CDE - CIOOS Exploration and Data Discovery

Testing a dataset

If you just want to see how a dataset is harvested by CDE:

1. Start your python environment environment, conda create -n cde python=3.10;conda activate cde
2. pip install -e ./harvester
3. python -m cde_harvester --urls https://data.cioospacific.ca/erddap --dataset_ids ECCC_MSC_BUOYS
4. See files in harvest folder

Starting using docker

1. Install Docker and Docker compose. New versions of Docker include docker-compose
2. Rename .env.sample to .env and change any settings if needed. If you are running on your local machine these settings don't need to change
3. docker-compose -f docker-compose.production.yaml up -d to start all services. This will take a few minute to download, build, create the database schema.
4. Start your python 3.10 or greater environment, eg conda create -n cde python=3.10;conda activate cde
5. pip install -e ./harvester -e ./db-loader
6. Run harvester to load data. From this directory, run: sh data_loader_test.sh to just load one dataset for testing purposes
7. See website at http://localhost:8098

Development

• To run CDE locally, you will need Docker, Python and Node and a few terminal windows

• Rename .env.sample from the root directory to .env and change any settings if needed. If you are running on your local machine these settings don't need to change

• Start a local database using docker: docker-compose up -d db

• Setup Python virtual env and install Python modules:

  conda create -n cde python=3.10
  conda activate cde
  pip install -e ./downloader -e ./download_scheduler -e ./harvester -e ./db-loader

• Start the API:

    cd web-api
    npm install
    npm start

• Start the download scheduler:

    python -m download_scheduler

• Start the frontend:

    cd frontend
    npm install
    npm start

• Harvest a single dataset and load CKAN data.

  sh data_loader_test.sh

• See website at http://localhost:8000

Handy docker commands

See which cde services are running: docker-compose ps

Start all containers, the first time this runs it will build containers: docker-compose up -d

Tail logs: docker-compose logs -f

(Re/)Build and (re/)start all containers that have code changes: docker-compose -f docker-compose.production.yaml up -d --build

Rebuild database: (this will erase all your data)

docker-compose stop db
docker volume rm cde_postgres-data
docker-compose up -d db

Redis CLI: docker exec -it cde_redis_1 redis-cli

Flush redis tile cache: docker exec -it cde_redis_1 redis-cli FLUSHALL

Production deployment

From the production server,

• rename .env.sample to production.env and configure.

• Delete old redis and postgres data (if needed): sudo docker volume rm cde_postgres-data cde_redis-data

• Start all services: sudo docker-compose -f docker-compose.production.yaml up -d --build

• Harvest data:

  conda create -n cde python=3.10
  conda activate cde
  pip install -e ./harvester -e ./db-loader
  sh data_loader.sh

• Add a crontab entry for the scheduler to run nightly.

• deploy frontend to Gitpages

  API_URL=https://explore.cioos.ca/api npm run deploy