Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
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
44 changes: 44 additions & 0 deletions .github/workflows/benchmark.yml
Original file line number Diff line number Diff line change
Expand Up @@ -130,6 +130,50 @@ jobs:
--parser-url http://localhost:8080 \
--parser-image "${{ steps.image_tag.outputs.value }}"

- name: Checkout predictions repo
if: github.ref == 'refs/heads/main'
uses: actions/checkout@v5
with:
repository: elifepathways/sciencebeam-eval-predictions
path: sciencebeam-eval-predictions
token: ${{ secrets.PREDICTIONS_REPO_PAT }}
fetch-depth: 1
filter: blob:none
sparse-checkout: .gitignore
sparse-checkout-cone-mode: false

- name: Push ScienceBeam Parser predictions
if: github.ref == 'refs/heads/main'
env:
IMAGE_TAG: ${{ steps.image_tag.outputs.value }}
SPLIT: ${{ env.BENCHMARK_SPLIT }}
MODE: ${{ env.BENCHMARK_MODE }}
run: |
VERSION="${IMAGE_TAG#sciencebeam-parser:}"
DEST="sciencebeam-eval-predictions/sciencebeam-parser/${VERSION}/${SPLIT}"
PREDICTIONS_SRC="${{ env.BENCHMARK_RUN }}/predictions"
for CORPUS_DIR in "${PREDICTIONS_SRC}"/*/; do
[ -d "${CORPUS_DIR}" ] || continue
CORPUS=$(basename "${CORPUS_DIR}")
VARIANT=$(python3 -c "import yaml; c=yaml.safe_load(open('benchmarks/eval.yml')); print(c['dataset']['splits'].get('${SPLIT}',{}).get('${CORPUS}',{}).get('variant','v1'))")
mkdir -p "${DEST}/${CORPUS}/${VARIANT}"
cp -r "${CORPUS_DIR}/." "${DEST}/${CORPUS}/${VARIANT}/"
done
[ -f "${PREDICTIONS_SRC}/manifest.jsonl" ] && cp "${PREDICTIONS_SRC}/manifest.jsonl" "${DEST}/"
DATE=$(date --utc +%Y-%m-%dT%H:%M:%SZ)
echo "{\"tool\":\"sciencebeam-parser\",\"version\":\"${VERSION}\",\"split\":\"${SPLIT}\",\"mode\":\"${MODE}\",\"generated_at\":\"${DATE}\"}" \
| python3 -m json.tool > "${DEST}/metadata.json"
cd sciencebeam-eval-predictions
git config user.name "github-actions[bot]"
git config user.email "github-actions[bot]@users.noreply.github.com"
git add --sparse .
if git diff --cached --quiet; then
echo "No new predictions to push"
else
git commit -m "Add sciencebeam-parser/${VERSION} ${{ env.BENCHMARK_SPLIT }} predictions (${{ env.BENCHMARK_MODE }})"
git push
fi

- name: Run score
run: |
uv run python -m benchmarks.score \
Expand Down
2 changes: 2 additions & 0 deletions benchmarks/eval.yml
Original file line number Diff line number Diff line change
Expand Up @@ -6,10 +6,12 @@ dataset:
biorxiv:
file: biorxiv-jats/train-00000-of-00001.parquet
id_column: ppr_id
variant: v1
validation:
biorxiv:
file: biorxiv-jats/validation-00000-of-00001.parquet
id_column: ppr_id
variant: v1

sampling:
smoke:
Expand Down
Loading