FileID format

Format

The key is constructed by using the file name including the path in the batch structure with the batchid (including round trip number) as the root element. The path separator is the '_' (underscore) char. The maximum length allowed is 250.

The fileID must conform to the regular expression: [a-zA-Z0-9\-_.] {5,250}

Example: The file

B400022028241-RT1/400022028241-14/1795-06-13-01/adresseavisen1759-1795-06-13-01-0006.jp2

in batch structure will be archived with fileID:

B400022028241-RT1_400022028241-14_1795-06-13-01_adresseavisen1759-1795-06-13-01-0006.jp2

Considerations

The defined format is based on the following considerations 

• File and directory name format: Most of the elements making up the directory and file names are either predefined name (UNMATCHED, ISO, Target) or number formats (0002, 1, 1890-10-18-01, 4000220289521). The only slightly complexed case is the newspaperID element, where the format is [a-zA-Z0-9] (this is apparently not written down anywhere).
• The bitrepository only accepts FileIDs of the format [a-zA-Z0-9\-_.] {5,250}.

Note that the constraint on newspaperIDs must be implemented in the file structure check and shared with the MFPak people. The ingester itself also needs to protect itself from invalid newspaperIDs, as the file structure checker will only be run on the received files after ingest (should have been run at ninestars though).