Adding N annotations takes O(N^2) time

[amari-ca]

Plotly is unreasonably slow to add more than a few hundred annotations to a plot. This code

import time

import plotly


fig = plotly.graph_objects.Figure()

NUM_ANNOTATIONS=1000

start_time = time.monotonic()
for i in range(NUM_ANNOTATIONS):
    fig.add_annotation(
        x=i,
        y=time.monotonic() - start_time,
        text=str(i),
    )
    print(i)
fig.update_layout(
    xaxis_range=[0,NUM_ANNOTATIONS],
    yaxis_range=[0,time.monotonic() - start_time],
    xaxis_title_text="# Annotations",
    yaxis_title_text="Cumulative Time [s]",
)
fig.show()

produces this plot

[AlonSpivack]

I've proposed a solution in #5514 that addresses this as part of a broader _as_dict mode
for go.Figure.

The root cause is _add_annotation_like doing self.layout["annotations"] += (new_obj,)
which copies the entire tuple on every call. In _as_dict mode, self._layout["annotations"]
is a plain list, so list.append() gives O(1) per call instead of O(N).

Benchmark results for 200 annotations:

• Default: 2,192ms
• _as_dict mode: 12ms (~189x faster)

For add_vline (which calls both add_shape AND add_annotation):

• Default: 6,550ms
• _as_dict mode: 0.58ms (~11,300x faster)

I'm ready to open a PR. See #5514 for full implementation details and benchmarks.