#!/usr/bin/env bash
# Uploads eth/data/*.pdf to the demo S3 bucket. Curate eth/data/ by hand
# first — the script never reads ~/Documents and has no filter logic of its
# own. Idempotent: aws s3 sync skips files already present at the same size.
set -euo pipefail

REPO_ROOT="$(cd "$(dirname "$0")/.." && pwd)"
DATA_DIR="${DATA_DIR:-$REPO_ROOT/data}"
PREFIX="${PREFIX:-2026/04/}"
BUCKET="${BUCKET_NAME:-eth-demo-reports-$(aws sts get-caller-identity --query Account --output text)}"

if [ ! -d "$DATA_DIR" ]; then
  echo "no such directory: $DATA_DIR" >&2
  exit 2
fi

pdf_count=$(find "$DATA_DIR" -maxdepth 1 -name "*.pdf" | wc -l)
if [ "$pdf_count" -eq 0 ]; then
  echo "no .pdf files in $DATA_DIR — curate first" >&2
  exit 2
fi

echo "syncing $pdf_count PDFs from $DATA_DIR to s3://${BUCKET}/${PREFIX}"
aws s3 sync "$DATA_DIR" "s3://${BUCKET}/${PREFIX}" \
  --exclude "*" --include "*.pdf"

echo
echo "done. verify with:"
echo "  aws s3 ls s3://${BUCKET}/${PREFIX} --human-readable | tail"