Plotly：如何對不同的跡線使用移位 colors？

Question

我正在使用 pandas 中的數據在 plotly 中繪制圖表。我有一個字典，其中鍵是日期，例如2020-10-26 ，值是pd.Dataframe ，它有一個time列和 2 個數據列。

.csv格式的數據示例：

datetime,SeriesA,SeriesB
2020-10-23 06:00:03.560996+01:00,0,0
2020-10-23 06:05:03.358855+01:00,0,0
2020-10-23 06:10:02.927629+01:00,0,0
2020-10-23 06:15:02.514076+01:00,0,0
2020-10-23 06:20:03.231765+01:00,0,0
2020-10-23 06:25:02.790677+01:00,0,0
2020-10-23 06:30:04.252621+01:00,0,0
2020-10-23 06:35:02.845816+01:00,0,0
2020-10-23 06:40:03.510711+01:00,0,0
2020-10-23 06:45:03.235390+01:00,0,0
2020-10-23 06:50:02.689276+01:00,0,0
2020-10-23 06:55:03.569083+01:00,0,0
2020-10-23 07:00:08.279973+01:00,0,0
2020-10-23 07:05:02.816612+01:00,0,0
2020-10-23 07:10:03.417112+01:00,2,0
2020-10-23 07:15:03.041706+01:00,2,0
2020-10-23 07:20:02.537567+01:00,3,0
2020-10-23 07:25:03.266320+01:00,5,0
2020-10-23 07:30:02.900625+01:00,5,0
2020-10-23 07:35:03.577497+01:00,5,0
2020-10-23 07:40:03.179229+01:00,7,0
2020-10-23 07:45:02.743349+01:00,10,0
2020-10-23 07:50:03.285283+01:00,11,0
2020-10-23 07:55:02.858642+01:00,11,0
2020-10-23 08:00:03.641784+01:00,10,0
2020-10-23 08:05:03.256219+01:00,10,0
2020-10-23 08:10:02.994154+01:00,9,0
2020-10-23 08:15:02.525091+01:00,9,0
2020-10-23 08:20:03.261529+01:00,9,0
2020-10-23 08:25:02.890515+01:00,9,0
2020-10-23 08:30:03.533912+01:00,9,0
2020-10-23 08:35:03.206510+01:00,7,0
2020-10-23 08:40:02.719998+01:00,7,0
2020-10-23 08:45:03.310976+01:00,7,0
2020-10-23 08:50:02.996382+01:00,4,0
2020-10-23 08:55:02.697139+01:00,7,0
2020-10-23 09:00:03.721763+01:00,7,0
2020-10-23 09:05:03.182091+01:00,7,0
2020-10-23 09:10:02.765971+01:00,8,0
2020-10-23 09:15:03.360654+01:00,8,0
2020-10-23 09:20:02.990297+01:00,10,0
2020-10-23 09:25:02.661501+01:00,11,0
2020-10-23 09:30:03.275004+01:00,14,0
2020-10-23 09:35:02.868574+01:00,14,0
2020-10-23 09:40:03.460406+01:00,15,0
2020-10-23 09:45:02.905538+01:00,18,0
2020-10-23 09:50:02.423377+01:00,20,0
2020-10-23 09:55:02.984470+01:00,24,0
2020-10-23 10:00:02.622322+01:00,24,0
2020-10-23 10:05:03.166971+01:00,26,0
2020-10-23 10:10:02.859788+01:00,27,0
2020-10-23 10:15:03.494884+01:00,27,0
2020-10-23 10:20:03.113251+01:00,28,0
2020-10-23 10:25:02.606379+01:00,29,0
2020-10-23 10:30:03.268436+01:00,30,0
2020-10-23 10:35:02.889803+01:00,35,0
2020-10-23 10:40:03.577297+01:00,37,0
2020-10-23 10:45:03.082449+01:00,37,0
2020-10-23 10:50:02.680122+01:00,37,0
2020-10-23 10:55:03.207217+01:00,40,0
2020-10-23 11:00:03.633144+01:00,40,0
2020-10-23 11:05:03.342871+01:00,39,0
2020-10-23 11:10:02.952711+01:00,40,0
2020-10-23 11:15:02.519122+01:00,42,0
2020-10-23 11:20:03.330055+01:00,42,0
2020-10-23 11:25:02.948754+01:00,44,0
2020-10-23 11:30:02.505908+01:00,43,0
