search

chrism0dwk / covid19uk

MIT License
11 stars 10 forks source link

Create Dockerfile #18

Open sdwfrost opened 3 years ago

sdwfrost commented 3 years ago

In order for the model to be used in the model runner, it needs to be containerized. Brief instructions are here. Here's an example for IDMs covasim:

FROM python:3

ARG COVASIM_VER=1.0.0
RUN pip install covasim==${COVASIM_VER} pandas

COPY . /app

CMD ["python", "/app/runsim.py", "/data/input/inputFile.json", "/data/output/data.json"]
chrism0dwk commented 3 years ago

In hand...

sdwfrost commented 3 years ago

Thanks @chrism0dwk! Let me know if you need anything on this or on other issues

chrism0dwk commented 3 years ago

Linking you to our sister repo covid-pipeline where we've moved all the generic boilerplate code (ruffus pipeline, summary scripts, etc). We have a working Dockerfile with accompanying image on dockerhub which you can pull and run. Use a fast machine with a beefy GPU for best results!

That said, we'll need to talk about the inputs and outputs. If you run the Docker container (see README in the repo above), you'll see the HDF5 and NetCDF4 outputs. The model's not designed for scenario testing, but is targeted towards spatiotemporal anomaly detection, Rt measurement, and "status quo" projection. Hope that fits in with your expectations...

chrism0dwk commented 3 years ago

@sdwfrost just bumping this to keep the ball in the air

sdwfrost commented 3 years ago

Thanks - I downloaded and ran the latest tagged build of the Docker image, but it failed after running for about 3 hours. Do you have some example outputs I can play with in the meantime?

chrism0dwk commented 3 years ago

Interesting. What was the failure, I wonder?