Reading The Outputs

The Quickstart run wrote six files under projects/minimal_grid_project/outputs/. This page reads them, so the shapes introduced in Foundation and Twin have a concrete example before you meet them in the abstract.

outputs/
  data/
    buses.csv
    lines.csv
    loads.csv
  figures/
    minimal_voltage_profile.png
  manifests/
    project_run_manifest.json
  reports/
    minimal_grid_report.json

The manifest — what ran, and in what order

python3 -m json.tool projects/minimal_grid_project/outputs/manifests/project_run_manifest.json

Two stages, in order, each with its own exit code:

"stages": [
  {"id": "prepare_workspace", "status": "completed", "exit_code": 0},
  {"id": "run_minimal_powerflow", "status": "completed", "exit_code": 0}
]

The overall "status": "completed" is true only because both stages exited zero — one non-zero exit anywhere would have marked both that stage and the run "failed". Under provenance, three facts are recorded that the run itself does not print: which power-flow backend actually solved it ("backend_id": "pandapower_native"), which macro load model was available ("macro_model": {"expected": "lgbm", "lightgbm_runtime": true}), and the declared RNG seed ("seeds": {"base": 7}). Two runs on two machines can be compared on these three facts without re-running anything.

The report — the governed summary

python3 -m json.tool projects/minimal_grid_project/outputs/reports/minimal_grid_report.json

The eight fields Foundation describes are all here — report_id, schema_version, created_at, source_domain, inputs, artifacts, summary, validation — plus governance and project. Two parts matter most for a first read:

• artifacts lists every file this run produced, each with its byte count and SHA-256 — not a description of the files, a fingerprint of them.
• summary is where the actual physics landed: "bus_count": 5, "converged": true, "min_voltage_pu": 1.0097973462536918, "max_line_loading_percent": 2.834697084369379. "validation": {"valid": true, "errors": [], "warnings": []} is what a sense check (Projects) actually decided about this run.

The data — the network itself

head -3 projects/minimal_grid_project/outputs/data/buses.csv

bus_id,name,vn_kv,type,zone,in_service,geo,vm_pu,va_degree,p_mw,q_mvar
0,bus_00,12.47,b,,True,"{""coordinates"":[0.0,0.0], ""type"":""Point""}",1.01,0.0,-0.15002099163838176,-0.03726438609421171
1,bus_01,12.47,b,,True,"{""coordinates"":[1.0,0.15], ""type"":""Point""}",1.0099224498179593,-0.0014955731625605326,0.029971536292357482,0.0074928840730893705

Five buses, four lines, four loads — the summary's counts and this table are the same fact told twice, once as a number and once as rows you can open in a spreadsheet. vm_pu is the solved voltage in per-unit; both buses shown here sit at essentially 1.01 p.u., which is why min_voltage_pu in the report reads just under 1.01 rather than something alarming.

The figure

outputs/figures/minimal_voltage_profile.png plots exactly the vm_pu column you just read, per bus, so the number in the report and the shape of the line are two views of the same solve — open it and confirm the flat line matches the tight, near-1.01 voltages above.

Where this leads

The tables you just read (buses, lines) are the same shape Twin describes for the full base — this project's buses.csv is a five-row instance of the schema that page states in general. Read Twin next, or Run Demo Projects to see the same six-artifact shape scale up to a real study.