#!/bin/zsh
# =============================================================================
#  Project   : anomaly-atlas
#  File      : web/sync-content.sh
#  Purpose   : Snapshot the research repo into web/content/ + build-info.json
#  Author    : Simon-Pierre Boucher
#  Contact   : contact@spboucher.ai
#  Data src  : hfmarketdata.io (sole data source)
#  Created   : 2026-08-12
#  Modified  : 2026-08-12
#  Platform  : macOS / Apple Silicon (arm64)
#  License   : All rights reserved (research code)
# =============================================================================
set -euo pipefail

REPO_ROOT="$(cd "$(dirname "$0")/.." && pwd)"
CONTENT="$REPO_ROOT/web/content"

mkdir -p "$CONTENT"
# NOTE: rsync applies filters in order — excludes MUST precede the broad
# includes, or large artifacts slip through.
rsync -a --delete \
  --exclude='*.parquet' --exclude='*.duckdb' --exclude='*.log' \
  --exclude='__pycache__' --exclude='.git' --exclude='.venv' \
  --exclude='node_modules' --exclude='web' --exclude='data' \
  --include='CLAUDE.md' --include='README.md' --include='CITATION.cff' \
  --include='Makefile' --include='pyproject.toml' \
  --include='research/***' --include='experiments/***' --include='benchmarks/***' \
  --include='src/***' --include='tools/***' --include='results/***' \
  --include='atlas/***' --include='docs/***' \
  --exclude='*' \
  "$REPO_ROOT/" "$CONTENT/"

COMMIT=$(git -C "$REPO_ROOT" rev-parse HEAD 2>/dev/null || echo "unknown")
COMMITS=$(git -C "$REPO_ROOT" rev-list --count HEAD 2>/dev/null || echo 0)
cat > "$CONTENT/build-info.json" <<EOF
{
  "author": "Simon-Pierre Boucher",
  "contact": "contact@spboucher.ai",
  "data_source": "hfmarketdata.io",
  "commit": "$COMMIT",
  "commits": $COMMITS,
  "generated": "$(date -u +%Y-%m-%dT%H:%M:%SZ)"
}
EOF
echo "content synced at commit ${COMMIT:0:7} ($COMMITS commits)"