Page History

...

These classes are generally classified by those that are responsible for either reading a source file or accessing an API that reads the file for you and those that map objects created from that file into LexEVS coding scheme metadata, entity and relationship objects. 

In summary:

• Read source
• Map to objects related to source structure
• Map from source structure objects to LexEVS coding scheme object

Mapping Entry Point

In the lbConverter project  the edu.mayo.informatics.lexgrid.convert.directConversions.medDRA package contains the classes that do much of the work of the MedDRA load into LexGrid.  edu.mayo.informatics.lexgrid.convert.directConversions.medDRA.MedDRA2LGMain provides a central kickoff point with some methods that can be wrapped for load and validation responsibilities down further up the execution chain.

public class MedDRA2LGMain {

	//providing parameters for the source directory, the UMLS CUI file and a logging object from LexEVS
   	public CodingScheme map(URI cuiUri, URI sourceDir, LgMessageDirectorIF lg_messages)
	
	//Validate choice for the MedDRA source only
	private boolean validateSourceDir(URI sourceDir)

Read the Data to Source Model Objects

This package also contains edu.mayo.informatics.lexgrid.convert.directConversions.medDRA.MedDRAMapToLexGrid with a readMedDRAFile method that reads from a CSV file and persists it to objects defined in package edu.mayo.informatics.lexgrid.convert.directConversions.medDRA.Data.  While this data package defines beans that model the content of lines read from the CSV file, it also organizes them into structures that are easier for the mapping code to use. The individual manner of implementation is left up to the developer but this is a good example of how a third party library was used to process the file and how the resulting objects were stored in a data structure that's easy for the mapping code to consume.

    public void readMedDRAFiles() {
        String input;
        
        for(int i=0; i < meddraMetaData.length; i++){
            input = medDRASourceDir.getPath() + meddraMetaData[i].filename();
            try {
              
                FileReader fileReader = new FileReader(input);
                CSVReader reader = new CSVReader(fileReader, '$');
                ColumnPositionMappingStrategy<DatabaseRecord> strat = new ColumnPositionMappingStrategy<DatabaseRecord>();
                strat.setType(meddraMetaData[i].classname());
                String[] columns = getFields(meddraMetaData[i].classname()); 
            
                strat.setColumnMapping(columns);
    
                CsvToBean<DatabaseRecord> csv = new CsvToBean<DatabaseRecord>();
                List<DatabaseRecord> list = csv.parse(strat, reader);
                meddraDatabase.add(meddraMetaData[i].tablename(), list);
            } catch (FileNotFoundException e) {
                messages_.error("MedDRA input file missing.", e);
            } catch (Exception e) {
                messages_.error("Failed to read MedDRA files.");
            }
        }
    }

CSVReader is a third party CSV reader. 

Map the Model Objects to LexGrid Model Objects

Once the source is read an persisted to the appropriate model objects the MedDRAMapToLexGrid class can map these data objects derived from the MedDRA source into a complete coding scheme object. 

    public void mapToLexGrid(CodingScheme csclass) {
        try {
            loadCodingScheme(csclass);
            loadConcepts(csclass);
            loadRelations(csclass);
        } catch (Exception e) {
            messages_.error("Failed to map MedDRA data to LexEVS.");
        }     
        
        messages_.info("Mapping completed, returning to loader");
    }

Pass Control Back to BaseLoader

The tasks are to read the file into some kind of logical model or bean class object, organize these objects or make them available to be mapped into LexEVS objects, tie all objects together as a coding scheme, and pass this potentially large coding scheme object to the LexEVS BaseLoader to be persisted to the database.  Back up the execution chain in MeDRALoaderImpl the doLoad method first calls the mapping method of the conversion class MedDRA2LGMain to get the reading and mapping done, then passes control of the resulting coding scheme to BaseLoader by calling this.persistCodingSchemeToDatabase method on the BaseLoader super class. 

    if(codingScheme != null){
            messages.info("Completed mapping.  Now saving to database");
            this.persistCodingSchemeToDatabase(codingScheme);

            messages.info("Saved to database.  Now constructing version pairs");            
            return this.constructVersionPairsFromCodingSchemes(codingScheme);
        }

While it's not necessary that things are done this way, many required, most loaders have been written in this fashionway.