如何为Pandas绘图标签添加LaTeX文本并精准控制颜色？

When working with pandas df.plot() and needing to add transparent LaTeX text above each column name in the legend (with column names matching their line colors), you can use a custom matplotlib legend handler to achieve this without plotting each line individually. Here's how to do it efficiently, even for multiple DataFrames:

Step-by-Step Implementation #

First, let's set up the necessary imports and sample DataFrame:

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from matplotlib.legend_handler import HandlerBase
from matplotlib.text import Text

# Optional: Enable LaTeX rendering for better math text appearance (requires LaTeX installed)
plt.rcParams['text.usetex'] = True

# Sample DataFrame
df = pd.DataFrame(np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9]]), columns=['a', 'b', 'c'])

Custom Legend Handler #

We'll create a handler that draws two text elements per legend entry:

1. Transparent LaTeX text ($V_{sd}$) above the column name
2. The column name colored to match its corresponding line

class CustomLegendHandler(HandlerBase):
    def create_artists(self, legend, orig_handle, xdescent, ydescent, width, height, fontsize, trans):
        # LaTeX text: positioned above the column name, fully transparent
        latex_label = Text(
            xdescent + width/2,  # Horizontal center
            ydescent + height + fontsize/2,  # Position above the column name
            r'$V_{sd}$',
            ha='center', va='bottom',
            fontsize=fontsize,
            alpha=0  # Set to 0 for full transparency; adjust to 0.3 for semi-transparent
        )
        
        # Column name: colored to match the line, centered vertically
        column_label = Text(
            xdescent + width/2,
            ydescent + height/2,
            orig_handle.get_label(),  # Get the original column name
            ha='center', va='center',
            fontsize=fontsize,
            color=orig_handle.get_color()  # Match the line's color
        )
        
        return [latex_label, column_label]

Plot with Custom Legend #

Now, plot your DataFrame and replace the default legend with our custom one:

# Plot the DataFrame
ax = df.plot()

# Update the legend using our custom handler
ax.legend(
    handles=ax.get_lines(),  # Use the existing line objects from the plot
    handler_map={plt.Line2D: CustomLegendHandler()}  # Apply our handler to all Line2D elements
)

plt.show()

Reusable Function for Multiple DataFrames #

To avoid repeating code for many DataFrames, wrap the logic in a function:

def plot_with_latex_labels(df, latex_text=r'$V_{sd}$', latex_alpha=0):
    """
    Plot a DataFrame with custom legend labels (LaTeX text above column names).
    
    Args:
        df: pandas DataFrame to plot
        latex_text: LaTeX string to add above column names
        latex_alpha: Transparency of the LaTeX text (0 = fully transparent)
    
    Returns:
        matplotlib Axes object
    """
    ax = df.plot()
    
    class CustomLegendHandler(HandlerBase):
        def create_artists(self, legend, orig_handle, xdescent, ydescent, width, height, fontsize, trans):
            latex_label = Text(
                xdescent + width/2,
                ydescent + height + fontsize/2,
                latex_text,
                ha='center', va='bottom',
                fontsize=fontsize,
                alpha=latex_alpha
            )
            column_label = Text(
                xdescent + width/2,
                ydescent + height/2,
                orig_handle.get_label(),
                ha='center', va='center',
                fontsize=fontsize,
                color=orig_handle.get_color()
            )
            return [latex_label, column_label]
    
    ax.legend(handles=ax.get_lines(), handler_map={plt.Line2D: CustomLegendHandler()})
    return ax

# Usage example for any DataFrame
ax = plot_with_latex_labels(df)
plt.show()

Key Notes #

• LaTeX Support: If you don't have LaTeX installed, matplotlib's built-in mathtext will still render $V_{sd}$ correctly (just remove the plt.rcParams['text.usetex'] = True line).
• Transparency Adjustment: Change alpha=0 to a value between 0 and 1 if you want the LaTeX text to be semi-transparent instead of fully invisible.
• Color Matching: The column name color automatically matches the line color from pandas' default plotting style (blue for first column, orange for second, etc.).

内容的提问来源于stack exchange，提问作者Ivanovitch