feat(metadata): add CivitAI AutoV3 hash support across all storage layers

- Three-state autov3 field (not-checked / checked-unavailable / 12-hex value)
  in .metadata.json sidecars, in-memory ModelHashIndex, and SQLite
  (models.autov3 column + autov3_index table) with column-presence migration
- Background self-terminating backfill for legacy rows: per-model-type
  concurrency guard, executor-offloaded I/O, Civitai-first resolution
  (SHA256-matched version file) falling back to the embedded safetensors
  header hash
- Civitai-first propagation on metadata refresh, scan, and download paths;
  reject the empty-string SHA256 placeholder and strip OneTrainer 0x prefix
- List API hash filters and hash index lookups accept 12-char AutoV3
- Cap safetensors header reads at 64 MiB to prevent crafted-file allocation
- Prevent stale AutoV3 mappings on file replacement while preserving them on
  same-file re-registration (lazy-hash completion)

py/utils/constants.py

@@ -80,6 +80,17 @@ CIVITAI_USER_MODEL_TYPES = [
 # Default chunk size in megabytes used for hashing large files.
 DEFAULT_HASH_CHUNK_SIZE_MB = 4
 
+# Upper bound for a safetensors header block (bytes). Real headers are at most
+# a few MB (tensor name/shape lists); the cap prevents a crafted file with an
+# absurd 64-bit header length from forcing a multi-GB allocation during scan.
+MAX_SAFETENSORS_HEADER_BYTES = 64 * 1024 * 1024
+
+# First 12 chars of the SHA256 of an empty byte string. Some (re-packaging)
+# training tools write this placeholder into safetensors metadata instead of a
+# real hash; it must never be treated as a valid AutoV3 — several broken
+# models sharing it would collide in the hash index and falsely match recipes.
+INVALID_AUTOV3_EMPTY_HASH = "e3b0c44298fc"
+
 # Auto-organize settings
 AUTO_ORGANIZE_BATCH_SIZE = (
     50  # Process models in batches to avoid overwhelming the system