如何使用Bokeh为一组线条应用颜色映射（附Matplotlib参考）

Got it! Converting your matplotlib color-mapped line plot to Bokeh is straightforward—here's how you can replicate that exact functionality step by step:

Key Concepts to Translate #

In matplotlib, you use cm(colors_l[i]) to map a numerical value to a color from the Jet colormap. In Bokeh, we'll use LinearColorMapper to handle this same mapping, or directly index into the Jet palette if you prefer.

Full Code Example #

First, make sure you have the necessary imports, then adapt your logic to Bokeh's API:

import numpy as np
from bokeh.plotting import figure, show
from bokeh.models import LinearColorMapper
from bokeh.palettes import Jet256

# Assume your data and state_means variables are already defined
step = 15
xi = np.linspace(data[data.columns[0]].min(), data[data.columns[0]].max(), 2)
colors_l = np.linspace(0.1, 1, len(state_means[::step]))

# 1. Set up the color mapper (equivalent to plt.cm.get_cmap('jet'))
color_mapper = LinearColorMapper(palette="Jet256", low=0.1, high=1.0)

# 2. Create your Bokeh figure
p = figure(title="Color-Mapped Lines with Bokeh", x_axis_label=data.columns[0], y_axis_label="Y Value")

# 3. Loop through your state means and plot each line
for i, beta in enumerate(state_means[::step]):
    # Calculate the line's y-values (same as your original code)
    y_values = beta[0] * xi + beta[1]
    # Get the mapped color for this line (equivalent to cm(colors_l[i]))
    line_color = color_mapper.to_rgba(colors_l[i])
    # Plot the line with matching alpha and line width
    p.line(xi, y_values, alpha=0.2, line_width=1, color=line_color)

# Show the plot
show(p)

Alternative: Direct Palette Indexing #

If you'd rather work directly with the palette list (no mapper), you can normalize your colors_l values to index into the Jet256 palette:

palette = Jet256
# Normalize colors_l to 0-1 range (since your values go from 0.1 to 1)
normalized_colors = (colors_l - 0.1) / (1.0 - 0.1)
# Convert to integer indices for the palette
color_indices = (normalized_colors * (len(palette) - 1)).astype(int)

# Then in your loop:
line_color = palette[color_indices[i]]

Notes #

• Bokeh's Jet256 palette matches matplotlib's jet colormap closely, with 256 distinct colors.
• The to_rgba method returns an RGBA tuple (e.g., (255, 0, 0, 1.0)), which Bokeh accepts directly as a color value.
• Adjust the low and high parameters of LinearColorMapper if your colors_l range changes—this ensures your values map correctly across the full palette.

内容的提问来源于stack exchange，提问作者Parikshit Bhinde