Dev Environment Setup

Dev Environment Setup#

Pre-reqs and system setup are covered in server setup. You will then need to fork and clone the repo.

Local Testing#

The following instructions are for once you’ve made changes to your local forked repo and you’re ready to test.

  1. cd to root folder of scout

  2. Build the docker image of your altered code. You’ll want to tag the image for easy reference later. For this demo, we use the tag yourimage.

    docker build -t yourimage .

  3. Verify you have a successful build. Check your terminal logs for the following:

    Successfully built [IMAGE-ID]
Successfully tagged yourimage:latest

  4. Run a container with your newly built docker image.

    docker run --privileged -p 1337:1337 --rm -it --gpus all --mount type=bind,source=/data,target=/data -e ENV_IP="`ip route get 1 | sed 's/^.*src \([^ ]*\).*$/\1/;q'`" -v /data/scout/db:/data/db -v /data/scout/tmp:/tmp/scout-tmp yourimage:latest

  5. Provide the image directory information you created in server setup. This will only be necessary for your first setup.

  6. Check your work! Use your browser to access localhost:1337 and see if your changes are working as desired.

  7. Continue making changes or create your PR.

Run without GPU#

If you are doing development work on a feature that is not impacted by image availability or ML results, you can go from storage setup to setting up your code repo.

docker run --privileged -p 1337:1337 --rm -it --mount type=bind,source=/data,target=/data -e ENV_IP="`ip route get 1 | sed 's/^.*src \([^ ]*\).*$/\1/;q'`" -v /data/scout_data:/data/db -v /data/scout/tmp:/tmp/scout-tmp yourimage:latest

Run image that includes data#

To simplify development, we have provided a Docker image that comes pre-loaded with image data and tasks. Example images are from the WAID dataset from Applied Sciences.

Note that this version of Scout does not include Scoutbot and therefore does not require a GPU to function. This also means, however, that it cannot process images with ML classifiers.

One time setup#

From root of the code directory (/scout), run npm install to make sure all necessary libraries are installed.

Usage#

  1. cd to the /develop-with-data directory

  2. Run docker-compose up to fetch that docker image and run Scout.

  3. Once running, use a browser to open http://localhost:1337

  4. Login with admin/admin as username/password.

Troubleshooting#

  • If you encounter a docker error urllib3.exceptions.URLSchemeUnknown: Not supported URL scheme http+docker, try running pip install requests=2.31.0

  • If your process fails, you should run docker-compose down to stop the container when you shut down your scout instance (Ctrl+C)