v0.4.15

• Add preprocessor for Langmuir collection CSV files (#667)
  The preprocessor accepts existing Langmuir CSVs with one file per row and returns a new csv with each work and fileset in their own row.
  All of the rows for a single work are grouped together
  Each group begins with a row containing the work-level metadata
  All files associated with a fileset are listed in the same row
  Filesets are listed in sequence order and given appropriate labels
  Blank lines in the source CSV are ignored
  The preprocessor adds a deduplication_key field
• Change Langmuir fileset labelling from 'Side #' to 'Image #' (#671)
• Adds preservation events (#659)
• Add update logic to ModularImporter (#670)
  This continues the use of the decorator pattern to add the logic for the update options to the importer.
  The import_langmuir_from_csv spec has been updated to test for each of the new conditions.
• Add pull list mappings to the YellowbackPreprocessor (#662)
  Adds the fields that map directly from the pull list to the final import manifest.
  Some fields in the pull list like "CSV Call Number" are provided for reference only and will be mapped later from authoritative data in Alma records instead.