Classes and their responsibilities

See the diagram at https://github.com/runningwave/openmrs-groovy-data-importer/blob/master/docs/architecture_diagram.png for an overview.

Classes in org.openmrs.* are meant for reuse by other projects.

• org.openmrs.tools.importer.Launcher launches the script. It reads command line arguments, initializes the OpenMRS context, instantiates the requested class of importer, and kicks off the import. It closes the context and does any other cleanup when the import is finished.

• org.openmrs.tools.importer.BaseEncounterImporter is an abstract class that directs the import process. initComponents() is a factory that instantiates the assembler and source you're using for this import.

importEncounters() is where all the action happens. It iterates over the data source. It calls the assembler to create OpenMRS objects. It's responsible for saving OpenMRS objects. A record may include several OpenMRS objects that need to be saved (to get ID numbers) in a particular order. importEncounters() decides what to do if a record import fails, including logging enough information for you to identify and retry a failed record.

• org.openmrs.tools.importer.source is responsible for getting field values from the raw data. This is a good place to encapsulate any cleanup or transformation. Examples of cleanup: making values lowercase, trimming whitespace, substituting meaningful terms for constants or internal IDs.

An implementation of a csv file source is included. You could create one to read from a database, or from multiple sources. A data source must be able to indicate that there are more records available to consume.

• org.openmrs.tools.importer.assembler puts the pieces together. It's responsible for expressing the current line of data from the Source as something the DSL can understand. It hands an openMRS object to the Importer when asked.

• org.openmrs.dsl.* - a Domain Specific Language written in Groovy, to help build OpenMRS objects in a somewhat more natural way.