#!/bin/bash
# Smoke test for hayabusa: fetch sample EVTX, run the container, check that
# all expected output artefacts landed and contain detections, print stats.
#
# Defaults to the DeepBlueCLI subset (21 files, ~30s).
#
# Available SUBSET values (under test-data/sample-evtx/):
#   DeepBlueCLI           21 files, ~30s   — well-known PowerShell + auth attacks
#   YamatoSecurity        16 files, ~20s   — Yamato Security's own samples
#   EVTX-ATTACK-SAMPLES  278 files, ~5min  — sbousseaden's MITRE-mapped corpus
#   EVTX-to-MITRE-Attack 284 files, ~5min  — mdecrevoisier's MITRE-mapped corpus
#   ""                   599 files, ~10min — full bundle (set SUBSET="" or empty)
#
# Env vars:
#   TAG=ls-hayabusa:test       which image to test
#   SUBSET=DeepBlueCLI         which subdir of sample-evtx to scan (or "" for all)
#   KEEP_DATA=1                don't delete test-data/sample-evtx after run
set -u
TAG="${TAG:-ls-hayabusa:test}"
SUBSET="${SUBSET:-DeepBlueCLI}"
KEEP_DATA="${KEEP_DATA:-0}"

cd "$(dirname "$0")"
ROOT="$(pwd)"
DATA="$ROOT/test-data/sample-evtx"
OUT="$(mktemp -d)"
trap 'rm -rf "$OUT"; [ "$KEEP_DATA" = 0 ] && rm -rf "$ROOT/test-data"' EXIT

pass=0; fail=0
ok()  { echo "PASS  $1"; pass=$((pass+1)); }
bad() { echo "FAIL  $1"; fail=$((fail+1)); }

# 1. Image present?
if docker image inspect "$TAG" >/dev/null 2>&1; then
  ok "image $TAG present"
else
  bad "image $TAG not present (build it first)"; exit 1
fi

# 2. Sample EVTX available (clone on demand).
if [ ! -d "$DATA" ]; then
  echo "Fetching sample EVTX (Yamato bundle, ~250 MB)..."
  ./fetch-test-data.sh >/dev/null
fi
SCAN_PATH="$DATA/$SUBSET"
[ -z "$SUBSET" ] && SCAN_PATH="$DATA"
n_evtx=$(find "$SCAN_PATH" -name "*.evtx" | wc -l | tr -d ' ')
if [ "$n_evtx" -gt 0 ]; then
  ok "found $n_evtx EVTX files in ${SUBSET:-<full bundle>}"
else
  bad "no EVTX in $SCAN_PATH"; exit 1
fi

# 3. Run scan.
echo "Running scan against $SUBSET → $OUT ..."
if docker run --rm --network=none \
     -v "$SCAN_PATH:/data:ro" \
     -v "$OUT:/output" \
     "$TAG" >"$OUT/.run.log" 2>&1; then
  ok "container exited cleanly"
else
  bad "container exited non-zero"
  tail -20 "$OUT/.run.log"
fi

# 4. Expected artefacts.
html=$(ls "$OUT"/hayabusa_*.html 2>/dev/null | head -1)
csv=$(ls "$OUT"/hayabusa_*.ts.csv 2>/dev/null | head -1)
jsonl=$(ls "$OUT"/hayabusa_*_takajo.jsonl 2>/dev/null | head -1)
takajo="$OUT/takajo"

[ -s "$html"  ] && ok "HTML report exists ($(du -h "$html" | cut -f1))"  || bad "HTML report missing/empty"
[ -s "$csv"   ] && ok "Timesketch CSV exists ($(du -h "$csv" | cut -f1))" || bad "CSV missing/empty"
[ -s "$jsonl" ] && ok "Takajo JSONL exists ($(du -h "$jsonl" | cut -f1))" || bad "JSONL missing/empty"
[ -d "$takajo" ] && [ -n "$(ls "$takajo" 2>/dev/null)" ] && ok "takajo/ directory has $(ls "$takajo" | wc -l | tr -d ' ') files" || bad "takajo/ missing"

# 5. Detection count from CSV (skip header row).
if [ -s "$csv" ]; then
  detections=$(($(wc -l < "$csv") - 1))
  echo
  echo "Detections in timeline: $detections"
fi

# 6. TTP coverage.
if [ -s "$takajo/MitreTTP-Heatmap.json" ]; then
  ttp_count=$(grep -oE '"T[0-9]+' "$takajo/MitreTTP-Heatmap.json" | sort -u | wc -l | tr -d ' ')
  echo "Distinct MITRE TTPs surfaced: $ttp_count"
fi

echo
echo "Summary: $pass pass, $fail fail"
[ "$fail" -eq 0 ]