Fixed missing cdm_schema attribute in cdm_reference when using remote databases (e.g., SQL Server).
Implemented dynamic, unique schema names (prefixed with cdm_testgenerator_) to prevent database collisions.
Fixed patientsCDM() to ensure the correct remote CDM reference is returned during the injection process.
Added regression tests to verify cdm_schema attribute presence and CohortConstructor compatibility on SQL Server.
Removed DARWIN_ prefix from environment variables to run tests on SQL Server and Postgres.
Support for PostgreSQL as a DBMS in patientsCDM()
Added RPostgres to Imports
Added skip_on_cran() to several tests to avoid network dependencies during R CMD check
Updated .gitignore and .Rbuildignore
Support other database types next to duckdb: sqlserver and databricks
Remove dependency on arrow
generateTestTables() creates the most common used tables in the Excel sheet by default
readPatients.xl(), that table then is pushed to "other tables" in the CDM reference using patientsCDM().generateTestTable() creates an Excel file with sheets that correspond to OMOP-CDM tables.
patientsCDM() now accepts cdmName as argument to allow for custom cdm name.
getEmptyCDM() returns an empty cdm object.
readPatients() now has a parameter to select either Excel or CSV files as an input.
readPatients.xl() and readPatients.csv are also exported functions for convenience.
graphCohort() provides a visualisation of cohort timelines.
JSONS are saved in the testthat/testCases folder as default for better test self-containment.
Tested with MIMIC database.