Bulk entity access for table.joinByKey's tag/property strategies,
which reuse @ifc-lite/mutations' csv-match.ts index builder verbatim
(issue #5230) instead of re-implementing tag/property matching. That
builder scans a whole EntityTable once (O(entities + rows)) rather than
once per candidate entity, so it needs the raw table, not BimContext's
per-EntityRef accessors.
Optional on FlowHost: a host that cannot cheaply provide this (nothing
today besides the CLI's HeadlessBackend) simply cannot run those two
match strategies, and table.joinByKey reports that plainly rather than
falling back to a slow re-implementation.
Bulk entity access for
table.joinByKey'stag/propertystrategies, which reuse@ifc-lite/mutations'csv-match.tsindex builder verbatim (issue #5230) instead of re-implementing tag/property matching. That builder scans a wholeEntityTableonce (O(entities + rows)) rather than once per candidate entity, so it needs the raw table, notBimContext's per-EntityRefaccessors.Optional on
FlowHost: a host that cannot cheaply provide this (nothing today besides the CLI'sHeadlessBackend) simply cannot run those two match strategies, andtable.joinByKeyreports that plainly rather than falling back to a slow re-implementation.