lsst/rubin_scheduler: v3.21.0

v3.21.0 The primary focus for these updates include better handling for ToO events. The ToO observations are now labelled with the superevent ID as the identifier, instead of a simple integer. Detailers are available to better cover the footprint for large ToOs (prioritizing setting regions of the footprint first, and chunking coverage by hour angle). A bugfix handles ToO dithering night-to-night, even if only one observation is requested in a given night. The options to automatically split long exposures into snaps have been removed; snaps are not supported for observations any longer; observation exposure times should just be set explicitly for the ToO surveys. A detailer now handles substituting bandpasses based on whether z or y band is available, instead of handling this within the ToO event. This enables better matching of acquired observations, as well as explicitly making the band substitution clear. An additional update from a new contributor now allows retry and resume after partially successful download attempts. This is very welcome - data download has been somewhat unreliable and this should help mitigate the situation. The skybrightness test file was also updated, removing a small gap when transitioning from the test data to a larger data file for simulations. A bugfix incompatibility with numpy 2.4 was found and fixed - 3.21.0 is now compatible with numpy 2.4. What's Changed Add retry and partial download resumption to data downloader by @GUIpsp in https://github.com/lsst/rubin_scheduler/pull/241 Tickets/sp 2790 - remove split_long options and add BandSubstituteDetailer by @yoachim in https://github.com/lsst/rubin_scheduler/pull/243 ToO dithering even if only 1 position by @yoachim in https://github.com/lsst/rubin_scheduler/pull/247 update small sybrightness file to not have a gap by @yoachim in https://github.com/lsst/rubin_scheduler/pull/245 Make sure detailers are note applied twice in ScriptedSurvey by @yoachim in https://github.com/lsst/rubin_scheduler/pull/246 numpy in1d deprecated by @yoachim in https://github.com/lsst/rubin_scheduler/pull/250 SP-2789 - handle banana ToO footprints better by @yoachim in https://github.com/lsst/rubin_scheduler/pull/244 SP-2890: fix DDF filter thrashing from ToO survey updates by @yoachim in https://github.com/lsst/rubin_scheduler/pull/252 New Contributors @GUIpsp made their first contribution in https://github.com/lsst/rubin_scheduler/pull/241 Full Changelog: https://github.com/lsst/rubin_scheduler/compare/v3.20.1...v3.21.0