#!/usr/bin/env bash
# v3 vs v3.1 battle test, tests 7 and 9 (shell: curl + sha256sum). Usage: tests.sh 2026-10-29 18 [dir]
# Prints PASS or FAIL for each check and exits 1 if any check FAILs.
set -euo pipefail
day=$1 hh=$2 dir=${3:-./battle-$day-$hh}; h=${day}T$hh; dl=https://dl.pendulumflow.com
mkdir -p "$dir/v3" "$dir/v3.1"; cd "$dir"
fails=0
verdict() { if [ "$2" = ok ]; then echo "PASS $1"; else echo "FAIL $1"; fails=$((fails + 1)); fi; }

# 7. SHA256SUMS. Each archive's SHA256SUMS.txt lists every file of the hour (paths are relative to v3/ or v3.1/).
#    v3.1's list has 2 more lines than the manifest's own list: the manifest itself and the verify report,
#    which the manifest cannot list (a file cannot hold its own checksum). In a typical hour that is 38 and 36.
#    Expected: every listed file downloads and checks OK, in all three lists.
curl -fsS --retry 5 --retry-delay 30 "$dl/v3/SHA256SUMS.txt"   | grep "  $day/$hh/" > v3/SUMS   || true
curl -fsS --retry 5 --retry-delay 30 "$dl/v3.1/SHA256SUMS.txt" | grep "  $day/$hh/" > v3.1/SUMS || true
check() {  # check <label> <dir> <sums file>: every line present and OK
  local label=$1 d=$2 sums=$3 n ok
  n=$(grep -c . "$sums" || true)
  ok=$( (cd "$d" && sha256sum -c "$sums" 2>/dev/null || true) | grep -c ': OK$' || true)
  echo "$label: $ok of $n OK"
  if [ "$n" -gt 0 ] && [ "$ok" = "$n" ]; then verdict "7 $label" ok; else verdict "7 $label" bad; fi
}
for a in v3 v3.1; do
  awk '{print $2}' "$a/SUMS" | while read -r p; do
    mkdir -p "$a/$(dirname "$p")"; curl -fsS --retry 5 --retry-delay 30 -o "$a/$p" "$dl/$a/$p" || echo "missing: $a/$p"
  done
  check "$a SHA256SUMS" "$a" "$PWD/$a/SUMS"
done
#    and the manifest's own list (files[].sha256):
if [ -s "v3.1/$day/$hh/$h.manifest.json" ]; then
  python3 -c "import json,sys; sys.stdout.reconfigure(newline=chr(10)); m=json.load(open(sys.argv[1])); [print(f['sha256']+'  '+f['file']) for f in m['files']]" \
    "v3.1/$day/$hh/$h.manifest.json" > v3.1/MANIFEST_SUMS
  check "v3.1 manifest files[]" "v3.1/$day/$hh" "$PWD/v3.1/MANIFEST_SUMS"
else
  verdict "7 v3.1 manifest files[] (no manifest)" bad
fi

# 9. Range reads and the cache. The same 64 KiB range twice. Expected: 206 every time; v3.1 cf-cache-status
#    HIT on the second try (a MISS is allowed on the first only); v3 DYNAMIC (v3 is not served from the cache).
#    The v3.1 file is the manifest's first lean file, so the test never guesses a file name.
lean1=$(python3 -c "import json,sys; m=json.load(open(sys.argv[1])); print(sorted(f['file'] for f in m['files'] if f['role']=='lean')[0])" \
  "v3.1/$day/$hh/$h.manifest.json" 2>/dev/null | tr -d '\r' || true)
[ -n "$lean1" ] || { verdict "9 (no lean file in the manifest)" bad; lean1=missing; }
for u in "$dl/v3/$day/$hh/$h.parquet" "$dl/v3.1/$day/$hh/$lean1"; do
  for i in 1 2; do
    out=$(curl -sS -r 0-65535 -o /dev/null -D - -w 'time_total=%{time_total}s code=%{http_code}\n' "$u" | tr -d '\r' || true)
    code=$(printf '%s\n' "$out" | sed -n 's/.*code=\([0-9]*\).*/\1/p' | tail -1)
    cache=$(printf '%s\n' "$out" | sed -n 's/^cf-cache-status: *//Ip' | tail -1)
    echo "${u#$dl/} try $i: code=$code cf-cache-status=${cache:-none} $(printf '%s\n' "$out" | grep -o 'time_total=[^ ]*')"
    want=ok
    [ "$code" = 206 ] || want=bad
    case "$u" in
      */v3.1/*) [ "$i" = 1 ] && [ "$cache" = MISS ] || [ "$cache" = HIT ] || want=bad ;;
      *)        [ "$cache" = DYNAMIC ] || want=bad ;;
    esac
    verdict "9 ${u#$dl/} try $i" "$want"
  done
done

echo "== $fails check(s) FAILED"
[ "$fails" = 0 ]