Standardization of embedding cohort definitions in a study package

[schuemie]

There are currently different ways to embed cohort definitions in study packages used throughout HADES. I propose the following standard:

A CSV file called 'Cohorts.csv' in the 'inst/' folder, with at least these fields:

• cohortId: the cohort ID used throughout in the study package. This is also used to generate the JSON and SQL file names.
• cohortName: the cohort name used throughout the study package. This name is for example displayed in Shiny apps
• webApiCohortId: only used by ROhdsiWebApi::insertCohortDefinitionSetInPackage when grabbing the cohort definitions from the WebAPI.

In the 'inst/cohorts' folder one JSON file per cohort definition named '[cohortId].json', so for example '1.json', '2.json', etc.

In the 'inst/sql/sql_server' folder one SQL file per cohort definition named '[cohortId].sql', so for example '1.sql', '2.sql', etc.

(Note: I'm proposing we not store all this information (all JSON, SQL, meta-data) in a single RDS because that could become a very big file that might not fit in memory, and because we couldn't use GitHub's version control to see what has changed.)