mirror of
https://github.com/ipnet-mesh/meshcore-hub.git
synced 2026-06-22 19:15:07 +02:00
Louis King
e1199a42cd
Two read-only query optimisations, no schema changes. node-count history: replace the per-day COUNT(*) loop (up to 90 full scans of the unindexed created_at column) with two queries — a baseline count of nodes created before the window plus one GROUP BY date() aggregate, accumulated into the running total in Python. Results are identical; the baseline seed keeps pre-window nodes counted from day 0. sender-name resolution: add resolve_sender_names() to observer_utils, batching all pubkey prefixes into two queries (names + name tags) via an OR of indexable LIKE 'prefix%' terms instead of two queries per prefix. Wire it into list_messages (was ~2xN per page) and the dashboard channel-messages loop (nested per channel x per prefix). The dashboard recent-ads block already batches on full public keys via IN(), so it is left as-is. Tests: add cumulative+baseline correctness for node-count and a multi-sender batched-resolution case for messages. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>