HCAI-Lab/w2-consensus-deepdive-unlearning-artifacts / social-data-attribution-w2 /scripts /modal /run_materialize_sample.sh
| set -euo pipefail | |
| # Materialize a sample and upload to R2, HF Buckets, and/or HF Datasets. | |
| # | |
| # Usage: | |
| # bash scripts/modal/run_materialize_sample.sh data/samples/sample_500_docs | |
| # bash scripts/modal/run_materialize_sample.sh data/samples/sample_500_docs 128 --upload-only | |
| # UPLOAD_TO=hf-buckets bash scripts/modal/run_materialize_sample.sh data/samples/sample_500_docs | |
| # UPLOAD_TO=r2,hf-buckets bash scripts/modal/run_materialize_sample.sh data/samples/sample_500_docs | |
| # UPLOAD_TO=r2,hf bash scripts/modal/run_materialize_sample.sh data/samples/sample_500_docs | |
| SAMPLE_DIR="${1:?Usage: run_materialize_sample.sh <sample_dir> [chunk_count] [--upload-only]}" | |
| CHUNK_COUNT="${2:-128}" | |
| HF_ORG="${HF_ORG:-HCAI-Lab}" | |
| UPLOAD_TO="${UPLOAD_TO:-hf-buckets}" | |
| EXTRA_FLAGS="" | |
| for arg in "$@"; do | |
| if [ "$arg" = "--upload-only" ]; then | |
| EXTRA_FLAGS="$EXTRA_FLAGS --upload-only" | |
| fi | |
| done | |
| SAMPLE_NAME=$(basename "$SAMPLE_DIR") | |
| VOLUME_NAME="soc134-output-$(echo "$SAMPLE_NAME" | tr '_' '-')" | |
| HF_REPO="${HF_ORG}/dolma3-6t-${SAMPLE_NAME}" | |
| HF_BUCKET_ID="${HF_BUCKET_ID:-${HF_ORG}/dolma3-6t-$(echo "$SAMPLE_NAME" | tr '_' '-')}" | |
| MANIFEST="${SAMPLE_DIR}/working_sample_manifest.parquet" | |
| if [ ! -f "$MANIFEST" ]; then | |
| echo "ERROR: manifest not found: $MANIFEST" | |
| exit 1 | |
| fi | |
| echo "Sample: $SAMPLE_NAME" | |
| echo "Volume: $VOLUME_NAME" | |
| echo "Upload to: $UPLOAD_TO" | |
| echo "Chunks: $CHUNK_COUNT" | |
| if echo "$UPLOAD_TO" | grep -q "hf-buckets"; then | |
| echo "HF bucket: $HF_BUCKET_ID" | |
| elif echo "$UPLOAD_TO" | grep -q "hf"; then | |
| echo "HF repo: $HF_REPO" | |
| fi | |
| if echo "$UPLOAD_TO" | grep -q "r2"; then | |
| echo "R2 prefix: soc134-samples/$SAMPLE_NAME/" | |
| fi | |
| echo "" | |
| SOC134_OUTPUT_VOLUME="$VOLUME_NAME" \ | |
| uv run modal run scripts/modal/materialize_working_sample.py \ | |
| --sample-manifest "$MANIFEST" \ | |
| --sample-name "$SAMPLE_NAME" \ | |
| --hf-repo "$HF_REPO" \ | |
| --hf-bucket-id "$HF_BUCKET_ID" \ | |
| --chunk-count "$CHUNK_COUNT" \ | |
| --upload-to "$UPLOAD_TO" \ | |
| $EXTRA_FLAGS | |
Xet Storage Details
- Size:
- 2.05 kB
- Xet hash:
- b077fe8d1ba44473b29c7da8d5210c34f41d6da10e4496f8c868a33501225bbd
·
Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.