179
190

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?

More than 5 years have passed since last update.

Plotly Dashで簡単に可視化ができるWebアプリを作る

Last updated at Posted at 2017-08-05

Plotly Dashとは

Plotly Dash(正式名称はDashらしいですが、他にもDashという名前のアプリが色々あるのでこの表記とします)はPython製のWebアプリケーションのフレームワークです。
FlaskやBottleなど、Python製のWebフレームワークは色々ありますが、Plotly DashはPlotlyを使用した可視化されたデータを組み込むことができます。

インストール

pipでインストールします。condaでインストールできるかどうかは確認していません。

pip install dash==0.17.7  # The core dash backend
pip install dash-renderer==0.7.4  # The dash front-end
pip install dash-html-components==0.7.0  # HTML components
pip install dash-core-components==0.11.0  # Supercharged components
pip install plotly==2.0.12  # Plotly graphing library used in examples

とりあえずかいてみる

import dash
import dash_core_components as dcc
import dash_html_components as html
import numpy as np

x = np.linspace(-np.pi, np.pi, 10)
y = np.sin(x)
app = dash.Dash()

app.layout = html.Div(children=[
    html.H1(children='H1の文章'),

    html.Div(children='''
        divの文章
    '''),

    dcc.Graph(
        id='example-graph',
        figure={
            'data': [
                {'x': x, 'y': np.sin(x), 'type': 'line', 'name': 'line'},
                {'x': x, 'y': np.cos(x), 'type': 'bar', 'name': 'bar'},
            ],
            'layout': {
                'title': 'グラフのタイトル'
            }
        }
    )
])

if __name__ == '__main__':
    app.run_server(debug=True)

Screenshot at 2017-08-05 13-45-20.png

上記のようにHTMLを書く必要はなく、Pythonのコードだけで完結できます。
グラフの描画はPlotlyが担っています。

Markdown

Markdown記法に対応しています。

import dash
import dash_html_components as html
import dash_core_components as dcc

markdown_doc = dcc.Markdown('''
# 見出し1
## 見出し2
### 見出し3
---
* 箇条書き
* 箇条書き
---

[りんく](http://commonmark.org/help)
![いめーじ](https://images.plot.ly/static/marketing/dash/dash-logo.png)

> 引用
''')

app = dash.Dash()
app.layout = html.Div(markdown_doc)

if __name__ == '__main__':
    app.run_server(debug=True)

image.png

筆者の環境ではtableが表現できませんでした。もしうまくいった方がいらっしゃいましたら教えてください。

Live Updates

動的なグラフにも対応しています。

下記の例では psutil を使用してマシンのリソース状況をリアルタイムにモニタします。

import dash
from dash.dependencies import Input, Output, Event
import dash_core_components as dcc
import dash_html_components as html
import datetime
import plotly
import numpy as np
import pandas as pd
import psutil

colors = {
    'background': '#111111',
    'text': '#7FDBFF'
}

app = dash.Dash(__name__)
app.layout = html.Div(children=[
    html.Div([
        html.H4('オレオレシステムモニタ'),
        html.Div(id='live-update-text'),
        dcc.Graph(id='live-update-graph'),
        html.H4('プロセスリスト'),
        html.Div(id='live-update-proc'),
        dcc.Interval(
            id='interval-component',
            interval=1 * 1000  # in milliseconds
        )
    ])
],
    style={'backgroundColor': colors['background'], 'color': colors['text']}
)


class Context:
    def __init__(self):
        self.t = []
        self.cpu = []
        self.per_cpu = [[] for x in range(psutil.cpu_count())]
        self.mem = []

    @classmethod
    def append_data(cls, d1, d2):
        n = len(d1)
        if n > 100:
            del d1[0:n - 99]
        d1.append(d2)


context = Context()

# The `dcc.Interval` component emits an event called "interval"
# every `interval` number of milliseconds.
# Subscribe to this event with the `events` argument of `app.callback`


