Bug summary

Right now, it looks like the type hints for Axes.annotate only accept a tuple of floats as input for the xy and xycoords arguments, but that is too narrow. Specifically, if I've plotted something with a datetime on the x axis (a not-unusual practice), I am most likely going to specify the annotation position with an acceptable datetime-like object.

There are no runtime issues, but both mypy and basedpyright complain about the annotate line in the following code (ignore the messy output, it's just for example purposes). t is truly a pd.Timestamp, but even if I try using a native datetime object the problem occurs.

Code for reproduction

from typing import cast

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd

idx = pd.date_range(start='01/01/2014', end='03/01/2014', freq='1h')
n = len(idx)
rng = np.random.default_rng()
data = rng.normal(size=n)
df = pd.DataFrame({"data": data}, index=idx)

daily = df.resample("D")
maxs = daily.max()
mins = daily.min()

ax = maxs.plot()
mins.plot(ax=ax)
for t, x in maxs["data"].items():
    t = cast(pd.Timestamp, t)
    x = cast(float, x)
    ax.annotate(str(x), (t, x))

plt.show()

Actual outcome

$ basedpyright dummy1.py
/home/andrea/Code/playground/dummy1.py
/home/andrea/Code/playground/dummy1.py:26:26 - error: Argument of type "tuple[Timestamp, float]" cannot be assigned to parameter "xy" of type "tuple[float, float]" in function "annotate"
"Timestamp" is not assignable to "float" (reportArgumentType)
1 error, 0 warnings, 0 notes

$ mypy dummy1.py
dummy1.py:26: error: Argument 2 to "annotate" of "Axes" has incompatible type "tuple[Timestamp, float]"; expected "tuple[float, float]" [arg-type]
Found 1 error in 1 file (checked 1 source file)

Expected outcome

No type errors.

Additional information

Versions:
python: 3.12
matplotlib: 3.10.3
basedpyright: 1.31.0
mypy: 1.17.0

Operating system

Debian (Linux Mint)

Matplotlib Version

3.10.3

Matplotlib Backend

qtagg

Python version

3.12

Jupyter version

No response

Installation

pip