0.16.1

Features

• Extract user roles from a string property in user claims #496 (#497)

Fixes

• fix vulnerability issues

0.16.0

The version contains breaking changes in DIAL encryption settings. Please update the settings before use.

Features

• Add mapping for iconUrl when publish applications #468 (#473)
• Add resource type to user metadata of a blob #465 (#466)
• Allow interceptors to pass authZ header along with API key #478
• Rename encryption properties in DIAL settings #454 (#455)
• allow to pass extra data to upstreams endpoints (#470)
• move resources across buckets (#474)

Fixes

• 'URL bucket doesn't match' error appears if to create an unpublish request for some conversations #463 (#467)
• Core allows to upload files with invalid path #459 (#460)
• GCP Core deployment start to fail after some time #472 (#475)

Other

• Update README.md (#462)
• bump epam/ai-dial-ci from 1.6.0 to 1.9.0 (#457)
• bump peter-evans/slash-command-dispatch from 3.0.1 to 4.0.0 (#249)
• update README.md (#464)
• upgrade dependencies versions (#443)

0.15.2

Fixes

• GCP Core deployment start to fail after some time #472 (#475)

0.15.0

Features

• Add new feature flag allowResume #444 (#445)
• Implement secured API Keys #435 (#436)
• Support attachment auto-sharing via responses #404 (#416)
• enable cors (#433)
• implement advanced load balancing (#422)
• implement periodic heartbeats for notifications endpoint (#439)
• return createdAt/updatedAt for files (#418)
• return etag in resource notifications (#429)
• support if-none-match for files (#426)

Fixes

• Chat completion request with per request API key fails to model adapter #449
• Core returns chunks of streaming response in a wrong order #420 (#421)
• Fix assembled response in prompt logs #447 (#448)
• DIAL Interceptor should check responses (#434)
• Incorrect token calculation for temp-api-key based on JWT token #438 (#440)
• Token usage stats are failed to calculate when deployment has interceptors (#424)
• TokenUsage stats is not calculated properly for apps when Core have multiple instances #393 (#407)
• assemble streaming response with a 'delimiter' inside content (#453)
• fix NPE at collecting request data function (#415)
• handle blank content encoding (#412)
• log errors on file upload/download (#410)
• minor updates in the docs (#409)
• passing query params to interceptor (#430)
• return permissions in metadata response (#414)
• unit tests
• update links in README.md (#441)

Other

• remove linebreaks at multiline logs (#408)
• renaming
• set the feature allowResume to true by default

0.14.3

Fixes

• Chat completion request with per request API key fails to model adapter #449

0.14.2

Features

• Implement secured API Keys #435 (#436)

Fixes

• Token usage stats are failed to calculate when deployment has interceptors (#424)
• passing query params to interceptor (#430)

Other

• remove linebreaks at multiline logs (#408)

0.14.1

Fixes

• handle blank content encoding (#412)
• log errors on file upload/download (#410)
• return permissions in metadata response (#414)

0.9.2

Other

• improve logging (#384)
• remove linebreaks at multiline logs (#408)

0.14.0

Features

• Add support of interceptors for models, applications #379 (#392)
• implement deployment reference (#403)
• implement etag and file read-write caching (#389)
• subscription for resources (#394)
• supported collection of attached files for multi-modal embeddings request (#395)

Fixes

• custom application URL handling (#396)
• fix big file upload to s3 (#401)
• use the result of AccessControlBaseController#handle (#397)

0.13.0

Features

• Implement write access sharing (#377)
• support custom applications (#350)(#305)
• support feature endpoints for custom applications (#385)
• update gflog to 3.0.6 (#391)

Fixes

• publication rules validation (#390)
• use Redisson API which doesn't require Redis 7+ (#387)