#!/usr/bin/env bash
set -euo pipefail

# Implementation for cameron-site-content-sync.service. Normal operators should
# start the credential-provisioned systemd unit instead of invoking this script
# from a shell that may deploy Fly and then lack the PDS credential.

REPO=${CAMERON_SITE_REPO:-/home/cameron/code/cameron-site-tangled}
STATE_DIR=${CAMERON_SITE_SYNC_STATE_DIR:-/home/cameron/.local/state/cameron-site}
STATE_FILE="$STATE_DIR/content-source-fingerprint"
LOCK_FILE="$STATE_DIR/content-sync.lock"
REPORT_PATH="$STATE_DIR/last-run.json"
DEPENDENCY_STATE_FILE="$STATE_DIR/dependency-source-fingerprint"

if [[ -z ${CAMERON_BSKY_APP_PASSWORD:-} ]]; then
  echo "Refusing credential-dark content sync before Git, Fly, or PDS mutation."
  echo "Start cameron-site-content-sync.service so the private worker environment is loaded."
  exit 2
fi
readonly PDS_APP_PASSWORD=$CAMERON_BSKY_APP_PASSWORD
unset CAMERON_BSKY_APP_PASSWORD

mkdir -p "$STATE_DIR"
exec 9>"$LOCK_FILE"
flock -n 9 || {
  echo "Another Cameron.site content sync is already running."
  exit 0
}

cd "$REPO"

retry() {
  local attempt delay
  delay=3
  for attempt in 1 2 3 4 5; do
    if "$@"; then return 0; fi
    if [[ $attempt -eq 5 ]]; then return 1; fi
    sleep "$delay"
    delay=$((delay * 3))
  done
}

changed_paths() {
  git status --porcelain | sed -E 's/^.. //' | sort -u
}

assert_receipt_only_dirty() {
  local unexpected
  unexpected=$(changed_paths | grep -Ev '^(content/about-atproto-manifest\.json|knowledge/atproto-manifest\.json)$' || true)
  if [[ -n "$unexpected" ]]; then
    echo "Refusing dirty canonical checkout; unexpected paths:"
    printf '%s\n' "$unexpected"
    exit 2
  fi
}

commit_receipts() {
  if git diff --quiet -- content/about-atproto-manifest.json knowledge/atproto-manifest.json; then
    return 0
  fi
  assert_receipt_only_dirty
  git add content/about-atproto-manifest.json knowledge/atproto-manifest.json
  git commit -m 'Record ATProto content projection receipts.

👾 Generated with [Letta Code](https://letta.com)

Co-Authored-By: Letta Code <noreply@letta.com>'
  retry git push origin main
}

if [[ $(git branch --show-current) != main ]]; then
  echo "Canonical content sync requires the main branch."
  exit 2
fi

# Recover a prior successful PDS write whose receipt commit or push was interrupted.
if [[ -n $(changed_paths) ]]; then
  assert_receipt_only_dirty
  commit_receipts
fi

retry git fetch origin main
local_head=$(git rev-parse HEAD)
remote_head=$(git rev-parse origin/main)
if [[ "$local_head" != "$remote_head" ]]; then
  merge_base=$(git merge-base HEAD origin/main)
  if [[ "$merge_base" == "$remote_head" ]]; then
    unexpected=$(
      git diff --name-only origin/main..HEAD \
      | grep -Ev '^(content/about-atproto-manifest\.json|knowledge/atproto-manifest\.json)$' \
      || true
    )
    if [[ -n "$unexpected" ]]; then
      echo "Local main is ahead with non-receipt changes; refusing automatic push:"
      printf '%s\n' "$unexpected"
      exit 2
    fi
    retry git push origin main
    remote_head=$local_head
  elif [[ "$merge_base" != "$local_head" ]]; then
    echo "Canonical main diverged from origin/main; refusing automatic reconciliation."
    exit 2
  else
    git merge --ff-only origin/main
  fi
fi

if [[ -n $(changed_paths) ]]; then
  echo "Canonical checkout became dirty during fast-forward."
  exit 2
fi

if [[ ! -f package-lock.json ]]; then
  echo "The deploy checkout requires package-lock.json for reproducible dependency installation."
  exit 2
fi

