Update jtokkit library from 0.6.1 to 1.1.0 to support text-embedding-3-small model

Description

The jtokkit library (used for tokenizing text before sending to OpenAI's embedding API) is currently pinned at version 0.6.1 in bom/application/pom.xml. This version does not include support for the text-embedding-3-small and text-embedding-3-large model family (released January 2024).

As a result, when dotCMS attempts to generate embeddings using text-embedding-3-small, the EncodingUtil class cannot find a matching encoding in the registry and logs the following warning:

INFO  util.EncodingUtil - Model [text-embedding-3-small] is not suitable for encoding, marking it as invalid and falling back to other models

The model is then marked as ModelStatus.INVALID and the system falls back to other configured models, or fails entirely if no valid fallback exists.

Version 1.1.0 of jtokkit adds support for the text-embedding-3-* model family. Upgrading resolves this issue.

Acceptance Criteria

• Update jtokkit version from 0.6.1 to 1.1.0 in bom/application/pom.xml
• Verify that EncodingUtil.getEncoding() correctly resolves an encoding for text-embedding-3-small after the upgrade
• Verify that no "is not suitable for encoding" warning is logged when text-embedding-3-small is the configured embedding model
• Confirm embedding generation works end-to-end with text-embedding-3-small
• Run existing AI-related integration/unit tests to confirm no regressions

Related

• Epic: #34058
• Regression Test Task: #34072

Start in bom/application/pom.xml and inspect EncodingUtil, then run the existing AI-related unit and integration tests. Update the jtokkit dependency to 1.1.0 and verify that text-embedding-3-small resolves an encoding, avoids the warning, and supports end-to-end embedding generation without regressions.