Skip to content

Fix uplift eval run isolation #27

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
04kash merged 2 commits into from
Sep 3, 2026
Merged

Fix uplift eval run isolation #27

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
93472ae
fix eval uplift run isolation
04kash Sep 3, 2026
b3fa66f
style: format uplift runner test
04kash Sep 3, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 5 additions & 4 deletions eval/rhdh-templates/uplift/run-local.sh
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,18 +24,19 @@ fi
cd "${REPO_ROOT}"
export UV_CACHE_DIR="${UV_CACHE_DIR:-/tmp/rhdh-aeh-uv-cache}"
export AGENT_EVAL_RUNS_DIR="${REPO_ROOT}/eval/runs/rhdh-templates-uplift"
prefix="uplift-$(date -u +%Y%m%dT%H%M%SZ)"
for arm in skill baseline; do
config="${SCRIPT_DIR}/${arm}.eval.yaml"
for index in $(seq 1 "${runs}"); do
run_id="uplift-${arm}-${index}"
run_id="${prefix}-${arm}-${index}"
uv run --project "${aeh_checkout}" python "${SCRIPT_DIR}/run.py" --aeh-dir "${aeh_checkout}" --config "${config}" --model "${MODEL}" --run-id "${run_id}"
done
done
comparison_dir="/tmp/rhdh-templates-uplift-comparison"
comparison_dir="/tmp/rhdh-templates-uplift-comparison-${prefix}"
comparison_input="$(mktemp -d /tmp/rhdh-templates-uplift-runs.XXXXXX)"
for index in $(seq 1 "${runs}"); do
cp -a "${AGENT_EVAL_RUNS_DIR}/rhdh-templates/uplift-skill-${index}" "${comparison_input}/"
cp -a "${AGENT_EVAL_RUNS_DIR}/rhdh-templates-uplift-baseline/uplift-baseline-${index}" "${comparison_input}/"
cp -a "${AGENT_EVAL_RUNS_DIR}/rhdh-templates/${prefix}-skill-${index}" "${comparison_input}/"
cp -a "${AGENT_EVAL_RUNS_DIR}/rhdh-templates-uplift-baseline/${prefix}-baseline-${index}" "${comparison_input}/"
done
uv run --project "${aeh_checkout}" python "${aeh_checkout}/skills/eval-compare/scripts/compare.py" generate "${comparison_input}" --output "${comparison_dir}" --title "rhdh-templates uplift" --overview "Observed comparison only; no generalization beyond this reviewed case set."
echo "Comparison: ${comparison_dir}"
4 changes: 4 additions & 0 deletions tests/unit/test_rhdh_templates_nunjucks_uplift.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -77,5 +77,9 @@ def test_nunjucks_uplift_runs_three_local_attempts_per_arm_with_case_comparison(
assert "runs=3" in command
assert "for arm in skill baseline; do" in command
assert 'for index in $(seq 1 "${runs}"); do' in command
assert 'prefix="uplift-$(date -u +%Y%m%dT%H%M%SZ)"' in command
assert 'run_id="${prefix}-${arm}-${index}"' in command
assert "${prefix}-skill-${index}" in command
assert "${prefix}-baseline-${index}" in command
assert "compare.py" in command
assert "Observed comparison only" in command
Loading