Story #4030

Updated by Peter Amstutz almost 5 years ago

The science team has expressed the need for Need a command line tool to provide comprehensive editing of collections/manifests. Some possible features include: collections/manifests:

* Delete files or streams from a collection
* Filter collection (include/exclude files by regex)
* Rename files or streams (by name or by regex)
* Move files/streams from one collection to another
*
Combine manifests to get a new manifest
* Add file(s) from another collection
* Upload files using arvados.command.put module and add to collection
* Optimize manifest Possibly optimize by re-packing small files (or files that are split across made up of many more blocks than necessary) small ranges) into continuous ranges that fills whole blocks (possibly with some logic to avoid starting new files at the very end of a block)

User should be able to specify whether to modify collection in place, create new collection, or operate on manifest text read in from stdin.

Back