2020-10-23 11:35:03.091022+01:00,42,0
2020-10-23 11:40:02.570567+01:00,38,0
2020-10-23 11:45:03.214173+01:00,38,0
2020-10-23 11:50:02.909830+01:00,38,0
2020-10-23 11:55:02.493098+01:00,40,4
2020-10-23 12:00:03.200475+01:00,41,5
2020-10-23 12:05:02.851384+01:00,45,6
2020-10-23 12:10:03.545193+01:00,45,7
2020-10-23 12:15:03.191269+01:00,42,8
2020-10-23 12:20:02.726863+01:00,45,9
2020-10-23 12:25:03.444214+01:00,47,9
2020-10-23 12:30:03.266972+01:00,49,9
2020-10-23 12:35:02.904323+01:00,53,9
2020-10-23 12:40:03.548200+01:00,53,10
2020-10-23 12:45:03.065578+01:00,49,10
2020-10-23 12:50:02.517495+01:00,49,10
2020-10-23 12:55:03.228082+01:00,48,10
2020-10-23 13:00:03.344813+01:00,53,12
2020-10-23 13:05:02.999236+01:00,54,12
2020-10-23 13:10:02.555469+01:00,54,14
2020-10-23 13:15:03.424480+01:00,62,14
2020-10-23 13:20:03.178713+01:00,62,14
2020-10-23 13:25:02.856786+01:00,64,14
2020-10-23 13:30:03.538784+01:00,66,17
2020-10-23 13:35:03.226000+01:00,65,17
2020-10-23 13:40:02.738893+01:00,65,20
2020-10-23 13:45:03.216859+01:00,68,20
2020-10-23 13:50:02.801997+01:00,71,22
2020-10-23 13:55:03.325790+01:00,73,22
2020-10-23 14:00:03.032139+01:00,76,24
2020-10-23 14:05:02.713786+01:00,77,23
2020-10-23 14:10:03.329970+01:00,84,23
2020-10-23 14:15:02.891348+01:00,81,23
2020-10-23 14:20:02.587012+01:00,84,23
2020-10-23 14:25:03.119747+01:00,82,25
2020-10-23 14:30:03.751337+01:00,82,25
2020-10-23 14:35:03.332507+01:00,88,23
2020-10-23 14:40:03.049341+01:00,85,23
2020-10-23 14:45:02.630885+01:00,85,20
2020-10-23 14:50:03.287446+01:00,86,19
2020-10-23 14:55:02.915400+01:00,86,18
2020-10-23 15:00:07.137616+01:00,85,16
2020-10-23 15:05:02.683116+01:00,82,14
2020-10-23 15:10:03.373464+01:00,85,14
2020-10-23 15:15:03.128489+01:00,88,14
2020-10-23 15:20:02.846043+01:00,86,14
2020-10-23 15:25:03.462925+01:00,87,15
2020-10-23 15:30:03.131532+01:00,92,15
2020-10-23 15:35:02.721796+01:00,97,15
2020-10-23 15:40:03.375007+01:00,102,15
2020-10-23 15:45:03.020813+01:00,101,15
2020-10-23 15:50:02.563966+01:00,91,15
2020-10-23 15:55:03.235962+01:00,89,15
2020-10-23 16:00:03.040490+01:00,87,19
2020-10-23 16:05:02.467856+01:00,90,19
2020-10-23 16:10:03.099486+01:00,92,19
2020-10-23 16:15:02.637764+01:00,94,18
2020-10-23 16:20:03.153888+01:00,94,18
2020-10-23 16:25:02.736936+01:00,91,18
2020-10-23 16:30:03.359875+01:00,93,17
2020-10-23 16:35:02.974661+01:00,90,19
2020-10-23 16:40:02.513056+01:00,91,19
2020-10-23 16:45:03.092756+01:00,92,20
2020-10-23 16:50:02.659552+01:00,96,21
2020-10-23 16:55:03.478308+01:00,96,21
2020-10-23 17:00:04.117017+01:00,91,24
2020-10-23 17:05:02.685190+01:00,95,24
2020-10-23 17:10:03.267694+01:00,96,24
2020-10-23 17:15:03.019771+01:00,101,26
