feat(preprocess): add deployment jobs and S3 source

scripts/preprocess-smoke.sh

@@ -0,0 +1,15 @@
+#!/bin/sh
+set -eu
+cd "$(dirname "$0")/.."
+
+rendered=$(docker compose -f compose.yaml -f deploy/compose.preprocess.yaml --profile preprocess config)
+printf '%s' "$rendered" | grep -q 'preprocess-evidence:'
+printf '%s' "$rendered" | grep -q 'preprocess-dwh:'
+if printf '%s' "$rendered" | grep -qi 'access-secret\|secret-key'; then
+  echo "preprocess Compose rendered secret material" >&2
+  exit 1
+fi
+(cd harness && .venv/bin/pytest -q \
+  tests/test_corpus_pipeline.py::test_unchanged_documents_skip_acquire_normalize_chunk_and_embed \
+  tests/test_corpus_pipeline.py::test_retention_bounds_generations_and_purges_vectors_after_publish)
+echo "preprocess unchanged rerun and modified-generation smoke passed."