Migrate to the PowerIO parser backend for broad non-MATPOWER format support

[samtalki]

This PR routes ExaModelsPower parsing through PowerIO.

This PR's branch currently retains ExaPowerIO as a dependency for PGLib artifact lookup, pending the guidance of @hfytr, @sshin23, @frapac .

The main change that this PR offers is broad format compatibility for ExaModelsPower beyond MATPOWER: through PowerIO, ExaModelsPower immediately supports PSS/E .raw, PowerWorld .aux, PSLF .epc, PowerModels .json, and more.

PowerIO has also already shipped distribution systems support as of v0.3.0 through the powerio-dist crate, creating a straightforward pathway for future distribution systems support in ExaModelsPower. (note that PowerIO has already been adopted as a dependency by BMOPFTools.jl)

At the same time, this PR branch is smaller than the current parser path.

Changes:

• add PowerIO as a parser dependency
• keep ExaPowerIO for PGLib fallback, if desired by the maintainers
• add from passthrough for ambiguous .json formats such as PowerModels vs. Egret
• update multi period load rewriting for NamedTuple bus rows

Current checks:

• implementation branch diff: net -39 LOC
• testing branch: https://github.com/samtalki/ExaModelsPower.jl/tree/agent/powerio-parser-backend-tests
• focused parser checks passed: PowerIO matched ExaPowerIO on MATPOWER and storage cases
• generated RAW, AUX, and PowerModels JSON companions parse and build AC/DC/MPOPF models
• local ACTIVSg200 RAW, AUX, and EPC companion cases parse and build DCOPF models
• Pkg.test("ExaModelsPower") passed locally

cc @hfytr: ExaPowerIO’s fantastic PowerData layout made this integration extremely straightforward. I would greatly value your review and expertise on this, and I would be happy to coordinate on the parser surface in PowerIO as well.

Unrelated Aside

The full local ExaModelsPower test suite passed, but took >25 minutes on my machine. Is this expected behavior?