, H, bins, yerr=err_array)                     # your own errors

keyword	meaning
histtype	"step" (default), "fill", "errorbar"
yerr	True for √N, or an array / list of arrays
stack	cumulative sums; fills are painted top-down so every layer stays visible, the legend keeps your order
density	divide by the integral
edges	False to drop the vertical edges at the first and last bin
label, color, linewidth	one value or one per histogram
row=2	draw into the lower panel of a ratio figure

Line widths and marker sizes are mplhep's points converted to pixels at 100 dpi, so a step histogram is 1.5 pt wide here as there.

Hover

Every bin answers the cursor with its edges, content and error:

[80, 85)
412 ± 20.3

The drawn outline, fill or error segments never respond themselves (hoverinfo="skip"); a transparent marker sits at each bin centre and carries the customdata and hovertemplate. That is why hovering a stacked histogram tells you about the layer under the cursor and not about the outline above it. php.plot.bin_hover(edges, values, err, name) returns the same customdata/hovertemplate pair for a trace you build yourself, and any Plotly hovertemplate you pass through **kw overrides it.

2D histograms

H, xe, ye = np.histogram2d(x, y, bins=[40, 30])
php.hist2dplot(fig, H, xe, ye, colorscale="Viridis")

H is indexed [x, y] as numpy.histogram2d returns it; the heatmap shows bin edges on hover.