Command line¶
pynrpf run readings.csv --out results/
Reads a CSV with the four columns (substation_id, timestamp, net_load_MW, solar_MW
by default) and writes four files to the output folder:
| file | content |
|---|---|
site_days.csv |
the site-day table |
intervals.csv |
the interval table |
summary.json |
days per outcome, MWh flipped, the control and calibration used |
provenance.json |
the package version and the settings that produced the tables: calibration, evidence floor, control |
Show details
Options¶
| option | meaning | default |
|---|---|---|
--c |
the control | 0.7 |
--calibration a b |
intercept and slope replacing the release calibration, two numbers | release pair |
--phi |
the evidence floor, MW | release value |
--site, --timestamp, --net-load, --solar |
column names | substation_id, timestamp, net_load_MW, solar_MW |
pynrpf run readings.csv --out results/ --c 0.9 --site site_id --net-load mw --solar pv
pynrpf run readings.csv --out results/ --calibration -4.2 2.16 # a refitted intercept
pynrpf --version
Reproducing the paper's rows¶
The same command reproduces the package's part of the paper: run it on
publication/2_journal_article/dataset/final/dataset_beta.parquet exported to CSV, and the
site-day rows equal the reference run's M9 rows for the release calibration.