Release v0.3.0

• If interrupted, always cancel the extract job (if one is in progress) and clean up any extracted files in the bucket
• Extract selected properties only - this will significantly improve overall performance if only a couple of properties are required
• Reduce minimum required Python version to 3.8 (previously set to 3.9 for no good reason)
• Additionally support google_application_credentials config as an object
• Misleading setting filter_datasets renamed to filter_tables
• Support shell pattern matching for table names with filter_tables

Release v0.2.0

• Initial implementation of bulk load using Google Storage bucket. Currently always emits batch messages.

Release v0.1.0

• Initial release with latest versions of singer-sdk and python 3.9
• Added support for Google Application Credentials by json and file path
• Improved support for structured json
• Implemented dataset filtering to resolve issues when there's a large number of datasets in a schema
• Added support for REPEATED types
• Fix 'ValueError: Out of range float values are not JSON compliant' by dropping these unsupported values