2 Commits

Author	SHA1	Message	Date
Harry	65e89520c0	refactor: unify download item types and eliminate extension-based branching ... Merge AssetDownloadItem, AssetInlineItem into SandboxDownloadItem with optional 'content' field. All consumers now follow a clean pipeline: get items → accessor.resolve_items() → AppAssetService.to_download_items() → download Key changes: - SandboxDownloadItem gains content: bytes | None (entities.py) - ZipSandbox.download_items() handles both inline (base64 heredoc) and remote (curl) via a single pipeline — no structural branching - AssetDownloadService.build_download_script() takes unified list - CachedContentAccessor.resolve_items() batch-enriches items from DB (extension-agnostic, no 'if md' checks needed) - AppAssetService.to_download_items() converts AssetItem → SandboxDownloadItem - DraftAppAssetsInitializer, package_and_upload, export_bundle simplified - file_upload/node.py switched to SandboxDownloadItem - Deleted AssetDownloadItem and AssetInlineItem classes	2026-03-10 17:12:15 +08:00
Harry	a8074f4f4a	feat: add DB inline content cache for app asset draft files ... Introduce app_asset_contents table as a read-through cache over S3 for text-like asset files (e.g. .md skill documents). This eliminates N individual S3 fetches during SkillBuilder builds — bulk_load pulls all content in a single SQL query with S3 fallback on miss. Key components: - CachedContentAccessor: DB-first read / dual-write / S3 fallback - AssetContentService: static DB operations (get, get_many, upsert, delete) - should_mirror(): single source of truth for extension-based policy - Alembic migration for app_asset_contents table Modified callers: - SkillBuilder uses accessor.bulk_load() instead of per-node S3 reads - AppAssetService.get/update_file_content route through accessor - delete_node cleans both DB cache and S3 - draft_app_assets_initializer uses should_mirror() instead of hardcoded .md	2026-03-10 17:12:14 +08:00