eddytor-object-store

name	eddytor-object-store
description	Browses and manages raw files in a registered Eddytor storage configuration — list, upload, download, delete, organise into folders, and move objects between configs. Activates for listing files, uploading/downloading objects, folders, moving data between buckets, or seeding a demo table — even without "object store."
license	CC-BY-NC-4.0
metadata	{"author":"eddytor","version":"1.0"}

Object Store

These tools operate on the raw objects in a storage configuration (the bucket/container behind a config_id), independent of Delta tables. Get the config_id from list_storage_configs (see eddytor-storage-registration).

Default procedure

1. list_storage_configs → pick a config_id.
2. list_objects(config_id, delimiter=true) → browse folders and files.
3. Upload/download/delete/move as needed with the tools below.

Tool usage examples

# Browse: delimiter=true gives directory-style listing (folders in common_prefixes)
list_objects(config_id, path="master-data", delimiter=true, extensions="csv,parquet", limit=100)
# paginate with the returned continuation_token

# Download (returns { path, filename, size_bytes, content_base64 }); max 8 MiB
download_object(config_id, path="master-data/products.csv")

# Upload from base64; path = destination directory, filename = object name. Max 100 MiB
upload_object(config_id, path="master-data", filename="products.csv", content_base64="<...>")

# Organise
create_folder(config_id, path="master-data/archive")
delete_object(config_id, path="master-data/old.csv")

# Move every object under a prefix (optionally to a different config)
move_objects(source_config_id, source_path="staging/products",
  destination_config_id, destination_path="master-data/products")

# Seed a demo table to explore domains
create_demo_table(config_id)        # creates demo_products at the config root

Gotchas

• list_objects limit defaults to 100, max 1000; use continuation_token to page. delimiter=true returns folders separately in common_prefixes.
• download_object rejects objects over 8 MiB — use the REST download endpoint for larger files. upload_object caps decoded content at 100 MiB.
• upload_object filename must have no path separators; the destination is path/filename.
• delete_object targets a single object, not a prefix. To move/remove a whole prefix use move_objects.
• move_objects deregisters any Delta table at the source and re-runs discovery at the destination — paths shift, so re-check list_tables afterward. To relocate a table specifically (not raw objects), prefer move_table (see eddytor-table-lifecycle).
• Paths are validated: no .., no leading /, no //, no null bytes.

Guidelines

• Use extensions to filter noisy listings (e.g. "csv").
• For tabular data you intend to query, import it (import_csv) rather than leaving it as a raw object — raw files are read-only even when discovered.
• create_demo_table is the fastest way to get a constrained table to experiment with.