TODO: Publish Docker to Docker Hub

[callahantiff]

Publish Docker to Docker Hub, when creating v2.0.0 release.

[callahantiff]

Update README for how to use project Docker. Good examples from @bill-baumgartner included below:

https://github.com/UCDenver-ccp/craft-shared-tasks/wiki/Concept-Annotation-Task-Evaluation

https://github.com/UCDenver-ccp/craft-shared-tasks/wiki/Building-the-CRAFT-evaluation-Docker-container

[callahantiff]

@bill-baumgartner - I am drafting the content that we will ultimately add to the README for utilizing Docker. I thought I would run it by you here. Would you mind taking a look at let me know what you think?

Notes

• The DockerHub link is not yet active and will be updated with the correct link once our container is uploaded.
• References to "above" will make sense once this information is added back to the README.

---

Docker

In order to utilize the Dockerized version of the code, please make sure that you have downloaded the newest version of Docker. There are two ways to utilize Docker with this repository:

• Obtain the pre-built Docker container from [DockerHub]()
• Build the Container

Build the Container
To build the PheKnowLator Docker container:

• Download a stable release of this repository or clone this repository to get the most up-to-date version
• Unpack the repository downloaded (if necessary), then execute the following commands to build the container:

cd /path/to/PheKnowLator (Note, this is the directory containing the Dockerfile file)
docker build -t pkt:[VERSION] .

Running the Container
The following code can be used to run PheKnowLator from outside of the container (after obtaining a prebuilt container or after building the container locally):

docker run --name pkt_example --rm -ti -v /local/path/to/PheKnowLator/resources/knowledge_graphs/:/PheKnowLator/resources/knowledge_graphs/ pkt:2.0.0 --app subclass --kg full --nde yes --rel yes --owl no --kgm yes

NOTES:

• The Docker container cannot write to an encrypted filesystem, however, so please make sure /local/path/to/PheKnowLator/resources/knowledge_graphs references a directory that is not encrypted
• The example shown above builds a full version of the knowledge graph using the subclass construction approach with node metadata, inverse relations, and decoding of OWL classes. See the Running the pkt Library section for more information on the parameters that can be passed to PheKnowLator.

---

[bill-baumgartner]

Looks good. I have three comments: 1) There is a missing dot at the end of the docker build command, it should be: docker build -t pkt:[VERSION] . 2) The docker run command can work from any directory so "run the following code from the project root (i.e. PheKnowLator/):" is not necessary, and may not be possible if the user doesn't download the PheKnowLator code. 3) I think the Dockerfile needs to be changed at this point such that it doesn't download the code from GitHub, but instead uses the COPY command to import into the container. This way, any local changes to the code will be automatically imported. I think this is the expected behavior and may lead to confusion if we are instead pulling in the code from GitHub each time.

[callahantiff]

Looks good. I have three comments:

1. There is a missing dot at the end of the docker build command, it should be: docker build -t pkt:[VERSION] .
2. The docker run command can work from any directory so "run the following code from the project root (i.e. PheKnowLator/):" is not necessary, and may not be possible if the user doesn't download the PheKnowLator code.
3. I think the Dockerfile needs to be changed at this point such that it doesn't download the code from GitHub, but instead uses the COPY command to import into the container. This way, any local changes to the code will be automatically imported. I think this is the expected behavior and may lead to confusion if we are instead pulling in the code from GitHub each time.

Thanks @bill-baumgartner! I have fixed the first two points and will work on the third. I need to think about what assumptions we will impose when utilizing the COPY approach. I will put it on the list for our meeting tomorrow.

[callahantiff]

TASK

Task Type: INFRASTRUCTURE Publish Docker container to DockerHub

TODO

• [ ] Decide if a single Docker container is published
• [ ] Publish container and update README

Questions:

• @bill-baumgartner - Do we need to have two versions 1 for the CI/CD and one that people can download independently and run on their own? Perhaps we can just publish one to DockerHub and that can fulfill both use cases?