fix(evidence): close S3 and smoke safety gaps

scripts/preprocess-smoke.sh

@@ -4,10 +4,15 @@ cd "$(dirname "$0")/.."
 
 tmp=$(mktemp -d "${TMPDIR:-/tmp}/thoth-preprocess.XXXXXX")
 project="thoth-preprocess-$$"
+compose=""
 cleanup() {
-  docker compose -f compose.yaml -f deploy/compose.local-vector.yaml \
-    -f deploy/compose.preprocess.yaml -f "$tmp/smoke.yaml" --project-name "$project" \
-    --profile local-vector --profile preprocess down --volumes >/dev/null 2>&1 || true
+  trap - EXIT HUP INT TERM
+  if [ -n "$compose" ]; then
+    $compose down --volumes >/dev/null
+  fi
+  test -z "$(docker ps -aq --filter "label=com.docker.compose.project=$project")"
+  test -z "$(docker volume ls -q --filter "label=com.docker.compose.project=$project")"
+  test -z "$(docker network ls -q --filter "label=com.docker.compose.project=$project")"
   rm -rf "$tmp"
 }
 trap cleanup EXIT HUP INT TERM
@@ -56,18 +61,36 @@ $compose build preprocess-evidence
 $compose up -d vector-db mock-embeddings
 $compose run --rm vector-reconcile >/dev/null
 $compose run --rm --no-deps vector-migrate >/dev/null
+if [ "${1:-}" = "--cleanup-failure" ]; then
+  cleanup
+  test ! -e "$tmp"
+  echo "injected preprocessing failure cleanup passed."
+  exit 0
+fi
+generation_count() {
+  $compose run --rm --no-deps --entrypoint sh preprocess-evidence -c \
+    'find /data/workspaces/preprocess-evidence/corpus/generations -mindepth 1 -maxdepth 1 -type d 2>/dev/null | wc -l'
+}
+before=$(generation_count)
 first=$($compose run --rm preprocess-evidence)
+after_first=$(generation_count)
 second=$($compose run --rm preprocess-evidence)
+after_second=$(generation_count)
 printf '%s' 'generation two' >>"$tmp/source/evidence/a.md"
 third=$($compose run --rm preprocess-evidence)
+after_third=$(generation_count)
 dwh=$($compose run --rm preprocess-dwh)
-python3 - "$first" "$second" "$third" <<'PY'
+python3 - "$first" "$second" "$third" "$before" "$after_first" "$after_second" "$after_third" <<'PY'
 import json, sys
-a, b, c = map(json.loads, sys.argv[1:])
+a, b, c = map(json.loads, sys.argv[1:4])
+before, first_count, second_count, third_count = map(int, sys.argv[4:])
 assert len(a["changed"]) == 1 and not a["unchanged"]
 assert len(b["unchanged"]) == 1 and not b["changed"]
 assert len(c["changed"]) == 1 and c["generation"] != a["generation"]
 assert all(row["published"] for row in (a, b, c))
+assert first_count == before + 1
+assert second_count == first_count
+assert third_count == second_count + 1
 PY
 python3 - "$dwh" <<'PY'
 import json, sys
@@ -80,3 +103,5 @@ import json, sys
 assert json.loads(sys.argv[1])["generation"] == sys.argv[2].strip()
 PY
 echo "real Compose preprocessing unchanged rerun, mutation, DWH job, and ACTIVE publish passed."
+cleanup
+test ! -e "$tmp"