Add comprehensive ANN benchmarking suite

Extend benchmarks-ann/ with results database (SQLite with per-query detail
and continuous writes), dataset subfolder organization, --subset-size and
--warmup options. Supports systematic comparison across flat, rescore, IVF,
and DiskANN index types.

benchmarks-ann/datasets/nyt-768/Makefile

@@ -0,0 +1,37 @@
+MODEL ?= bge-base-en-v1.5-768
+K ?= 100
+BATCH_SIZE ?= 512
+DATA_DIR ?= ../nyt/data
+
+all: base.db
+
+# Reuse data from ../nyt
+$(DATA_DIR):
+	$(MAKE) -C ../nyt data
+
+# Distill model (separate step, may take a while)
+$(MODEL):
+	uv run distill-model.py
+
+contents.db: $(DATA_DIR)
+	uv run build-contents.py --data-dir $(DATA_DIR) -o $@
+
+base.db: contents.db queries.txt $(MODEL)
+	uv run ../nyt/build-base.py \
+		--contents-db contents.db \
+		--model $(MODEL) \
+		--queries-file queries.txt \
+		--batch-size $(BATCH_SIZE) \
+		--k $(K) \
+		-o $@
+
+queries.txt:
+	cp ../nyt/queries.txt $@
+
+clean:
+	rm -f base.db contents.db
+
+clean-all: clean
+	rm -rf $(MODEL)
+
+.PHONY: all clean clean-all