Tools

A tool is one modelling framework's view of a record. Tools live under datarecord/tools/, are reached by importing them (no registry), and nothing in the core imports one — so importing the record layer pulls in no framework (design, module layout).

from datarecord.tools.pypsa import PyPSA

missing = PyPSA.verify(record)
if missing:
    raise RuntimeError(missing.describe())

n = PyPSA.build(record)  # a pypsa.Network
n.optimize()

record_back = PyPSA.to_datarecord(n)  # the inverse; a `Record`
results = PyPSA.results(n)  # Frames, in the long schema

verify catches what the record layer cannot — a component type the framework has no registry entry for, a connection role it cannot place, a partial set that breaks its constant-versus-varying split — and returns a falsy Requirements when the record is usable. build raises UnsupportedRecordError rather than producing a partial model.

Solving and writing results back

results returns the same Frames type Record.outputs presents, so a solve's output goes straight back:

w = WorkingRecord(record, con)
w.set("p_max_pu", 0.8, entity=["wind1"])
n = PyPSA.build(w)  # a WorkingRecord is a Record
n.optimize()
for attr, frame in PyPSA.results(n).items():
    w.set(attr, frame, kind="outputs")
w.commit(NewChild())  # one layer, inputs and results together

Two checks are skipped for kind="outputs": the attribute need not be schema-declared, and a result's entity need not resolve to a declared member — a solve may produce rows for a component type it derived rather than read (design).