Shiny for Python app fails to load

ShinyProxy
#1

Hi,

I try to run a Shiny for Python app on ShinyProxy 2.6.1.

While the Docker container works fine locally (and when run on the server directly as well), I just get a “Container did not respond in time” on ShinyProxy.

The Dockerfile, I use to create the image is

FROM <CUSTOM-BASE-IMAGE>

USER docker
WORKDIR /home/docker 
COPY requirements.txt /home/docker
RUN pip3 install --no-cache-dir --upgrade -r requirements.txt
RUN pip3 install gunicorn
COPY . /home/docker
EXPOSE 8080
CMD ["gunicorn", "app:app", "--bind", "0.0.0.0:8080", "-k", "uvicorn.workers.UvicornWorker"]

and the app configuration for ShinyProxy looks like:

  - id: pyshiny-demo
    container-image: <REGISTRY>/shinyproxy-pyshiny-airmass
    port: 8080
    container-env:
      SCRIPT_NAME: "#{proxy.getRuntimeValue('SHINYPROXY_PUBLIC_PATH').replaceFirst('/$','')}"
    target-path: "#{proxy.getRuntimeValue('SHINYPROXY_PUBLIC_PATH')}"

Using uvicorn as recommended in the docs didn’t work either.

Any idea how to debug further?

#2

Hi

We tested Shiny For Python in ShinyProxy and this is working for us. You can see the demo code here: https://github.com/openanalytics/shinyproxy-shiny-for-python-demo .
I think that the main difference with your setup is that we didn’t add the SCRIPT_NAME and target-path settings.

Can you let us know whether your app works with the above demo code?

#3

Thanks for the example, that works fine from e. But unfortunately I get a “Unsupported upgrade request.” from the airmass app still. Other apps working though, so not sure what actually causes that particular issue.
Thanks again for helping!

#4

That’s weird, this app is working fine for me.
I’m using this Dockerfile:

FROM python:3.10

WORKDIR /app

EXPOSE 8080

COPY . .

RUN pip3 install gunicorn
RUN pip3 install -r requirements.in

CMD ["gunicorn", "app:app", "--bind", "0.0.0.0:8080", "-k", "uvicorn.workers.UvicornWorker"]

Note that I had to install the dependencies from the requirements.in file, since it seems a wrong Shiny version is pinned in the requirements.txt file (and python would give an error at startup).

I pushed this image to Docker hub using ledfan/airmass, can you test the following config:

proxy:
  title: Open Analytics Shiny Proxy
  logo-url: https://www.openanalytics.eu/shinyproxy/logo.png
  landing-page: /
  heartbeat-rate: 10000
  heartbeat-timeout: 60000
  port: 8080
  authentication: simple
  admin-groups: scientists
  users:
  - name: jack
    password: password
    groups: scientists
  - name: jeff
    password: password
    groups: mathematicians
  specs:
  - id: pyshiny-demo
    container-image: ledfan/airmass
    port: 8080
logging:
  file:
    shinyproxy.log

Finally, I see that this example is downloading a file from the internet, maybe this fails? Can you check the logs of the container when it fails, e.g. using docker ps -a, docker logs -f <id>