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

usage() {
  cat <<'EOF'
Usage:
  ./scripts/cir/run_cir_replay_rpki_client.sh \
    --cir <path> \
    --repo-bytes-db <path> \
    --out-dir <path> \
    --reference-ccr <path> \
    [--build-dir <path> | --rpki-client-bin <path>] \
    [--keep-db] \
    [--real-rsync-bin <path>]
EOF
}

ROOT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")/../.." && pwd)"
CIR=""
REPO_BYTES_DB=""
OUT_DIR=""
REFERENCE_CCR=""
BUILD_DIR=""
RPKI_CLIENT_BIN=""
KEEP_DB=0
REAL_RSYNC_BIN="${REAL_RSYNC_BIN:-/usr/bin/rsync}"
CIR_MATERIALIZE_BIN="${CIR_MATERIALIZE_BIN:-$ROOT_DIR/target/release/cir_materialize}"
CIR_EXTRACT_INPUTS_BIN="${CIR_EXTRACT_INPUTS_BIN:-$ROOT_DIR/target/release/cir_extract_inputs}"
CCR_TO_COMPARE_VIEWS_BIN="${CCR_TO_COMPARE_VIEWS_BIN:-$ROOT_DIR/target/release/ccr_to_compare_views}"
WRAPPER="$ROOT_DIR/scripts/cir/cir-rsync-wrapper"

while [[ $# -gt 0 ]]; do
  case "$1" in
    --cir) CIR="$2"; shift 2 ;;
    --repo-bytes-db) REPO_BYTES_DB="$2"; shift 2 ;;
    --out-dir) OUT_DIR="$2"; shift 2 ;;
    --reference-ccr) REFERENCE_CCR="$2"; shift 2 ;;
    --build-dir) BUILD_DIR="$2"; shift 2 ;;
    --rpki-client-bin) RPKI_CLIENT_BIN="$2"; shift 2 ;;
    --keep-db) KEEP_DB=1; shift ;;
    --real-rsync-bin) REAL_RSYNC_BIN="$2"; shift 2 ;;
    -h|--help) usage; exit 0 ;;
    *) echo "unknown argument: $1" >&2; usage; exit 2 ;;
  esac
done

[[ -n "$CIR" && -n "$REPO_BYTES_DB" && -n "$OUT_DIR" && -n "$REFERENCE_CCR" ]] || {
  usage >&2
  exit 2
}
if [[ -z "$BUILD_DIR" && -z "$RPKI_CLIENT_BIN" ]]; then
  usage >&2
  exit 2
fi
if [[ -z "$RPKI_CLIENT_BIN" ]]; then
  RPKI_CLIENT_BIN="$BUILD_DIR/src/rpki-client"
fi
if [[ ! -x "$RPKI_CLIENT_BIN" ]]; then
  echo "rpki-client binary not executable: $RPKI_CLIENT_BIN" >&2
  exit 2
fi

mkdir -p "$OUT_DIR"
if [[ ! -x "$CIR_MATERIALIZE_BIN" || ! -x "$CIR_EXTRACT_INPUTS_BIN" || ! -x "$CCR_TO_COMPARE_VIEWS_BIN" ]]; then
  (
    cd "$ROOT_DIR"
    cargo build --release --bin cir_materialize --bin cir_extract_inputs --bin ccr_to_compare_views
  )
fi

TMP_ROOT="$OUT_DIR/.tmp"
TALS_DIR="$TMP_ROOT/tals"
META_JSON="$TMP_ROOT/meta.json"
MIRROR_ROOT="$TMP_ROOT/mirror"
CACHE_DIR="$TMP_ROOT/cache"
OUT_CCR_DIR="$TMP_ROOT/out"
RUN_LOG="$OUT_DIR/rpki-client.log"
ACTUAL_VRPS="$OUT_DIR/actual-vrps.csv"
ACTUAL_VAPS="$OUT_DIR/actual-vaps.csv"
ACTUAL_VAPS_META="$OUT_DIR/actual-vaps-meta.json"
ACTUAL_VRPS_META="$OUT_DIR/actual-vrps-meta.json"
REF_VRPS="$OUT_DIR/reference-vrps.csv"
REF_VAPS="$OUT_DIR/reference-vaps.csv"
SUMMARY_JSON="$OUT_DIR/compare-summary.json"

rm -rf "$TMP_ROOT"
mkdir -p "$TMP_ROOT"

