如何更新IntSlider和InsSlider box自动适配样本量

Question

我真的希望我能找到解决问题的办法。我一直在努力寻找一个，但看起来我肯定被卡住了。

我有下面的一段代码，可以生成两种营养素的交互式散点图。

它会根据下拉框中的一对营养素进行更新，还会根据另一个下拉框中的产品类型进行更新，更重要的是会根据所选人群的 nsamples 子集进行更新（如果为“全部”，则整个人群被选中）。

问题是 nsample 的数量取决于每种产品类型的样本数量或如果选择“全部”则取决于总人口。我正在努力寻找如何可靠地传递给 IntSlider 及其框，即在下拉列表中选择产品类型时更新的最大样本数。

非常感谢

class App:
    
    def __init__(self, df):
        self._df = df
        self._x_dropdown = self._create_indicator_dropdown(all_nutriments, 1, "Nutriment1")
        self._y_dropdown = self._create_indicator_dropdown(all_nutriments, 0, "Nutriment2")        
        self._pnns_dropdown = self._create_indicator_dropdown(all_products, 0, "Product type")
        self._plot_container = widgets.Output()
        if self._pnns_dropdown.value == "All":
            vmax = df.shape[0]
        else:
            vmax = df[df["pnns_groups_2"] == self._pnns_dropdown.value].shape[0]
        self._nsample_slider, self._nsample_slider_box = self._create_sample_slider(200, 100, vmax, 100)

        _app_container = widgets.VBox([
            widgets.HBox([self._x_dropdown, self._y_dropdown]),
            self._plot_container,
            self._nsample_slider_box,
            self._pnns_dropdown], 
            layout=widgets.Layout(align_items='center', flex='3 0 auto'))
        self.container = widgets.VBox([
            widgets.HTML(('<h1>Nutriment indicators for product categories</h1>'), 
                         layout=widgets.Layout(align_items='center', margin='1 0 5em 0')),
            widgets.HBox([_app_container, 
                          widgets.HTML(USAGE, layout=widgets.Layout(margin='0 0 0 2em'))])], 
            layout=widgets.Layout(flex='1 1 auto', margin='0 auto 0 auto', max_width='1024px'))
        self._update_app()
        
    @classmethod
    def from_url(cls, url):
        df = pd.read_csv(url, sep=";")
        return cls(df)
        
    def _create_indicator_dropdown(self, indicators, initial_index, description):
        dropdown = widgets.Dropdown(options=indicators, value=indicators[initial_index], description=description)
        dropdown.observe(self._on_change, names=['value'])
        return dropdown
    
    def _create_sample_slider(self, value, min_sample, max_sample, step):
        sample_slider_label = widgets.Label('Number of samples: ')
        sample_slider = widgets.IntSlider(value=value, min=min_sample, max=max_sample, step=step,
                                          layout=widgets.Layout(width='500px'))
        sample_slider.observe(self._on_change, names=['value'])        
        sample_slider_box = widgets.HBox([sample_slider_label, sample_slider])
        sample_slider_box.observe(self._on_change, names=['value'])
#        sample_slider.observe(self._update_nsample, names=['value'])
        return sample_slider, sample_slider_box

#    def _update_nsample(change):
#        ns = change.new
#        self._nsample_slider.max = ns

    def _create_plot(self, x_indicator, y_indicator, pnns_group, nsample):
        if pnns_group == "All":
            df = self._df
        else:
            df = self._df[self._df['pnns_groups_2'] == pnns_group]
        xs = df[x_indicator]
        ys = df[y_indicator]
        plt.rcParams["figure.figsize"] = [12, 12]
        plt.rcParams.update({'font.size': 18})
        fig,ax = plt.subplots()
        colorpalette=["#008a4b", "#7fc241","#feca07","#f58221","#ef3e23"]
        sns.set_palette(sns.color_palette(colorpalette))
        sns.scatterplot(x=xs, y=ys, data=df.sample(nsample), 
                        hue="nutriscore_grade", s=80, hue_order=["a","b","c","d","e"],alpha=0.9)
        ax.set_xlabel(x_indicator.split("_")[0].capitalize() + " content (g) for 100g" if x_indicator != "energy_100g" \
                      else x_indicator.split("_")[0].capitalize() + " (kcal) for 100g")
        ax.set_ylabel(y_indicator.split("_")[0].capitalize() + " content (g) for 100g" if y_indicator != "energy_100g" \
                      else y_indicator.split("_")[0].capitalize() + " (kcal) for 100g")
        
    def _on_change(self, _):
        self._update_app()
        
    def _update_app(self):
        x_indicator = self._x_dropdown.value
        y_indicator = self._y_dropdown.value
        pnns_group = self._pnns_dropdown.value
        self._nsample_slider.observe(self._update_nsample, names=['value'])
        nsample = self._nsample_slider.value
        self._plot_container.clear_output(wait=True)
        with self._plot_container:
            self._create_plot(x_indicator, y_indicator, pnns_group, nsample)
            plt.show()

Answer 1

我没有处理样本数量，而是定义了一个占总样本量百分比的滑块：

self._nsample_slider, self._nsample_slider_box = self._create_sample_slider(10, 1, 100, 1)

然后在 _update_app 函数中，这个 % 用于获取要绘制的实际样本数（我不得不再次使用数据框）：

def _update_app(self):
    df = self._df
    x_indicator = self._x_dropdown.value
    y_indicator = self._y_dropdown.value
    pnns_group = self._pnns_dropdown.value
    if self._pnns_dropdown.value != "All":
        nsample = int((self._nsample_slider.value / 100) * df[df["pnns_groups_2"] == self._pnns_dropdown.value].shape[0])
    else:
        nsample = int((self._nsample_slider.value / 100) * df.shape[0])
    self._plot_container.clear_output(wait=True)
    with self._plot_container:
        self._create_plot(x_indicator, y_indicator, pnns_group, nsample)
        plt.show()

如何更新IntSlider和InsSlider box自动适配样本量

How to update IntSlider and InsSlider box to automatically adapt to sample size

jupyter-notebook

jupyter-lab

ipywidgets