|
说明
此篇为 一次python画图经历——涉及 matplotlib 的折线图，散点图，图中标注文字信息，设置横坐标密度，调整横坐标旋转角度，文字大小、各部分颜色、保存形式、美化调整 的延续，当时自以为很好地处理了标注信息的问题，结果发现，当需要标注的点很多很密集的时候，画出来的图会很糟糕：
 因此重新想了一个办法，加入鼠标交互功能，当鼠标指针指向标注点时，将标注信息显示出来，不指向时不显示标注信息，以下为实现过程。
标注信息动态显示
数据
代码
import pandas as pd
import matplotlib.pyplot as plt
import matplotlib.ticker as ticker
# ---------- 数据预处理 ----------
file_path = './example_csdn_processed.csv'
df_all = pd.DataFrame(pd.read_csv(file_path, encoding='gbk'))
df_all = df_all.set_index(pd.DatetimeIndex(pd.to_datetime(df_all.time)))
x = df_all['time']
y1 = df_all['data']
y2 = df_all["buy"]
y3 = df_all["sell"]
y_condition = list(df_all["操作"])
text1 = []
for i in range(len(x)):
if (df_all['操作'][i] != 0) and (df_all['操作'][i] == "买"):
str_info = '价格: ' + str(df_all['价格'][i]) + '\n' + '数量: ' + str(df_all['数量'][i]) + '\n' + '操作: ' + str(df_all['操作'][i])
text1.append(str_info)
else:
text1.append(" ")
text2 = []
for i in range(len(x)):
if (df_all['操作'][i] != 0) and (df_all['操作'][i] == "卖"):
str_info = '价格: ' + str(df_all['价格'][i]) + '\n' + '数量: ' + str(df_all['数量'][i]) + '\n' + '操作: ' + str(df_all['操作'][i])
text2.append(str_info)
else:
text2.append(" ")
# ---------- 画图 ----------
fig, ax = plt.subplots()
# 折线图
ax.plot(x, y1, color='royalblue', lw=2.5, label='data')
# 折线图上的散点
ax.scatter(x, y2, marker='o', c='darkgreen', label='买')
ax.scatter(x, y3, marker='o', c='firebrick', label='卖')
# 一些小设置
# 设置 x 轴显示密度
tick_spacing = 8
ax.xaxis.set_major_locator(ticker.MultipleLocator(tick_spacing))
# 设置 x 坐标轴标签的显示内容和大小
plt.xlabel('时间', fontsize=16)
# 设置 x 坐标轴刻度的旋转方向和大小
plt.xticks(rotation=90, fontsize=16)
# 设置 y 坐标轴刻度大小
plt.yticks(fontsize=18)
# 坐标轴中文显示
plt.rcParams['font.sans-serif'] = ['SimHei']
# 调整图的位置
plt.subplots_adjust(top=0.9, bottom=0.22)
# 设置尺寸
fig.set_size_inches(24, 12)
# po_annotation1 为 ‘买’ 时需要显示的标注信息
# 该 list 内部存放多个子 list，每个子 list 为 [标注点的坐标, 标注]
po_annotation1 = []
for i in range(len(x)):
# 标注点的坐标
point_x = x[i]
point_y = y2[i]
point, = plt.plot(point_x, point_y, 'o', c='darkgreen')
# 标注框偏移量
offset1 = 80
offset2 = 80
# 标注框
bbox1 = dict(boxstyle="round", fc='lightgreen', alpha=0.6)
# 标注箭头
arrowprops1 = dict(arrowstyle="->", connectionstyle="arc3,rad=0.")
# 标注
annotation = plt.annotate(text1[i], xy=(x[i], y2[i]), xytext=(-offset1, offset2), textcoords='offset points',
bbox=bbox1, arrowprops=arrowprops1, size=15)
# 默认鼠标未指向时不显示标注信息
annotation.set_visible(False)
po_annotation1.append([point, annotation])
po_annotation2 = []
for i in range(len(x)):
# 标注点的坐标：
point_x = x[i]
point_y = y3[i]
point, = plt.plot(point_x, point_y, 'o', c='firebrick')
# 标注框偏移量
offset1 = 30
offset2 = 80
# 标注框
bbox2 = dict(boxstyle="round", fc='salmon', alpha=0.6)
# 标注箭头
arrowprops2 = dict(arrowstyle="->", connectionstyle="arc3,rad=0.")
# 标注
annotation = plt.annotate(text2[i], xy=(x[i], y3[i]), xytext=(-offset1, offset2), textcoords='offset points',
bbox=bbox2, arrowprops=arrowprops2, size=15)
# 默认鼠标未指向时不显示标注信息
annotation.set_visible(False)
po_annotation2.append([point, annotation])
# 定义鼠标响应函数
def on_move(event):
visibility_changed = False
for point, annotation in po_annotation1:
should_be_visible = (point.contains(event)[0] == True)
if should_be_visible != annotation.get_visible():
visibility_changed = True
annotation.set_visible(should_be_visible)
for point, annotation in po_annotation2:
should_be_visible = (point.contains(event)[0] == True)
if should_be_visible != annotation.get_visible():
visibility_changed = True
annotation.set_visible(should_be_visible)
if visibility_changed:
plt.draw()
# 鼠标移动事件
on_move_id = fig.canvas.mpl_connect('motion_notify_event', on_move)
plt.show()
效果
|