[Idea] Hierarchical categories or tags for forecasts

# Rationale

* Easily classify forecasts into groups and filter in the UI
* It seems very plausible that the user's calibration (how accurately they forecast) depends a lot on the domain e.g. they could be very well calibrated for local elections in Germany but badly calibrated for elections in the US. Or well calibrated for forecasts concerning the reliability of their friends but badly calibrated for work-related project deadlines. Therefore seeing the calibration per category would be valuable.
* Making the categories hierarchical allows drilling down and comparing "politics" forecasts as a whole to other categories but also "politics / germany" to "politics / netherlands".

# Notes

Forecasts can belong to multiple categories. Forecasts can also belong to no category (default) that should maybe not be displayed if the user does not use categories.

The categories should be user defined, but maybe it's valuable to provide some default categories (that can be deleted).

The issue names no files, tests, or entry points. Start by tracing how forecasts, calibration results, and UI filtering are currently represented, then determine where hierarchical, user-defined categories would fit. Done would require an agreed design covering multiple or uncategorized forecasts, hierarchy navigation, filtering, and per-category calibration.

Written by the indexing model from the issue text.