Python3中adjust_text调整文本异常，如何修复重叠标签？

问题描述 #

使用Matplotlib绘制散点图时，手动添加的文本标签出现严重重叠；尝试用adjust_text调整后，标签被集体移至左下角，完全失去标注作用。

原始重叠代码：

import matplotlib.pyplot as plt
from adjustText import adjust_text
x = [12,471,336,1300]
y = [2,5,4,11]
z = [0.1,0.2,0.3,0.4]
im = plt.scatter(x, y, c = z, cmap = "gist_rainbow", alpha = 0.5)
plt.colorbar(im)
texts = []
texts.append(plt.text(783, 7.62372448979592, 'TRL1'))
texts.append(plt.text(601, 6.05813953488372, 'CFT1'))
texts.append(plt.text(631, 4.28164556962025, 'PTR3'))
texts.append(plt.text(665, 7.68018018018018, 'STT4'))
texts.append(plt.text(607, 5.45888157894737, 'RSC9'))
texts.append(plt.text(914, 4.23497267759563, 'DOP1'))
texts.append(plt.text(612, 7.55138662316476, 'SEC8'))
texts.append(plt.text(766, 4.1264667535854, 'ATG1'))
texts.append(plt.text(681, 3.80205278592375, 'TFC3'))
plt.show()

添加adjust_text后失效的代码：

import matplotlib.pyplot as plt
from adjustText import adjust_text
x = [12,471,336,1300]
y = [2,5,4,11]
z = [0.1,0.2,0.3,0.4]
im = plt.scatter(x, y, c = z, cmap = "gist_rainbow", alpha = 0.5)
plt.colorbar(im)
data = [
    (783, 7.62372448979592, 'TRL1'),
    (601, 6.05813953488372, 'CFT1'),
    (631, 4.28164556962025, 'PTR3'),
    (665, 7.68018018018018, 'STT4'),
    (607, 5.45888157894737, 'RSC9'),
    (914, 4.23497267759563, 'DOP1'),
    (612, 7.55138662316476, 'SEC8'),
    (766, 4.1264667535854, 'ATG1'),
    (681, 3.80205278592375, 'TFC3')
]

texts = [plt.text(x, y, l) for x, y, l in data]
adjust_text(texts)
plt.savefig('adjust.text.png', bbox_inches='tight', pad_inches = 0.1)

解决方案 #

问题核心是adjust_text默认不知道需要避开散点元素，导致算法将标签移至空白角落。需明确指定要避开的散点，并调整参数优化布局：

修正后的代码 #

import matplotlib.pyplot as plt
from adjustText import adjust_text

# 散点数据
x_scatter = [12, 471, 336, 1300]
y_scatter = [2, 5, 4, 11]
z = [0.1, 0.2, 0.3, 0.4]

im = plt.scatter(x_scatter, y_scatter, c=z, cmap="gist_rainbow", alpha=0.5)
plt.colorbar(im)

# 标签数据
data = [
    (783, 7.62372448979592, 'TRL1'),
    (601, 6.05813953488372, 'CFT1'),
    (631, 4.28164556962025, 'PTR3'),
    (665, 7.68018018018018, 'STT4'),
    (607, 5.45888157894737, 'RSC9'),
    (914, 4.23497267759563, 'DOP1'),
    (612, 7.55138662316476, 'SEC8'),
    (766, 4.1264667535854, 'ATG1'),
    (681, 3.80205278592375, 'TFC3')
]

texts = [plt.text(x, y, l) for x, y, l in data]

# 关键配置：指定避开的散点，添加箭头，调整排斥力与距离
adjust_text(
    texts,
    x=x_scatter,
    y=y_scatter,
    arrowprops=dict(arrowstyle="-", color='gray', lw=0.5),  # 添加箭头关联原位置
    expand_points=(1.2, 1.2),  # 增大标签与散点的间距
    force_text=0.8,  # 增强标签间的排斥力，避免重叠
    force_objects=0.6,  # 增强标签与散点的排斥力
    only_move={'points':'y', 'text':'y'}  # 可选：限制移动方向，避免过度偏移
)

plt.savefig('adjust.text.png', bbox_inches='tight', pad_inches=0.1)
plt.show()

关键调整说明 #

1. 指定避开对象：通过x和y参数传入散点坐标，让adjust_text明确需要避开的元素，避免标签被移至无关区域。
2. 添加箭头：arrowprops配置箭头，让标签与原始标注位置保持关联，可读性更强。
3. 优化布局参数：
   • expand_points：扩大标签与散点的安全距离
   • force_text/force_objects：调整标签间、标签与散点间的排斥力度，平衡布局与可读性
4. 可选限制移动方向：only_move参数可限制标签仅在特定方向移动，防止过度偏离原位置。

内容的提问来源于stack exchange，提问作者con