Because I want to visualize trades in simulation runs.
Using Dash; a framework for building apps and dashboards with plotly.
The dashboard in this repo works by creating a dash app, and running it on a thread, so not to block the execution of training. The graph is updated every 100 milliseconds via a callback setup on the dash app.
Steps:
- import and create a dashboard in your training file or agent file:
import dashboard
import threading
ds = dashboard.DashboardServer()
t = threading.Thread(target=ds.run)
t.start()
-
Setup an environment with a
PlotlyTradingChartrenderer -
In your training loop, after you call
env.render()do:
dashboard.dashboard.update_figure(env.renderer.fig)
Run:
python3 example.py
To view the graph goto localhost:8050
