This gist is using the Python code at https://github.com/pymonger/prov_es to attempt to model the prov workflow for Open Core Data. This code use a first stab at simply modifying the prov_es code to the simpler OCD event.

ocd_etl_prov.py

from datetime import datetime, timedelta
from prov_es.model import ProvEsDocument

# create doc
doc = ProvEsDocument()

# input database...
# Janus database credit JRSO/TAMU
id = "Janus"  # not an id, just a NAME
downloadURL = 'http://iodp.org'  # no download URL, just an "info" URL
instrument = "measurement:thermalconductivity"  # the type of measurements being read from the database
# put an organization with this?
doc.dataset(id,  [downloadURL], [instrument])

# software
software = "eos:ocdJanus"  # The ETL code...
# put a github URl with this?  (have one)
# put an organization with this?
doc.software(software)

# output dataset
out_id = "9_83A_JanusCoreSummary_IAoFPcse.csv"
out_doi = "10.5067/ARIAMH/CSK/Interferogram"
out_accessURL = 'http://opencoredata.org/id/dataset/3671b954-d967-47ac-a21e-70bf9c62d134'
out_downloadURL = 'http://opencoredata.org/doc/dataset/3671b954-d967-47ac-a21e-70bf9c62d134.csv'
# put an organization with this?
doc.dataset(out_id, out_doi, [out_downloadURL], [instrument])


print doc.serialize(indent=2)