As discussed with @mkhorton, it would be good to have some new builders as we start to integrate new data sources. Here is a list of things to add:
AFLOW source update builder, to update a local AFLOW store with recent additions and deprecations to the main database. This should check the local last_updated field against the remote aflowlib_date field and update accordingly. In addition, we should keep track of new auid values in the database and those that have been removed (I believe that ones that are removed were incomplete calculations that get a new auid once they have finished).
AFLOW-style to propnet-style builder, to convert AFLOW data to propnet materials and quantities and store them in the converted format. This is done before graph evaluation to store a propnet representation of an AFLOW material.
MP-style to propnet-style builder, similar to above.
Convert the current MP builder into a graph evaluation builder, which takes in a store that looks like one the two above builders will build and runs propnet graph evaluation on them.
SNL builder for propnet-style databases to run automatic association of materials from different databases by pymatgen structure matching. See emmett for SNL database schema.
Aggregate propnet-style material db builder, which uses information from SNL db to aggregate materials properties from multiple sources into one aggregate material. Assign a unique propnet ID number.
I think this is a pretty exhaustive list, but feel free to add ideas.
As discussed with @mkhorton, it would be good to have some new builders as we start to integrate new data sources. Here is a list of things to add:
last_updatedfield against the remoteaflowlib_datefield and update accordingly. In addition, we should keep track of newauidvalues in the database and those that have been removed (I believe that ones that are removed were incomplete calculations that get a newauidonce they have finished).I think this is a pretty exhaustive list, but feel free to add ideas.