"$CIR_EXTRACT_INPUTS_BIN" --cir "$CIR" --tals-dir "$TALS_DIR" --meta-json "$META_JSON"
python3 - <<'PY' "$TALS_DIR"
from pathlib import Path
import sys
for tal in Path(sys.argv[1]).glob("*.tal"):
    lines = tal.read_text(encoding="utf-8").splitlines()
    rsync_uris = [line for line in lines if line.startswith("rsync://")]
    base64_lines = []
    seen_sep = False
    for line in lines:
        if seen_sep:
            if line.strip():
                base64_lines.append(line)
        elif line.strip() == "":
            seen_sep = True
    tal.write_text("\n".join(rsync_uris) + "\n\n" + "\n".join(base64_lines) + "\n", encoding="utf-8")
PY
materialize_cmd=("$CIR_MATERIALIZE_BIN" --cir "$CIR" --repo-bytes-db "$REPO_BYTES_DB" --mirror-root "$MIRROR_ROOT")
if [[ "$KEEP_DB" -eq 1 ]]; then
  materialize_cmd+=(--keep-db)
fi
"${materialize_cmd[@]}"

VALIDATION_EPOCH="$(python3 - <<'PY' "$META_JSON"
from datetime import datetime, timezone
import json, sys
vt = json.load(open(sys.argv[1]))["validationTime"]
dt = datetime.fromisoformat(vt.replace("Z", "+00:00")).astimezone(timezone.utc)
print(int(dt.timestamp()))
PY
)"
mapfile -t TAL_PATHS < <(python3 - <<'PY' "$META_JSON"
import json, sys
for item in json.load(open(sys.argv[1], encoding="utf-8"))["talFiles"]:
    print(item["path"])
PY
)
CLIENT_TAL_ARGS=()
for tal_path in "${TAL_PATHS[@]}"; do
  CLIENT_TAL_ARGS+=(-t "$tal_path")
done
COMPARE_TRUST_ANCHOR="unknown"

export CIR_MIRROR_ROOT="$(python3 - <<'PY' "$MIRROR_ROOT"
from pathlib import Path
import sys
print(Path(sys.argv[1]).resolve())
PY
)"
export REAL_RSYNC_BIN="$REAL_RSYNC_BIN"
export CIR_LOCAL_LINK_MODE=1

mkdir -p "$CACHE_DIR" "$OUT_CCR_DIR"
chmod -R 0777 "$TMP_ROOT"
"$RPKI_CLIENT_BIN" \
  -R \
  -e "$WRAPPER" \
  -P "$VALIDATION_EPOCH" \
  "${CLIENT_TAL_ARGS[@]}" \
  -d "$CACHE_DIR" \
  "$OUT_CCR_DIR" >"$RUN_LOG" 2>&1

if [[ -f "$OUT_CCR_DIR/rpki.ccr" ]]; then
  "$CCR_TO_COMPARE_VIEWS_BIN" \
    --ccr "$OUT_CCR_DIR/rpki.ccr" \
    --vrps-out "$ACTUAL_VRPS" \
    --vaps-out "$ACTUAL_VAPS" \
    --trust-anchor "$COMPARE_TRUST_ANCHOR"
else
  python3 - <<'PY' "$OUT_CCR_DIR/json" "$ACTUAL_VRPS" "$ACTUAL_VAPS" "$COMPARE_TRUST_ANCHOR"
import csv
import json
import sys
from pathlib import Path

