Crosswalk performance for large tabular input files

[gregjan]

The Research Labs of Archaeology has a tabular input file of 40,000 records. Each crosswalk run in these circumstances takes about ten minutes and leaves the user will little feedback. (A rolling beach ball on the Mac)

Warn users about the size of a large input file, suggest that they take an action:

• limit the crosswalk to a number range of lines

[gregjan]

Create separate files for each crosswalked MODS record. This keeps a large number of records out of the in-memory workbench METS model.