Closed robmcmullen closed 7 years ago
I think all the loading for process_extra_metadata should be in SegmentedDocument! And probably each task should have its own section. There should be a general section which applies to the document itself, and other sections that only the tasks need.
Should fix #88; if a .omnivore file exists and has a document mime type and parser, restore from there rather than parsing again. The segments parsed from the saved version of the parser can then be restored, including the origin.
Will require update to omnivore file header: currently at v1, change to v2