json_path = Path(sys.argv[1])
vrps_out = Path(sys.argv[2])
vaps_out = Path(sys.argv[3])
compare_ta = sys.argv[4]

if not json_path.is_file():
    raise SystemExit(f"rpki-client output has neither rpki.ccr nor json: {json_path}")

data = json.loads(json_path.read_text(encoding="utf-8"))

vrps_out.parent.mkdir(parents=True, exist_ok=True)
with vrps_out.open("w", newline="", encoding="utf-8") as fh:
    writer = csv.writer(fh)
    writer.writerow(["ASN", "IP Prefix", "Max Length", "Trust Anchor"])
    for roa in data.get("roas", []):
        writer.writerow([
            f"AS{roa['asn']}",
            roa["prefix"],
            str(roa["maxLength"]),
            compare_ta,
        ])

with vaps_out.open("w", newline="", encoding="utf-8") as fh:
    writer = csv.writer(fh)
    writer.writerow(["Customer ASN", "Providers", "Trust Anchor"])
    for aspa in data.get("aspas", []):
        providers = ";".join(f"AS{item}" for item in sorted(aspa.get("providers", [])))
        writer.writerow([
            f"AS{aspa['customer_asid']}",
            providers,
            compare_ta,
        ])
PY
fi

python3 - <<'PY' "$ACTUAL_VRPS" "$ACTUAL_VAPS" "$ACTUAL_VRPS_META" "$ACTUAL_VAPS_META"
import csv, json, sys
def count_rows(path):
    with open(path, newline="") as f:
        rows = list(csv.reader(f))
    return max(len(rows) - 1, 0)
json.dump({"count": count_rows(sys.argv[1])}, open(sys.argv[3], "w"), indent=2)
json.dump({"count": count_rows(sys.argv[2])}, open(sys.argv[4], "w"), indent=2)
PY

"$CCR_TO_COMPARE_VIEWS_BIN" --ccr "$REFERENCE_CCR" --vrps-out "$REF_VRPS" --vaps-out "$REF_VAPS" --trust-anchor "$COMPARE_TRUST_ANCHOR"

python3 - <<'PY' "$ACTUAL_VRPS" "$REF_VRPS" "$ACTUAL_VAPS" "$REF_VAPS" "$SUMMARY_JSON" "$META_JSON"
import csv, json, sys
def rows(path):
    with open(path, newline="") as f:
        return list(csv.reader(f))[1:]
actual_vrps = {tuple(r) for r in rows(sys.argv[1])}
ref_vrps = {tuple(r) for r in rows(sys.argv[2])}
actual_vaps = {tuple(r) for r in rows(sys.argv[3])}
ref_vaps = {tuple(r) for r in rows(sys.argv[4])}
meta = json.load(open(sys.argv[6], encoding="utf-8"))
summary = {
  "compareMode": "trust-anchor-agnostic",
  "talCount": len(meta["talFiles"]),
  "talPaths": [item["path"] for item in meta["talFiles"]],
  "vrps": {
    "actual": len(actual_vrps),
    "reference": len(ref_vrps),
    "match": actual_vrps == ref_vrps,
    "only_in_actual": sorted(actual_vrps - ref_vrps)[:20],
    "only_in_reference": sorted(ref_vrps - actual_vrps)[:20],
  },
  "vaps": {
    "actual": len(actual_vaps),
    "reference": len(ref_vaps),
    "match": actual_vaps == ref_vaps,
    "only_in_actual": sorted(actual_vaps - ref_vaps)[:20],
    "only_in_reference": sorted(ref_vaps - actual_vaps)[:20],
  }
}
with open(sys.argv[5], "w") as f:
    json.dump(summary, f, indent=2)
PY

if [[ "$KEEP_DB" -ne 1 ]]; then
  rm -rf "$TMP_ROOT"
fi

echo "done: $OUT_DIR"