feat(plots): show individual-fit forecasts on overlay plots

[seabbs-bot]

Summary

The forecast-overlay plots (plot_forecast_overlay) currently show only the joint model's forecasts against the persistence baseline. The stream_forecasts.csv asset carries per-fit forecasts (cases, deaths, confirmed, confirmed deaths, treatment, exports) but the overlay plot does not display them.

Current behaviour

The overlay plot iterates over role_order = ["baseline", "individual", "joint"] and colours each by role. The _fit_role function maps non-baseline, non-joint fits to "individual". So the code already supports individual fits — but the data may not carry them.

What's needed

Once the stream_forecasts.csv asset is available in releases (or the backfill is extended), the overlay plot should show the individual-fit forecasts alongside the joint and baseline. The role_order and role_colour already exist — the fix is likely:

1. Ensure the overlay data includes individual-fit rows (from stream_forecasts.csv)
2. Verify the _fit_role mapping correctly colours them
3. Add a legend entry for individual fits

Location

• src/plots.jl — plot_forecast_overlay function
• scripts/score_releases.jl — overlay table construction

Priority

Medium — blocked on releases carrying stream_forecasts.csv (see #463).