Skip to content

README.md

Latest commit

 

History

History
67 lines (45 loc) · 2.82 KB

README.md

File metadata and controls

67 lines (45 loc) · 2.82 KB

Develop locally

To set up your environment with a local editable install of simularium_model_utils:

cd [path to root of repo]
conda env create -f env.yml
conda activate myenv
pip install -e .[dev]
# or on mac
pip install -e .\[dev\]

Run example model, create outputs/ directory and logs:

cd [model]
python docker/src/[model].py template.xlsx 1 test

Build Docker Image

To install docker:
brew install --cask docker
then open the Docker app in Applications/ and give it privileged access.

To build the docker image (named readdy-actin and tagged v1.0):
cd [path to examples dir]/[model]/docker
docker build --no-cache -t readdy-actin:v1.0 ./

To run a docker container (named readdy-actin-test) using the image LOCALLY:
docker run --rm -v [path to examples dir]/[model]/:/working/ -e SIMULATION_TYPE='LOCAL' -e PARAM_SET_NAME='template' -e JOB_ARRAY_INDEX=0 -e REPLICATE=0 --name readdy-actin-test readdy-actin:v1.0

Upload Image to AWS ECR

If this is the first time, install the AWS command line tools with pip install awscli, and then set up your access keys. Go to AWS IAM web console and download keys for your user as json. To set up your keys, use aws configure, which will guide you through a wizard.

In the AWS ECR web console, create a repository (in this example called blairl/readdy-actin)

Log in to ECR/Docker:
aws ecr get-login-password --region us-west-2 | docker login --username AWS --password-stdin 108503617402.dkr.ecr.us-west-2.amazonaws.com

Add the repo's 'latest' tag to docker image:
docker tag readdy-actin:v1.0 108503617402.dkr.ecr.us-west-2.amazonaws.com/blairl/readdy-actin:latest

Push the docker image to the repo in ECR:
docker push 108503617402.dkr.ecr.us-west-2.amazonaws.com/blairl/readdy-actin:latest

Run Models on AWS Batch

Create a job definition in AWS Batch:
aws batch register-job-definition --cli-input-json file://[path to examples dir]/[model]/[model]_job_definition.json

Upload the parameters sheet (e.g. template.xlsx) to AWS S3 (or do this in the AWS S3 web console):
aws s3 cp [path to examples dir]/[model]/template.xlsx s3://readdy-working-bucket/parameters/template.xlsx

Go to the AWS Batch web console and submit a new job.

Visualize Model output

If a model didn't output .simularium visualization files, you can generate them manually from the output .h5 file(s).

Download the .h5 files from AWS S3 web console.

If you haven't already, create a conda environment with conda env create -f [path to examples dir]/[model]/docker/env.yml.

Activate the environment: conda activate myenv.

Use cd [path to examples dir]/[model] to navigate to the script directory.

Run the visualization script, e.g. python visualize_[model].py [path to directory containing .h5 file(s)] [box size] [total steps]