HashStore 1.1.0 Release

[doulikecookiedough]

Summary of Changes:

• Major refactor of HashStore involving storing objects with their content identifiers cid and creating reference files
  • Storing an object with a pid will move the object to where it belongs, create a pid reference file that contains the cid, and a cid reference file that contains pids that reference a cid
  • Storing an object without a pid will only store the object. The caller/client must call tag_object to create the reference files (else create an orphaned object), and optionally delete_if_invalid_object should they have data they want to validate the object_metadata returned from storing an object.
• Metadata is now stored using a object named with the hash of the pid+format_id and in a directory formed using the hash of the pid in /metadata
• New Public API Methods
  • tag_object, delete_if_invalid_object
• Improved thread safety and synchronization to address possible race conditions when working with data objects and added thread safety when working with metadata.
• The hashstore config file is now separated in two pieces, with the actual .yaml content being formed with a library
• HashStore Client Improvements
  • New script entry point after installing hashstore through poetry install
• Revised python docstrings into reStructuredText, a format compatible with sphinx-autodocumentation along with typehints
• New functionality to use python threads or processes via setting an environment variable
• Revised README.md to provide context to the HashStore library and how-to sections.
• Overall code clean up, revision to logging statements and various minor bug fixes

Greetings @artntek - When you have scope, can you please help review HashStore to assist me in releasing HashStore 1.1.0? Please do not hesitate to scrutinize any of the code or bring to my attention any concerns. Thank you 🙏!

[doulikecookiedough]

Thanks again @artntek for your review! I've made the updates in another branch and merged them into develop. I'm also going to hold off on the 1.1.0 release until the changes to deleteObject are implemented (and merged into develop). I will ping you again then to wrap this PR up.

[doulikecookiedough]

Greetings @jeanetteclark! When you have scope, can you please help review/perform a sanity check on HashStore?

Additional Changes Since Matthew's Review:

• Refactored delete_object functionality & scope
  • Clients can delete a single object (based on a given content identifier cid)
  • Clients can delete an object and all of its related metadata when given a pid
• Optimized tag_object flow with additional locks
  • Added custom exception classes and revised logging statements to improve clarity
• HashStore Client Updates
  • Revised the way in which the client is called via an executable installed by poetry
  • Fixed bug in storing metadata (incorrect name due to changes not accounted for)
• Improvements to synchronization process via usage of Conditions
  • Added multiprocessing locks & conditions
  • Notable Change: Test times executing in ~2s vs ~15s previously
• Updated README.md