Week from Dec 07 to Dec 13

Introduction

During the nineteenth week I started to work on the new implementation of the storer.py module which was included in the oc_ocdm package for the first time. The goal is to update its inner workings so to reflect the brand new APIs of the oc_ocdm package, which were largely extended with respect to those of the original graphlib.py module. I also had to apply some fixes here and there in order to polish the oc_ocdm codebase before the end of the grant.

Report

New Storer implementation

First of all, I started working on the new implementation of the Storer class contained inside the storer.py module. I included the old storer.py module together with its dependency reporter.py into the oc_ocdm package. I cleaned a little bit the code and I added type hints for every method. I analyzed the actual implementation to better understand how it works, what should be changed and what should be kept as it is now.

Various fixes

The method apply_changes of GraphEntity was refactored and a new method with the same name was added to the GraphSet class. Regarding entities marked as to be deleted, there’s a little difference between calling apply_changes on the single entity and on the entire GraphSet: in the first case, all triples will be removed from the entity (including the mandatory rdf:type), while in the second case the entity will also be completely removed from the GraphSet itself. I removed the optional boolean parameter update_entities (added last week) from the generate_provenance signature because it was a mistake. The method apply_changes should only be called after the execution of a “store” or “upload” operation of the Storer class.

I fixed the “shexc.txt” and “shexc_closed.txt” files: in both of them I add to complete the list of possible values for the property “datacite:usesIdentifierScheme” of the “IdentifierShape”.

I removed the triplestore_url parameter of the ProvSet constructor, also leading to the removal of the fields self.ts and self.triplestore_url which were not used by any part of the codebase (because of the new implementation of generate_provenance that does not require to execute queries on the online triplestore anymore).

Methods get_types and remove_type of the classes BibliographicResource, Citation, DiscourseElement and ResourceEmbodiment, which were are already defined in the GraphEntity superclass, were removed from their respective class since they were effectively useless and their implementation was older with respect to their corrispectives in the GraphEntity class.

I fixed the name of some methods from various entity classes:

• AgentRole:
  • from get_held_by to get_is_held_by
  • from remove_held_by to remove_is_held_by
• BibliographicResource:
  • from get_part_of to get_is_part_of
  • from remove_part_of to remove_is_part_of
  • from get_in_reference_lists to get_contained_in_reference_lists
  • from remove_in_reference_list to remove_contained_in_reference_list
  • from get_discourse_elements to get_contained_discourse_elements
  • from remove_discourse_element to remove_contained_discourse_element
• Citation:
  • from remove_creation_date to remove_citation_creation_date
  • from remove_time_span to remove_citation_time_span
  • from remove_characterization to remove_citation_characterization
• DiscourseElement:
  • from get_discourse_elements to get_contained_discourse_elements
  • from remove_contained_de to remove_contained_discourse_element
  • from get_context_of_rp to get_is_context_of_rp
  • from remove_context_of_rp to remove_is_context_of_rp
  • from get_context_of_pl to get_is_context_of_pl
  • from remove_context_of_pl to remove_is_context_of_pl
• ReferencePointer:
  • from remove_be to remove_denoted_be
• BibliographicEntity:
  • from has_id to has_identifier
  • from remove_id to remove_identifier