Tonic Textual release information

Add support for a LOCATION_COMPLETE_ADDRESS entity type. By default, the handling type for this entity type is Off, meaning that detected values are ignored. It is not currently supported for PDFs and images.

Improved the predications for the named-entity recognition entity type ORGANIZATION.

The new environment variable SOLAR_MAX_CONCURRENT_WORKER_JOBS allows you to configure the number of jobs that each worker can process at the same time. By default, each worker can process 8 jobs simultaneously.

You can now use the Tonic Textual SDK to redact audio files. When you send a request to redact an audio file, Textual first transcribes the audio into text, and then redacts the transcription. The request includes the entity type handling.

Bug fixes and other internal updates.

Bug fixes and other internal updates.

Added the Password entity type to the named-entity recognition (NER) model.

Bug fixes and other internal updates.

Added the Password entity type to the named-entity recognition (NER) model.

When you select an AWS Region, the dropdown list now includes the ISO Regions.

Additional consistency between Textual and Structural - A new environment variable allows you to configure a statistics seed to enable you to ensure that synthesized values in Textual match corresponding generated values in Structural. The statistics seed is also displayed in the Textual application. To ensure consistency between the application, provide the same statistics seed value in both.

The consistency applies to non-HIPAA addresses, to match the Address generator in Structural, and US telephone numbers, to match the US Phone Number generator. For these types of values, when the statistics seeds match, the same input value produces the same output value in both Textual and Structural.

Fixed an issue where quotes and whitespaces broke the LLM_MODEL setting for LLM synthesis.

Added a new entity type, NUMERIC_PII, for numeric values that act as identifiers.