The make_eml() function renders templates (and input
arguments) into EML metadata (click make_eml() for the docs
and edi.260.1.xml
for an example EML file).
make_eml() insists on access to the data objects to keep
users from the monotonous mind-numbing task of gathering physical
attributes (e.g. file size, number of rows, checksums, etc.) each time a
data object changes. This enables automatic rebuild of revised data
object EML as long as the structure remains constant (i.e. variable
types and definitions don’t change, only new records are added). This
highlights the benifit of creating data with stable attributes not to
mention for the consideration of downstream user workflows depending on
a specific input data structure. Additionally, automated rebuild of EML
streamlines the data publication process further by enabling automatic
upload of a data package through a repository API, which is common for
timeseries data or data derived from time series data packages
(e.g. synthesis products). See “Automate
data publication” for more on this topic.