vocabulary:
  - term: CDP
    label: Customer Data Platform
    definition: >
      A software system that creates a persistent, unified customer database
      accessible to other systems. Treasure Data is an enterprise CDP enabling
      data collection, unification, and activation.

  - term: database
    label: Database
    definition: >
      A logical container in Treasure Data that holds tables of data.
      Analogous to a schema in relational databases.

  - term: table
    label: Table
    definition: >
      A collection of records within a Treasure Data database. Tables store
      time-series event data and can be queried using Presto or Hive.

  - term: job
    label: Job
    definition: >
      An asynchronous unit of work in Treasure Data, typically a query
      execution (Presto or Hive) or bulk import operation. Jobs are identified
      by a unique job_id.

  - term: bulk_import
    label: Bulk Import
    definition: >
      A session-based mechanism for uploading large volumes of data to Treasure
      Data in msgpack.gz format. Sessions progress through uploading, frozen,
      and committed states.

  - term: bulk_load
    label: Bulk Load
    definition: >
      A scheduled or on-demand data ingestion session that pulls data from
      external sources (databases, cloud storage, SaaS apps) into Treasure Data.

  - term: presto
    label: Presto Query Engine
    definition: >
      A distributed SQL query engine used in Treasure Data for fast, interactive
      analytics on large datasets. Supports standard ANSI SQL syntax.

  - term: hive
    label: Hive Query Engine
    definition: >
      An Apache Hive-compatible query engine in Treasure Data for batch
      processing of large datasets using HiveQL.

  - term: connector
    label: Output Connector
    definition: >
      An integration endpoint that allows Treasure Data job results to be
      written to external destinations such as data warehouses, marketing
      platforms, or storage services.

  - term: workflow
    label: Workflow
    definition: >
      An orchestrated sequence of tasks defined using the Digdag workflow engine
      for automating data pipelines and transformations in Treasure Data.

  - term: digdag
    label: Digdag
    definition: >
      An open-source workflow orchestration engine used by Treasure Data for
      defining and executing complex multi-step data pipelines.

  - term: api_key
    label: API Key
    definition: >
      A secret credential used to authenticate requests to the Treasure Data
      API. Passed in the Authorization header as "TD1 <api-key>".

  - term: account_owner
    label: Account Owner
    definition: >
      A user with full administrative privileges over a Treasure Data account,
      including the ability to manage users, SSO settings, and billing.

  - term: sso
    label: Single Sign-On
    definition: >
      Identity federation capability allowing users to authenticate to Treasure
      Data using an external Identity Provider (IdP) via SAML.

  - term: msgpack
    label: MessagePack
    definition: >
      A binary serialization format used for efficient data transfer in Treasure
      Data bulk import operations. Files are typically gzip-compressed (.msgpack.gz).

  - term: postback_api
    label: Postback API
    definition: >
      A JSON-based HTTP event ingestion API for submitting individual records to
      Treasure Data from systems that cannot use the JavaScript SDK.

  - term: td_audit_log
    label: TD Audit Log
    definition: >
      A built-in Treasure Data database that records all account activity,
      including queries, user actions, and API calls for compliance and
      monitoring purposes.