Skip to content

cli.md

Latest commit

 

History

History
70 lines (47 loc) · 2.01 KB

cli.md

File metadata and controls

70 lines (47 loc) · 2.01 KB

CLI Reference

boa-forecaster ships with a Click-based command-line entry point registered as the console script boa-forecaster (also reachable as python -m boa_forecaster).

Installation

pip install -e .

After installation, the boa-forecaster command is available on PATH.

Subcommands

run

Optimise the active model declared in config.yaml and emit a forecast.

boa-forecaster run --config config.yaml --output out/

Outputs, written to --output:

File Contents
forecast.csv date,forecast — point forecast over the optimiser horizon
params.json best_params + best_score from the TPE study
plot.png Matplotlib plot (only when matplotlib is installed)

Flags:

  • --n-trials INT — override optimization.n_calls.

validate

Run walk-forward (expanding-window) cross-validation for the active model.

boa-forecaster validate --config config.yaml --output out/ \
    --n-folds 3 --test-size 12 --min-train-size 24

Writes folds.csv (one row per fold, with sMAPE/RMSLE) and metrics.csv (aggregated mean/std), plus the same forecast artefacts as run.

compare

Optimise every models.* entry with enabled: true plus zero-budget baselines (Seasonal-Naïve, ETS) and rank them on the same walk-forward folds.

boa-forecaster compare --config config.yaml --output out/

Writes metrics.csv with one row per model and columns model, sMAPE_mean, sMAPE_std, RMSLE_mean, RMSLE_std.

Use --no-baselines to skip Seasonal-Naïve and ETS.

Configuration

All subcommands consume a YAML file validated by boa_forecaster.config_schema.BoaConfig. Copy config.example.yaml and edit. Typos in field names are tolerated in v2.x (warns via Pydantic extra="allow") but will raise ValidationError once the v3.0 strict-mode switch flips.