Add vector-stores module: one dataset through VectorStore on pgvector, Redis, Qdrant and Elasticsearch

Ingest time, recall@10 against exact search, latency, metadata filtering and running cost, with
the defaults that cost recall reproduced and every figure written to output/.

vector-stores/scripts/run-all.sh

@@ -0,0 +1,9 @@
+#!/usr/bin/env bash
+# Regenerates every file in output/. Needs the four services: scripts/services-up.sh (no Docker).
+# A store whose port is closed is skipped, so the run still completes with fewer rows.
+set -euo pipefail
+cd "$(dirname "$0")/.."
+mkdir -p output && find output -type f ! -name "13-*" -delete   # 13 is a one-off capture, see its header
+mvn -B -q clean test
+scripts/capture-1x-compile.sh >/dev/null
+ls output