@app.callback(Output('live-update-text', 'children'),
              events=[Event('interval-component', 'interval')])
def update_metrics():
    now = datetime.datetime.now()
    hour, minute, second = now.hour, now.minute, now.second
    style = {'padding': '5px', 'fontSize': '16px'}
    return [
        html.Span('CPU: {}%'.format(context.cpu[-1]), style=style),
        html.Span('Memory: {}%'.format(context.mem[-1]), style=style)
    ]


# Multiple components can update everytime interval gets fired.
@app.callback(Output('live-update-graph', 'figure'),
              events=[Event('interval-component', 'interval')])
def update_graph_live():
    # global context
    context.append_data(context.t, datetime.datetime.now())
    context.append_data(context.cpu, psutil.cpu_percent())
    for data, pct in zip(context.per_cpu, psutil.cpu_percent(percpu=True)):
        context.append_data(data, pct)
    context.append_data(context.mem, psutil.virtual_memory().percent)

    # Create the graph with subplots
    fig = plotly.tools.make_subplots(rows=2, cols=1, vertical_spacing=0.2)
    fig['layout']['margin'] = {
        'l': 30, 'r': 10, 'b': 30, 't': 10
    }
    fig['layout']['plot_bgcolor'] = colors['background']
    fig['layout']['paper_bgcolor'] = colors['background']
    fig['layout']['font'] = {'color': colors['text']}
    fig['layout']['legend'] = {'x': 0, 'y': 1, 'xanchor': 'left'}
    fig['layout']['yaxis1'].update(range=[0, 100])
    fig['layout']['yaxis2'].update(range=[0, 100])

    fig.append_trace({
        'x': context.t,
        'y': context.cpu,
        'name': 'cpu',
        'mode': 'lines',
        'type': 'scatter',
    }, 1, 1)
    for i, y in enumerate(context.per_cpu):
        fig.append_trace({
            'x': context.t,
            'y': y,
            'name': 'cpu {}'.format(i),
            'mode': 'lines',
            'type': 'scatter',
        }, 1, 1)
    fig.append_trace({
        'x': context.t,
        'y': context.mem,
        'name': 'memory',
        'mode': 'lines',
        'type': 'scatter',
        'fill': 'tonexty',
    }, 2, 1)

    return fig


def get_proc_df():
    def get_proc(proc):
        try:
            pinfo = proc
        except psutil.NoSuchProcess:
            pass
        return (pinfo.pid, pinfo.name(), pinfo.memory_percent(), pinfo.cpu_percent())

    data = [get_proc(proc) for proc in psutil.process_iter()]
    df = pd.DataFrame(data, columns=['pid', 'name', 'memory', 'cpu'])
    df['memory'] = df['memory'].map(lambda x: '{:.2f}%'.format(x))
    df['cpu'] = df['cpu'] / psutil.cpu_count()
    df['cpu'] = df['cpu'].map(lambda x: '{:.2f}%'.format(x))
    return df.sort_values('cpu', ascending=False)


@app.callback(Output('live-update-proc', 'children'),
              events=[Event('interval-component', 'interval')])
def generate_table():
    df = get_proc_df()
    max_rows = 10
    return html.Table(
        # Header
        [html.Tr([html.Th(col) for col in df.columns])] +

        # Body
        [html.Tr([
            html.Td(df.iloc[i][col], style={'width': '8em'}) for col in df.columns
        ]) for i in range(min(len(df), max_rows))]
    )


if __name__ == '__main__':
    app.run_server(debug=True)

Peek 2017-08-05 16-52.gif

対話的な操作

UIを操作しながら対話的に出力内容をコントロールすることができます。
やり方はそのうち書きたいと思います。

179
190
3

Register as a new user and use Qiita more conveniently

  1. You get articles that match your needs
  2. You can efficiently read back useful information
  3. You can use dark theme
What you can do with signing up
179
190

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?