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

# summarize.sh <samples.csv>
# Produces a quick text summary + per-endpoint aggregates.

csv="${1:-}"
if [[ -z "$csv" || ! -f "$csv" ]]; then
  echo "Usage: summarize.sh <path/to/samples.csv>" >&2
  exit 2
fi

# CSV columns (see monitor.sh):
# 1 ts_utc
# 2 tick
# 3 endpoint_idx
# 4 url (JSON string)
# 5 http_status
# 6 total_time_ms

awk -F',' '
  NR==1 {next}
  {
    idx=$3;
    status=$5;
    ms=$6;
    # url is JSON string (quoted); keep raw
    url=$4;

    count[idx]++;
    sum_ms[idx]+=ms;
    if (!(idx in min_ms) || ms < min_ms[idx]) min_ms[idx]=ms;
    if (!(idx in max_ms) || ms > max_ms[idx]) max_ms[idx]=ms;

    url_by_idx[idx]=url;

    status_count[idx, status]++;
    total++;
    if (status ~ /^2/) ok++;
    else if (status ~ /^3/) redir++;
    else if (status ~ /^4/) c4++;
    else if (status ~ /^5/) c5++;
    else other++;
  }
  END {
    print "Samples:", total;
    print "2xx:", ok, "3xx:", redir, "4xx:", c4, "5xx:", c5, "other:", other;
    print "";

    # Per endpoint
    print "Per-endpoint aggregates:";
    for (idx in count) {
      avg = (count[idx] ? sum_ms[idx]/count[idx] : 0);
      printf("- idx=%s url=%s\n", idx, url_by_idx[idx]);
      printf("  n=%d avg_ms=%.1f min_ms=%d max_ms=%d\n", count[idx], avg, min_ms[idx], max_ms[idx]);

      # print top statuses seen
      printf("  statuses:");
      first=1;
      for (k in status_count) {
        split(k, parts, SUBSEP);
        if (parts[1]==idx) {
          if (!first) printf(",");
          printf(" %s=%d", parts[2], status_count[k]);
          first=0;
        }
      }
      printf("\n");
    }
  }
' "$csv"