Plotly Express添加OLS趋势线时出现ImportError且无法显示趋势线的问题求助

我最近在用Plotly Express绘制散点图并添加OLS趋势线时碰到了棘手的问题：只要在代码里加上trendline='ols'参数，程序就会抛出一大段ImportError错误，趋势线完全无法显示；但如果把趋势线相关的代码去掉，散点图就能正常渲染，也不会出现任何错误。我非常希望能让代码的所有功能都正常工作，麻烦各位帮忙分析一下问题所在！

我的环境 #

• 设备：Surface Pro 8
• 系统：Windows 11（已完全更新）
• 已安装的相关库：statsmodels、matplotlib、scipy，且已更新pip

问题代码 #

import pandas as pd
import plotly.express as px

# Load the dataset
df = pd.read_csv('insurance.csv')

# Create scatter plot with trendline and hover data
fig = px.scatter(
    df,
    x='bmi',
    y='charges',
    color='smoker',
    trendline='ols',  # Add linear regression trendline
    hover_data=['age'],  # Include age in hover tooltip
    title='BMI vs Charges by Smoker Status with Trendline',
    labels={'bmi': 'BMI', 'charges': 'Charges ($)', 'smoker': 'Smoker'},
    opacity=0.6
)

fig.show()

报错信息 #

ImportError Traceback (most recent call last)
Cell In[2], line 13
      8 df = pd.read_csv('insurance.csv')
     11 # Create scatter plot with trendline and hover data
---> 13 fig = px.scatter(
     14 
     15 df,
     16 
     17 x='bmi',
     18 
     19 y='charges',
     20 
     21 color='smoker',
     22 
     23 trendline='ols', # Add linear regression trendline
     24 
     25 hover_data=['age'], # Include age in hover tooltip
     26 
     27 title='BMI vs Charges by Smoker Status with Trendline',
     28 
     29 labels={'bmi': 'BMI', 'charges': 'Charges ($)', 'smoker': 'Smoker'},
     30 
     31 opacity=0.6
     32 
     33 )
     36 fig.show()

File ~\AppData\Local\Programs\Python\Python313\Lib\site-packages\plotly\express\_chart_types.py:69, in scatter(data_frame, x, y, color, symbol, size, hover_name, hover_data, custom_data, text, facet_row, facet_col, facet_col_wrap, facet_row_spacing, facet_col_spacing, error_x, error_x_minus, error_y, error_y_minus, animation_frame, animation_group, category_orders, labels, orientation, color_discrete_sequence, color_discrete_map, color_continuous_scale, range_color, color_continuous_midpoint, symbol_sequence, symbol_map, opacity, size_max, marginal_x, marginal_y, trendline, trendline_options, trendline_color_override, trendline_scope, log_x, log_y, range_x, range_y, render_mode, title, subtitle, template, width, height)
     14 def scatter(
     15 data_frame=None,
     16 x=None,
   (...)
     63 height=None,
     64 ) -> go.Figure:
     65 """
     66 In a scatter plot, each row of `data_frame` is represented by a symbol
     67 mark in 2D space.
     68 """
---> 69 return make_figure(args=locals(), constructor=go.Scatter)

File ~\AppData\Local\Programs\Python\Python313\Lib\site-packages\plotly\express\_core.py:2674, in make_figure(args, constructor, trace_patch, layout_patch)
   2671 elif args["ecdfnorm"] == "percent":
   2672     group = group.with_columns((nw.col(var) / group_sum) * 100.0)
-> 2674 patch, fit_results = make_trace_kwargs(
   2675     args, trace_spec, group, mapping_labels.copy(), sizeref
   2676 )
   2677 trace.update(patch)
   2678 if fit_results is not None:

File ~\AppData\Local\Programs\Python\Python313\Lib\site-packages\plotly\express\_core.py:430, in make_trace_kwargs(args, trace_spec, trace_data, mapping_labels, sizeref)
    427 trace_patch["x"] = trace_patch["x"].to_numpy()
    429 trendline_function = trendline_functions[attr_value]
--> 430 y_out, hover_header, fit_results = trendline_function(
    431     args["trendline_options"],
    432     sorted_trace_data.get_column(args["x"]), # narwhals series
    433     x.to_numpy(), # numpy array
    434     y.to_numpy(), # numpy array
    435     args["x"],
    436     args["y"],
    437     non_missing.to_numpy(), # numpy array
    438 )
    439 assert len(y_out) == len(
    440     trace_patch["x"]
    441 ), "missing-data-handling failure in trendline code"
    442 trace_patch["y"] = y_out

File ~\AppData\Local\Programs\Python\Python313\Lib\site-packages\plotly\express\trendline_functions\__init__.py:42, in ols(trendline_options, x_raw, x, y, x_label, y_label, non_missing)
     36 if k not in valid_options:
     37     raise ValueError(
     38         "OLS trendline_options keys must be one of [%s] but got '%s'"
     39         % (", ".join(valid_options), k)
     40     )
---> 42 import statsmodels.api as sm
     44 add_constant = trendline_options.get("add_constant", True)
     45 log_x = trendline_options.get("log_x", False)

File ~\AppData\Local\Programs\Python\Python313\Lib\site-packages\statsmodels\api.py:76
      1 __all__ = [
      2     "BayesGaussMI",
      3     "BinomialBayesMixedGLM",
   (...)
     72     "__version_info__"
     73 ]
---> 76 from . import datasets, distributions, iolib, regression, robust, tools
     77 from .__init__ import test
     78 from statsmodels._version import (
     79     version as __version__, version_tuple as __version_info__
     80 )

File ~\AppData\Local\Programs\Python\Python313\Lib\site-packages\statsmodels\distributions\__init__.py:7
      2 from .empirical_distribution import (
      3     ECDF, ECDFDiscrete, monotone_fn_inverter, StepFunction
      4 )
      5 from .

注：报错信息末尾有截断，但核心问题出现在导入statsmodels.api时

我已经确认自己安装了statsmodels，但还是出现这个错误，有没有哪位大佬能帮我找出问题所在呀？

内容来源于stack exchange