FastAPI + Panel

The Python library Panel is a high-level app and dashboarding solution for Python. It allows you to create custom interactive web apps and dashboards supporting nearly all plotting libraries. Sometimes though, it is also useful to embed a Panel app in large web applications, such as FastAPI. FastAPI is especially useful compared to others like Flask and Django because of it's lightning fast, lightweight framework. Using Panel with FastAPI requires a bit of configuration that we'll show you here below.

Configuration

Following FastAPI's Tutorial - User Guide make sure you first have FastAPI installed using: pip install "fastapi[all]". Also make sure Panel is installed pip install panel.

Next you'll need to create a file called main.py.

In main.py you'll need to import the following( which should all be already available from the above pip installs):

import panel as pn
from bokeh.embed import server_document
from fastapi import FastAPI, Request
from fastapi.templating import Jinja2Templates

Each of these will be explained as we add them in.

Next we are going to need to create an instance of FastAPI below your imports in main.py and set up the path to your templates like so:

app = FastAPI()
templates = Jinja2Templates(directory="templates")

We will now need to create our first rout via an async function and point it to the path of our server:

@app.get("/")
async def bkapp_page(request: Request):
    script = server_document('http://127.0.0.1:5000/app')
    return templates.TemplateResponse("base.html", {"request": request, "script": script})

As you can see in this code we will also need to create an html Jinja2 template. Create a new directory named templates and create the file base.html in that directory.

Now add the following to base.html. This is a minimal version but feel free to add whatever else you need to it.

<!DOCTYPE html>
<html>
    <head>
        <title>Panel in FastAPI: sliders</title>
    </head>
    <body>
        {{ script|safe }}
    </body>
</html>

Return back to your main.py file. We will use pn.serve() to start the bokeh server (Which Panel is built on). Configure it to whatever port and address you want, for our example we will use port 5000 and address 127.0.0.1. show=False will make it so the bokeh server is spun up but not shown yet. The allow_websocket_origin will list of hosts that can connect to the websocket, for us this is fastApi so we will use (127.0.0.1:8000). The createApp function call in this example is how we call our panel app. This is not set up yet but will be in the next section.

pn.serve({'/app': createApp},
        port=5000, allow_websocket_origin=["127.0.0.1:8000"],
        address="127.0.0.1", show=False)

You could optionally add BOKEH_ALLOW_WS_ORIGIN=127.0.0.1:8000 as an environment variable instead of setting it here. In conda it is done like this.

conda env config vars set BOKEH_ALLOW_WS_ORIGIN=127.0.0.1:8000

Sliders App

Based on a standard FastAPI app template, this app shows how to integrate Panel and FastAPI.

Create a new directory called panelApps. Inside your new directory, create a new file called app1.py and another new file called pn_app.py.

In app1.py add the following code for an example sliders app:

import numpy as np
import param
from bokeh.models import ColumnDataSource
from bokeh.plotting import figure


class SineWave(param.Parameterized):
    offset = param.Number(default=0.0, bounds=(-5.0, 5.0))
    amplitude = param.Number(default=1.0, bounds=(-5.0, 5.0))
    phase = param.Number(default=0.0, bounds=(0.0, 2 * np.pi))
    frequency = param.Number(default=1.0, bounds=(0.1, 5.1))
    N = param.Integer(default=200, bounds=(0, None))
    x_range = param.Range(default=(0, 4 * np.pi), bounds=(0, 4 * np.pi))
    y_range = param.Range(default=(-2.5, 2.5), bounds=(-10, 10))

    def __init__(self, **params):
        super(SineWave, self).__init__(**params)
        x, y = self.sine()
        self.cds = ColumnDataSource(data=dict(x=x, y=y))
        self.plot = figure(plot_height=400, plot_width=400,
                           tools="crosshair, pan, reset, save, wheel_zoom",
                           x_range=self.x_range, y_range=self.y_range)
        self.plot.line('x', 'y', source=self.cds, line_width=3, line_alpha=0.6)

    @param.depends('N', 'frequency', 'amplitude', 'offset', 'phase', 'x_range', 'y_range', watch=True)
    def update_plot(self):
        x, y = self.sine()
        self.cds.data = dict(x=x, y=y)
        self.plot.x_range.start, self.plot.x_range.end = self.x_range
        self.plot.y_range.start, self.plot.y_range.end = self.y_range

    def sine(self):
        x = np.linspace(0, 4 * np.pi, self.N)
        y = self.amplitude * np.sin(self.frequency * x + self.phase) + self.offset
        return x, y

Now navigate to pn_app.py and add the following code to create server document from an instance of your SineWave class.

Note: here is where createApp is created that is used by pn.serve() in main.py.

import panel as pn

from .app1 import SineWave

def createApp():
    sw = SineWave()
    return pn.Row(sw.param, sw.plot).servable()

We now need to return to our main.py and import the createApp function. Add the following import near the other imports:

from panelApps.pn_app import createApp

Your file structure should now be like the following:

fastapi
│   main.py
│
└───panelApps
│   │   app1.py
│   │   pn_app.py
│
└───templates
    │   base.html

And your finished main.py should look like this:

import panel as pn
from bokeh.embed import server_document
from fastapi import FastAPI, Request
from fastapi.templating import Jinja2Templates

from panelApps.pn_app import createApp

app = FastAPI()
templates = Jinja2Templates(directory="templates")


@app.get("/")
async def bkapp_page(request: Request):
    script = server_document('http://127.0.0.1:5000/app')
    return templates.TemplateResponse("base.html", {"request": request, "script": script})

pn.serve({'/app': createApp},
        port=5000, allow_websocket_origin=["127.0.0.1:8000"],
         address="127.0.0.1", show=False)

Get Ready To Start Up

Your app should be ready to run now. To run your newly created app. Type the following command or copy and paste into your terminal:

uvicorn main:app --reload

The output should give you a link to go to to view your app:

Uvicorn running on http://127.0.0.1:8000 (Press CTRL+C to quit)

Go to that address and your app should be there running!

Multiple Apps

This is just a basic single app example. To run multiple apps you will need to do the following:

  1. Create a new file in your panelApps directory (ex. app2.py) and add your new app code.
  2. Create another pn_app file in your panelApps directory (ex. pn_app2.py) That might look something like this:
import panel as pn

from .app2 import SineWave2

def createApp2(doc):
    sw = SineWave2()
    return pn.Row(sw.param, sw.plot).servable()
  1. Create a new html template (ex. app2.html) with the same contents as base.html
  2. Import your new app in main.py from panelApps.pn_app import createApp2
  3. Add your new app to the dictionary in bk_worker() Server
{'/app': createApp, '/app2': createApp2}
  1. Add a new async function to rout your new app (The bottom of main.py should look something like this now):
@app.get("/")
async def bkapp_page(request: Request):
    script = server_document('http://127.0.0.1:5000/app')
    return templates.TemplateResponse("base.html", {"request": request, "script": script})
    
@app.get("/app2")
async def bkapp_page2(request: Request):
    script = server_document('http://127.0.0.1:5000/app2')
    return templates.TemplateResponse("app2.html", {"request": request, "script": script})

pn.serve({'/app': createApp, '/app2': createApp2},
        port=5000, allow_websocket_origin=["127.0.0.1:8000"],
         address="127.0.0.1", show=False)

Conclusion

That's it! You now have embedded panel in FastAPI! You can now build off of this to create your own web app tailored to your needs. Visit https://github.com/t-houssian/FastAPI-Panel For all of the code above.

Select a repo