2020-10-23 17:20:02.526978+01:00,105,31
2020-10-23 17:25:03.206001+01:00,106,33
2020-10-23 17:30:02.865237+01:00,106,35
2020-10-23 17:35:03.366034+01:00,107,40
2020-10-23 17:40:02.992485+01:00,114,42
2020-10-23 17:45:02.552713+01:00,108,44
2020-10-23 17:50:03.218854+01:00,116,43
2020-10-23 17:55:02.755765+01:00,112,45
2020-10-23 18:00:03.442851+01:00,112,46
2020-10-23 18:05:03.125732+01:00,113,43
2020-10-23 18:10:02.650996+01:00,113,42
2020-10-23 18:15:03.224235+01:00,115,42
2020-10-23 18:20:02.852754+01:00,114,44
2020-10-23 18:25:03.375879+01:00,110,45
2020-10-23 18:30:03.504858+01:00,110,42
2020-10-23 18:35:03.023102+01:00,107,40
2020-10-23 18:40:02.560902+01:00,103,37
2020-10-23 18:45:03.183863+01:00,101,37
2020-10-23 18:50:02.776229+01:00,102,40
2020-10-23 18:55:03.366546+01:00,99,40
2020-10-23 19:00:05.100330+01:00,94,40
2020-10-23 19:05:02.577656+01:00,93,41
2020-10-23 19:10:03.234291+01:00,93,45
2020-10-23 19:15:02.765871+01:00,89,45
2020-10-23 19:20:03.269868+01:00,92,45
2020-10-23 19:25:02.891858+01:00,89,43
2020-10-23 19:30:03.859379+01:00,86,43
2020-10-23 19:35:03.410564+01:00,89,42
2020-10-23 19:40:03.126319+01:00,85,39
2020-10-23 19:45:02.686578+01:00,82,38
2020-10-23 19:50:03.409481+01:00,80,34
2020-10-23 19:55:03.000850+01:00,78,34
2020-10-23 20:00:04.680330+01:00,81,32
2020-10-23 20:05:03.269432+01:00,81,32
2020-10-23 20:10:02.847406+01:00,83,33
2020-10-23 20:15:03.422214+01:00,79,34
2020-10-23 20:20:03.025468+01:00,78,33
2020-10-23 20:25:02.639656+01:00,75,31
2020-10-23 20:30:03.241510+01:00,71,31
2020-10-23 20:35:02.787490+01:00,66,31
2020-10-23 20:40:03.433893+01:00,65,28
2020-10-23 20:45:03.029779+01:00,67,20
2020-10-23 20:50:02.478445+01:00,63,19
2020-10-23 20:55:03.064874+01:00,63,19
2020-10-23 21:00:11.388007+01:00,59,16
2020-10-23 21:05:02.915554+01:00,59,16
2020-10-23 21:10:02.508213+01:00,54,16
2020-10-23 21:15:03.024323+01:00,53,16
2020-10-23 21:20:02.577168+01:00,51,16
2020-10-23 21:25:03.403608+01:00,50,13
2020-10-23 21:30:03.193242+01:00,47,13
2020-10-23 21:35:03.329020+01:00,47,10
2020-10-23 21:40:02.906741+01:00,45,10
2020-10-23 21:45:02.359574+01:00,38,10
2020-10-23 21:50:03.115074+01:00,37,10
2020-10-23 21:55:02.770450+01:00,34,10
2020-10-23 22:00:03.658104+01:00,34,10
2020-10-23 22:05:03.182939+01:00,34,10
2020-10-23 22:10:02.834897+01:00,34,10
2020-10-23 22:15:03.545781+01:00,33,10
2020-10-23 22:20:03.130197+01:00,31,10
2020-10-23 22:25:02.644800+01:00,31,10
2020-10-23 22:30:03.311895+01:00,30,10
2020-10-23 22:35:02.958333+01:00,30,10
2020-10-23 22:40:02.527824+01:00,30,9
2020-10-23 22:45:03.243267+01:00,29,9
2020-10-23 22:50:02.808424+01:00,25,9
2020-10-23 22:55:03.338346+01:00,24,9

