File Groups

This page is WIP, expect more details in the future

Maintaining ordering is a core functionality of the service. Grouping is intended to improve the intuitiveness & scalability of this workflow for larger collections

• Groups captures the concept of files that should be ordered relative to each other
• The service guarantees a unique group + index combination for every file it manages and in doing so ensures related files are always projected in the correct order
• Groups can have properties that can then be used to generate directory structures

Creating a Group

A group is automatically created for every file or directory of files ingested

• Every file copied into a watched directory gets its own group
• All files created from a transform (eg. unarchiving) will be imported in a single group
• To manually create a group, create a hopper that watches for directories. The workflow is then as follows:
  1. Create a directory
  2. Fill it with files & tweak its files' lexical order
  3. Rename to a predefined name
  4. The service imports the entire directory as one single pre-ordered group

[hopper]
path = /home/user/downloads
include = source_type directory
include = file_name .commit

The service will watch for directories renamed to .commit at which point every file within the directory will be imported as a single group, with their lexical ordering intact.