面板小部件不更新绘图

Question

我正在尝试在 class 中使用面板小部件来更改情节。下面的示例输出一次绘图，但是当我 select 来自小部件的不同变量时，它不会更新绘图。

import pandas as pd
import param

import holoviews as hv
hv.extension('bokeh')

import panel as pn
pn.extension()

df = pd.DataFrame({'index':[1,2,3,4,5],
                   'signal1':[1,2,3,4,5],
                   'signal2':[5,4,3,2,1]}).set_index('index')

class Plot(param.Parameterized):
    def __init__(self, df):
        self.multi_select = pn.widgets.MultiSelect(name='Vars', value=['signal1'],
                                                   options=['signal1', 'signal2'])
        self.df = df
        self.watcher = self.multi_select.param.watch(self.dashboard, 'value')
        super().__init__()

    def dashboard(self, *events, **kwargs):
        self.plt = hv.Curve(self.df, [self.df.index.name, self.multi_select.value[0]])
        return self.plt

a = Plot(df)
pn.Row(a.multi_select,a.dashboard)

但是，如果我使用参数而不是面板小部件，它会按预期工作。

class Plot2(param.Parameterized):
    multi_select = param.ListSelector(default=['signal1'],
                                      objects=['signal1', 'signal2'])

    df = df

    def dashboard(self, **kwargs):
        self.plt = hv.Curve(self.df, [self.df.index.name, self.multi_select[0]])
        return self.plt

b = Plot2()
pn.Row(b.param,b.dashboard)

我是做错了什么还是根本不可能？谢谢

Answer 1

首先我稍微修改了你的代码。主要变化是：我添加了

Plot.dashboard() 中的 try-except 子句，带有关于事件和变量的打印语句 self.multi_select
使用带有 view = pn.Row(a.multi_select,a.dashboard) 和 view.app() 的外部散景服务器从 try-except-clause 中获取打印件。

完整的情节（）：

class Plot(param.Parameterized):

    def __init__(self, df, **params):
        super(Plot, self).__init__(**params)
        self.multi_select = pn.widgets.MultiSelect(name='Vars', value=['signal1'],
                                                   options=['signal1', 'signal2'])
        self.df = df
        self.watcher = self.multi_select.param.watch(self.dashboard, 'value')


    def dashboard(self, *events, **kwargs):
        try: 
            event = events[0] # get first event in events
            select = events[0].obj # obj is our multi_select

            # see, what event looks like, and if select and self.multi_select are the same
            print(event, '\n', select, '\n', self.multi_select)

        except IndexError: pass

        self.plt = hv.Curve(self.df, [self.df.index.name, self.multi_select.value[0]])
        return self.plt

a = Plot(df)
view = pn.Row(a.multi_select,a.dashboard)
view.app()

因此您的代码可以正常工作，但不会显示更改。

第二：将曲线图与多选图分开并摆弄显示，如果方法 view 取决于更改的参数（此处为用于绘图的列的名称），则曲线仅发生变化：

class Curve_(param.Parameterized):
    col = param.String()

    def __init__(self, df, col, **params):
        super(Curve_, self).__init__(**params)
        self.df = df
        self.col = col

    @param.depends('col', watch=True)
    def view(self): 
        return hv.Curve(self.df, [self.df.index.name, self.col])

t=Curve_(df, 'signal1')
pn.Column(t.view)

结合类 Plot 和 Curve_ 并使用参数 dashboard_trigger 来保持仪表板更新，现在一切都按预期工作。

class Plot(param.Parameterized):
    dashboard_trigger = param.Number(1)

    def __init__(self, df, **params):
        super(Plot, self).__init__(**params)
        self.multi_select = pn.widgets.MultiSelect(name='Vars', value=['signal1'],
                                                   options=['signal1', 'signal2'])
        self.curve = Curve_(df, self.multi_select.value[0])
        self.watcher = self.multi_select.param.watch(self.set_column, 'value')

    def set_column(self, *events):
        self.curve.col = events[0].obj.value[0]
        self.dashboard_trigger = self.dashboard_trigger + 1

    @param.depends('dashboard_trigger', watch=True)
    def dashboard(self, *events, **kwargs):
        return self.curve.view

a = Plot(df)
pn.Row(a.multi_select,a.dashboard)

面板小部件不更新绘图

Panel widgets do not update plot

python

panel

bokeh

holoviews

pyviz