meek0/clin-arranger

Arranger server for clin frontend

About

This is a clin specific extension of the arranger project . More specifically, It is an ExpressJs app on top of arranger with extra endpoints.

Usage

Variables

This application takes minimally the following variables as input:

• PORT: Port number you want to the server to listen to. Please note that for admin service, PORT is hardcoded.
• LOGS_REQUEST_INTERCEPTOR: Log or not incoming request.
• AUTH_CLIENT_ID: Keycloack config.
• AUTH_SERVER_URL: Keycloack config.
• AUTH_REALM: Keycloack config.
• ES_HOST: Elastic search config.
• ES_USER: Elastic search config.
• ES_PASS: Elastic search config.
• PATIENTS_INDEX_NAME: Index of patients in ES.
• VARIANTS_INDEX_NAME: Index of variants in ES.
• CNV_INDEX_NAME: Index of cnv in ES.
• COVERAGE_BY_GENE_INDEX_NAME: index of coverage by gene in ES
• GENES_INDEX_NAME: Index of genes in ES.
• HPO_INDEX_NAME: Index of HPOs in ES.
• ANALYSES_INDEX_NAME: Index of analyses in ES.
• SEQUENCINGS_INDEX_NAME: Index of sequencings in ES.
• NODE_TLS_REJECT_UNAUTHORIZED: Certificate validation: if equals to 0, certificate validation is disabled for TLS connections.
• LOG_LEVEL: Logs levels (info, debug, error, ...) defined by Winston (library)
• PROJECT_INDICES: Space-separated list of indices that make up the arranger project.

Development Setup with Docker

Before going further, make sure that docker and docker-compose are installed on your system.

# 1. clone the repository
  git clone [email protected]:Ferlab-Ste-Justine/clin-arranger.git

# 2. enter the project's folder
  cd clin-arranger

# 3. create an .env file (you may have to adjust the template to your needs)
  touch .env

# 4 in a terminal, run docker-compose from project's docker-compose file. 
  docker-compose --profile <target profile> up # for admin service

# 5 to clean up afterwards once your are done developing.
  docker-compose --profile  <target profile> down

Note: you can activate multiple profiles at once: docker-compose --profile a --profile b ... up

⚠️ With this setup, your host and the app's container share the project directory/volume.

Deactivating arranger-projects

As a rule of thumb, it is best to keep the number of arranger-projects small (at least in versions 2.X.X). Every active arranger project creates a self-contained Apollo server and emits various events. In order to deactivate some projects one can post a similar query to ES ¹:

https://<ES_HOSTS>/arranger-projects/_update_by_query
{
  "script": {
    "lang": "painless",
    "source": "if (params['projects'].contains(ctx._source.id)) { ctx._source.active = false }",
    "params": {
      "projects": ["project_to_deactivate_x1", "project_to_deactivate_x2", "project_to_deactivate_x3"]
    }
  }
}

Test your Dockerfile

If you want to replicate qa/staging/prod as faithfully as possible you can do:

# 1 make sure that vars in .env point to correct services
  cat .env

# 2 build the image
  docker build -t <your-tag> .
  
# 3 run a container
  docker run --rm -it --network=<es-net| or else> <your-tag>

# 4 delete the image if you want to
  docker rmi <your-tag>

Footnotes

1. Please note that this was tested on elasticsearch v7.17.0. ↩