# Evaluation helpers (sane defaults, two targets):
# make bench-lang # semantic language eval on lang-eval-dev @ k=10
# make bench-cluster # clustering eval on cluster-stress-dev
#
# Optional: override config file via CONFIG, e.g.:
# make bench-lang CONFIG=.chunkhound.json
.PHONY: bench-lang bench-cluster
bench-lang:
uv run python -m chunkhound.tools.eval_search \
--bench-id lang-eval-dev \
--mode mixed \
--search-mode semantic \
--languages all \
--k 10 \
$(if $(CONFIG),--config $(CONFIG),) \
--output .chunkhound/benches/lang-eval-dev/eval_semantic_k10.json
bench-cluster:
uv run python -m chunkhound.tools.eval_cluster \
--bench-id cluster-stress-dev \
$(if $(CONFIG),--config $(CONFIG),) \
--output .chunkhound/benches/cluster-stress-dev/cluster_eval.json
