hydradb-cli 0.2.0

hydradb-cli 0.2.0

hydradb-cli 0.1.1

The CLI now talks to the HydraDB v2 API and adopts a consistent vocabulary across every
command. Everything that worked in 0.1.0 still works: each renamed command and flag is
kept as an alias that prints a one-line deprecation warning naming its replacement, so
existing scripts keep running while you migrate.

Warnings always go to stderr, so --output json pipelines are unaffected.

Added

• New command names, one verb per action:
  query, ingest, list, inspect, delete, relations, verify, doctor, and
  database {create,delete,list,collections,stats,readiness,monitor}.
• --database / -d and --collection to scope any command, replacing
  --tenant-id / --sub-tenant-id.
• Environment variables HYDRADB_API_KEY, HYDRADB_DATABASE, HYDRADB_COLLECTION
  and HYDRADB_BASE_URL.
• Config keys database and collection, e.g. hydradb config set database my-db.
• hydradb database stats and hydradb database readiness as separate commands.
  Previously this data was only available merged together via monitor, which is still
  there as the combined view.
• hydradb doctor, which reports your resolved configuration and whether the API is
  reachable. It replaces whoami, which only ever showed local config.
• A dependency on the official hydradb-sdk package, which now handles all API calls.

Changed

• login saves the database and collection config keys, and its JSON output gains a
  database field. The existing tenant_id field is unchanged.
• login --output json and every other JSON shape keep their documented jq paths, with
  the two exceptions listed under Breaking.
• The "no database specified" error now points at --database and
  hydradb config set database.

Deprecated

Each of these still works and warns once, naming its replacement:

Deprecated	Use instead
memories add, knowledge upload, knowledge upload-text	ingest
memories list, fetch sources	list
memories delete, knowledge delete	delete
knowledge verify	verify
recall full, recall preferences, recall keyword	query
fetch content	inspect
fetch relations	relations
tenant …	database …
whoami	doctor
--tenant-id, --sub-tenant-id	--database, --collection
HYDRA_DB_* and HYDRADB_API_URL env vars	HYDRADB_* (see Added)

Breaking (--output json)

Two commands change their JSON shape as a direct result of the v2 API migration. The
documented query … | jq '.chunks[0].chunk_content' path is unchanged.

• list (and its memories list / fetch sources aliases): items are now under
  sources rather than user_memories. Memories and knowledge appear in the same
  sources array.
• database monitor (and the tenant monitor alias): returns a merged
  { "database", "stats", "readiness" } object. The individual pieces are now available
  from database stats and database readiness.

Fixed

• Ingesting no longer displays Source ID: unknown when the server returns the identifier
  as id.
• ingest --kind knowledge --text … --source-id foo now preserves the ID you supply, so a
  later delete foo matches. Previously the server minted its own ID and the delete
  silently failed.
• delete of an ID that does not exist no longer reports success — it exits non-zero and
  reports {"success": false, …} in JSON mode.
• Ingesting multiple files at once no longer drops files; each is uploaded and the results
  are merged.