Data Dictionary

This is a working space for collaborating on the ARGO Data Dictionary.

File Schemas are found in /schemas

Variables, such as recurring scripts, regex, or code lists are found in /references. These values can be used in the restrictions section of a schema by refering to them with this pattern: #/path/to/value.

Scripts for validating field values fall under /references/validationFunctions/{schema name}/. If a validation script is common in many schemas, it can be added under /references/validationFunctions/common/. The filename of the script should not contain dashes/dots (ideally be in camelCase).

Example variable usage:

{
  "name": "submitter_donor_id",
  "valueType": "string",
  "description": "Unique identifier of the donor, assigned by the data provider.",
  "meta": {
    "examples": "90234,BLD,donor_89,AML-90"
  },
  "restrictions": {
      "required": true,
      "regex": "#/regex/submitter_id",
  }
},
{
  "name": "cause_of_death",
  "valueType": "string",
  "description": "Indicate the cause of a donor's death.",
  "meta": {
    "dependsOn": "donor.vital_status"
  },
  "restrictions": {
      "script": "#/script/donor/ensureDeceased",
  }
}

More information on writing schemas can be found here.

Scripts

Scripts are provided in NodeJS, requiring npm and Node installed locally.

To use these commands first install the npm dependencies by running: npm install

Unit Testing

Unit tests go in the tests folder, and are created with Jest.

To run all unit tests, you can use : npm run test.

More information about unit tests.

Compile Whole Dictionary

1. Install dependencies via npm.

npm ci

2. Then run the dictionary building script:

npm run compile

This will check that all units tests pass, and then will collect all schemas and references and format them into a full dictionary object that could be uploaded to lectern. An abridged version of the dictionary will be printed to the console, and the full compiled dictionary will be output to the file ./dictionary.json

You will be prompted to provide the dictionary name and version number, or leave them blank and accept the defaults (ICGC-ARGO Data Dictionary and 0.0 respectively).

If all tests do not pass, the dictionary will not be compiled.

Note: The dictionary.json file is ignored by git. This file can be uploaded to Lectern, for example, but it shouldn't appear in this repo.

Testing your dictionary locally

node populateReferences.js will create populated_dictionary.json. The output of the script is not intended to be uploaded to Lectern. Instead, it is to aid in testing with the clinical repo. It can be used to add/overwrite the contents of the sample-schema, when the Lectern URL in the .env file is set to said file's path.