for day, df in data.items():
    day_name = datetime.datetime.strptime(day, "%Y-%m-%d").strftime("%A") # convert str to datetime, then to day
    traces.append(go.Scatter(x=df["time"], y=df["SeriesA"], name=f"{day_name} SeriesA"))
    traces.append(go.Scatter(x=df["time"], y=df["SeriesB"], name=f"{day_name} SeriesB"))

有什么辦法可以將每行 plot 作為逐漸變化的色標中的不同顏色？ 理想情況下，我希望星期一的所有痕跡都是不同的紅色陰影，而星期二的每條痕跡都是不同的黃色陰影。

我能想到的唯一方法是手動將十六進制值輸入列表，然后為一周中的每一天創建一個索引。 然后在繪制軌跡時：

# If statement for every day of the week
if day_name == "Monday":
    color = red_hex_values[monday_idx]
    monday_idx += 1

traces.append(go.Scatter(x=df["time"], y=df["SeriesA"], name=f"{day_name} SeriesA", color=color))

不過，這似乎是一種非常混亂的方法。

Answer 1

介紹

您的問題可以使用更好的焦點，因為它實際上是兩個問題：

1. 如何制作自定義色標
2. 如何在 plotly 的工作日使用移位 colors。

我冒昧地或多或少地專門關注后一部分。 盡管我將簡要介紹前者的一些選項 - 如果您願意，我們可以稍后再討論。

回答

這是一個有趣的問題，我想我已經找到了一個可行的解決方案，通過循環迭代指定的色標作為值存儲在字典daymap中，其中鍵是工作日名稱：

daymap = {'Monday':cycle(px.colors.sequential.Reds),
           'Tuesday':cycle(px.colors.sequential.Greens),
         }

px.colors.sequential.Reds和px.colors.sequential.Greens是內置的 plotly 色標。 你特別要求黃色。 但現在必須這樣做。

我已將其設置為在 plot 中分配您設置的任何定義的weekday, color scale對，並且為您選擇不定義的所有工作日名稱提供您使用自定義 function colfx()指定的單一自定義顏色:

def colfx(daymap, day_name, default_color):
    if day_name in daymap.keys():
        return next(daymap[day_name])
    else:
        return default_color

下面的完整代碼片段和一個 60 天的隨機數據樣本將產生下面的 Plot 1，除其他一些細節外，還使用上面提到的規范。

Plot 1

Plot 2 是通過在daymap中添加另一個鍵值對並將colfx()中的默認顏色更改為'rgba(150,150,150, 0.2)' 。 我發現使用rgba colors 作為默認顏色很實用，因為最后一項0.2可以讓您輕松設置顏色的不透明度。

Plot 2

完整代碼

import pandas as pd
import numpy as np
import plotly.graph_objects as go
import plotly.express as px
from itertools import cycle

# data
obs = 60
dates = pd.date_range(start='2020-01-01', periods=obs)
random = np.random.RandomState(0)
df = pd.DataFrame({'data': np.random.uniform(low=0, high=100, size=obs).tolist()}, index=dates)

# custom function to assign next color in a color sequence
# for each new day defined in the dict daymap
def colfx(daymap, day_name, default_color):
    if day_name in daymap.keys():
        return next(daymap[day_name])
    else:
        return default_color

# color specifications for weekday
daymap = {'Monday':cycle(px.colors.sequential.Reds),
           'Tuesday':cycle(px.colors.sequential.Greens),
           'Friday':cycle(px.colors.sequential.thermal)
         }
    
# container for weekday names added to the figure legend
# if the weekday names has ALREADY been added to the legend,
# then no further additions are made
fignames = []

fig=go.Figure()
for row in df.iterrows():
    day_name=row[0].strftime("%A")
    fignames = [d.name for d in fig.data]
    fig.add_bar(x=[row[0]],
                y=[row[1]['data']],
                marker_color=colfx(daymap=daymap, day_name = day_name, default_color='rgba(150,150,150, 0.2)'),
                name=day_name,
                legendgroup = day_name,
                showlegend = False if day_name in fignames else True
               )
    
fig.show()