v0.25.3

What's Changed

• chore: Generate Dockerfiles for v0.25.2 by @paradedb-github-bot[bot] in #5930
• fix: preserve AggregateScan partition_by for range co-partitioning by @paradedb-github-bot[bot] in #5938
• chore: Dedupe propagation of partition_by in the aggregate scan by @paradedb-github-bot[bot] in #5940
• fix: Declare only the assigned partition when scaling a PgSearchScan node by @paradedb-github-bot[bot] in #5941
• fix: Fix dropped pdb.agg filters and improve warnings by @paradedb-github-bot[bot] in #5942
• chore: Sync to rebased Tantivy by @paradedb-github-bot[bot] in #5946
• chore: Add range co-partitioned join benchmarks by @paradedb-github-bot[bot] in #5959
• ci: ship macOS 26 (Tahoe) binaries and drop macOS 14 (Sonoma) (#5812) by @philippemnoel in #5967
• fix: reject unsortable top_hits.sort fields at plan time by @paradedb-github-bot[bot] in #5966
• fix: Run final build merge even when the last insert flushed at the doc cap by @paradedb-github-bot[bot] in #5972
• feat: Attribute buffer hits/reads to routing vs probing for vector search by @paradedb-github-bot[bot] in #5970
• feat: solve Param/SubPlan expressions on leader before MPP dispatch by @paradedb-github-bot[bot] in #5973
• perf: Fast path for unfiltered / single term filter counts by @paradedb-github-bot[bot] in #5975
• perf: Lazy vischeck for cardinality aggs over string fields by @paradedb-github-bot[bot] in #5974
• feat: added aggregate pushdown for NUMERIC columns. by @paradedb-github-bot[bot] in #5979
• chore: Post 0.25.2 release. by @rebasedming in #5984
• chore: Prepare 0.25.3. by @paradedb-github-bot[bot] in #5983

Full Changelog: v0.25.2...v0.25.3