hotdata-framework 0.6.0

Added

• HotdataClient.add_managed_table(database, table, *, schema) declares a new table on an existing managed database (wrapping the SDK add_database_table endpoint). This allows additive schema evolution without recreating the database.

hotdata-framework 0.5.0

Changed

• Adopt the hotdata 0.5.0 SDK surface (dependency bumped from >=0.4.1 to >=0.5.0). The release is backward compatible for everything the framework uses; the only API changes are additive (a new optional format field on LoadManagedTableRequest and an optional format parameter on ResultsApi.get_result), so no framework code changes were required.

hotdata-framework 0.4.1

Fixed

• ManagedDatabaseClient.fetch_table now waits for the persisted result to reach ready before fetching it as Arrow on the synchronous query path (it previously only waited on the async path). This fixes failures on read-modify-write loads (merge/append) and state reads against the live backend, where the result is often still processing when the inline preview returns.

hotdata-framework 0.4.0

Changed

• Renamed the distribution from hotdata-runtime to hotdata-framework and the import package from hotdata_runtime to hotdata_framework. Consumers should depend on hotdata-framework and use import hotdata_framework. The GitHub repository is now sdk-python-framework.
• Added PyPI classifiers, keywords, and an updated description identifying the project as a Python framework.

hotdata-runtime 0.3.0

Added

• Adopt the hotdata 0.4.1 SDK surface.
• New typed error-handling public API: HotdataError, HotdataTerminalError, HotdataTransientError, and classify_sdk_error (hotdata_runtime/errors.py).
• ManagedDatabaseClient for managed database operations (hotdata_runtime/managed_client.py).
• py.typed marker so downstream consumers pick up inline type information.

Changed

• Bump the hotdata dependency pin to >=0.4.1.
• Add ruff and mypy tooling configuration and dev dependencies (ruff>=0.5, mypy>=1.5); apply ruff lint/format cleanup across the package.

hotdata-runtime 0.2.4

Changed

• Release 0.2.4

hotdata-runtime 0.2.3

Changed

• Release 0.2.3

hotdata-runtime 0.2.2

Changed

• Release 0.2.2

hotdata-runtime 0.2.1

Added

• execute_sql accepts an optional database keyword argument. When provided, the database name is resolved to an ID and sent as the X-Database-Id header so SQL can reference managed database tables as "default"."<schema>"."<table>". Behaviour is unchanged when database is omitted.

hotdata-runtime 0.2.0

Changed

• Switch managed database operations from the connections API to the dedicated /databases API (hotdata>=0.2.3 required).
• create_managed_database first parameter renamed from name to description (keyword-only).
• ManagedDatabase dataclass: replace name/source_type fields with description/default_connection_id.
• resolve_managed_database tries direct ID lookup first, then falls back to a description scan.
• list_managed_databases now fetches all databases regardless of source type.
• list_managed_tables, load_managed_table, and delete_managed_table use default_connection_id instead of database id for connection-scoped operations.

Added

• create_managed_database accepts an optional expires_at parameter.

Removed

• MANAGED_SOURCE_TYPE, build_managed_config, and create_connection_request removed from the public API.