dependency_fingerprint=$(
  sha256sum package.json package-lock.json \
  | sha256sum \
  | cut -d' ' -f1
)
installed_dependency_fingerprint=$(cat "$DEPENDENCY_STATE_FILE" 2>/dev/null || true)
if [[ ! -x node_modules/.bin/tsx || "$dependency_fingerprint" != "$installed_dependency_fingerprint" ]]; then
  npm ci --no-audit --no-fund
  printf '%s\n' "$dependency_fingerprint" >"$DEPENDENCY_STATE_FILE.tmp"
  mv "$DEPENDENCY_STATE_FILE.tmp" "$DEPENDENCY_STATE_FILE"
fi

pnpm typecheck
pnpm test:markdown
pnpm test:content
pnpm test:charts
pnpm test:worker
if ! pnpm --silent knowledge:check >"$STATE_DIR/knowledge-check.json"; then
  cat "$STATE_DIR/knowledge-check.json"
  exit 1
fi

source_fingerprint=$(
  git ls-files -z \
    'content/about.md' 'knowledge/published/**' 'knowledge/policy.json' \
    'src/**' 'public/**' 'package.json' 'package-lock.json' 'Dockerfile' 'fly.toml' \
    ':(exclude)content/about-atproto-manifest.json' \
    ':(exclude)knowledge/atproto-manifest.json' \
  | sort -z \
  | xargs -0 sha256sum \
  | sha256sum \
  | cut -d' ' -f1
)
last_fingerprint=$(cat "$STATE_FILE" 2>/dev/null || true)
if [[ "$source_fingerprint" != "$last_fingerprint" ]]; then
  cleanup_fly_agent() {
    timeout 10s fly agent stop >/dev/null 2>&1 || true
  }
  trap cleanup_fly_agent EXIT
  fly deploy --depot=false
  cleanup_fly_agent
  trap - EXIT
  printf '%s\n' "$source_fingerprint" >"$STATE_FILE.tmp"
  mv "$STATE_FILE.tmp" "$STATE_FILE"
fi

pnpm --silent about:sync >"$STATE_DIR/about-plan.json"
python3 - "$STATE_DIR/about-plan.json" <<'PY'
import json, sys
p = json.load(open(sys.argv[1]))
if p["counts"]["conflict"]:
    raise SystemExit(f'About sync has {p["counts"]["conflict"]} conflict(s)')
PY
CAMERON_BSKY_APP_PASSWORD="$PDS_APP_PASSWORD" \
  pnpm --silent about:sync --apply --from-origin-main >"$STATE_DIR/about-apply.json"
commit_receipts

pnpm --silent knowledge:sync --all >"$STATE_DIR/knowledge-plan.json"
python3 - "$STATE_DIR/knowledge-plan.json" <<'PY'
import json, sys
p = json.load(open(sys.argv[1]))
if p["counts"]["conflict"]:
    raise SystemExit(f'Knowledge sync has {p["counts"]["conflict"]} conflict(s)')
PY
CAMERON_BSKY_APP_PASSWORD="$PDS_APP_PASSWORD" \
  pnpm --silent knowledge:sync --all --apply --from-origin-main >"$STATE_DIR/knowledge-apply.json"
commit_receipts

head=$(git rev-parse HEAD)
python3 - "$STATE_DIR/about-plan.json" "$STATE_DIR/knowledge-plan.json" "$REPORT_PATH.tmp" "$head" "$source_fingerprint" <<'PY'
import datetime, json, sys
about = json.load(open(sys.argv[1]))
knowledge = json.load(open(sys.argv[2]))
report = {
    "completedAt": datetime.datetime.now(datetime.timezone.utc).isoformat(),
    "head": sys.argv[4],
    "sourceFingerprint": sys.argv[5],
    "about": about["counts"],
    "knowledge": knowledge["counts"],
    "liveUrl": "https://cameron.stream/",
}
with open(sys.argv[3], "w") as handle:
    json.dump(report, handle, indent=2)
    handle.write("\n")
PY
mv "$REPORT_PATH.tmp" "$REPORT_PATH"

git status --short --branch
cat "$REPORT_PATH"