Sourced from org.springframework.ai:spring-ai-vector-store's releases.

Spring AI 1.1.5 Release Notes

🎯 Highlights

This release includes 9 bug fixes, 3 documentation improvements, 11 other improvements.

⚠️ Upgrading Notes

• The Pixtral 12B model has been removed and Pixtral Large is deprecated. Update your model configuration to use the currently recommended Pixtral models to avoid issues in future releases. 447d2a4

📢 Noteworthy

• The Pixtral 12B model has been removed and the Pixtral Large model is now deprecated. Integration tests have been updated to use the recommended replacement models. Users relying on these models should migrate to the recommended alternatives. 447d2a4

🪲 Bug Fixes

• Fixed the CosmosDB vector store's doDelete method to properly parameterize queries, preventing potential SQL injection vulnerabilities and improving correctness. 6039e57
• Fixed an issue where conversationId was not correctly applied in the VectorStoreChatMemoryAdvisor filter, which could cause incorrect memory retrieval across conversations. 3cccfdf
• Corrected key handling in the vector store filter expression converter to ensure filter expressions are properly evaluated. 01386e2
• Resolved test non-determinism in the BedrockConverse streaming token usage tests, improving test reliability. 4747a3c
• Corrected the test class naming to properly apply the integration test suffix, ensuring proper test categorization and execution. #5853
• Corrected string parsing logic for the toolChoice field in OpenAiSdkChatModel to ensure proper handling of tool choice configurations. aeb33b0 via #5735
• Fixed an issue where the extra_body parameter was incorrectly included in outgoing OpenAI API requests, which could cause unexpected behavior. 4c0120c
• Resolved issues with Javadoc generation and configuration to ensure API documentation is correctly produced. 0a71804
• Corrected the test bypass condition so integration tests are properly skipped when required API keys are not configured in the environment. bc26dc1

📓 Documentation

• Updated the README to include a note about CPU architecture requirements or compatibility information. a21e988
• Added documentation explaining how MCP servers can re-publish tools from MCP clients, clarifying the tool propagation model in multi-server setups. #5778
• Improved documentation to clarify the intended usage and behavior of the extra_body parameter in OpenAI API requests. 3d4d75b

🔨 Dependency Upgrades

• Updated the Spring Boot dependency to version 3.5.14, incorporating the latest bug fixes and improvements from the Spring Boot team. eb4c9a5
• Updated the Spring Boot dependency to version 3.5.13 as an intermediate upgrade. 9b902f8
• Updated document parsing dependencies: Apache Tika upgraded to 3.3.0, jsoup to 1.22.1, and Apache PDFBox to 3.0.7 for improved document processing capabilities and bug fixes. f25fc52

🔩 Build Updates

• Updated GitHub Actions workflow dependencies to their latest versions to improve CI/CD reliability and security. 9b70b38
• Changed the PR check workflow to use mvn package instead of mvn test for more efficient pull request validation. 7d2e455
• Integration tests are now skipped in the CI pipeline to improve build performance, and the release notes generation workflow has been removed. #5688
• The project has been bumped to the next development version 1.1.5-SNAPSHOT following the release. 400dc42

🔐 Security

• Hardened the default cache directory used for transformer models to prevent unauthorized access or tampering with cached model files. aac6b80
• Fixed a potential denial-of-service vulnerability where a malformed PDF could cause excessive memory allocation during document parsing. b61ac6a

🙏 Contributors

Thanks to all contributors who made this release possible:

• Daniel Garnier-Moiroux (@​Kehrlann)
• David Frizelle (@​dafriz)
• Emmanuel Essien-nta (@​essien)
• Eric Bottard (@​ericbottard)

... (truncated)

• 3d66270 Release version 1.1.5
• eb4c9a5 Upgrade to latest Spring Boot to 3.5.14
• aac6b80 Secure the default cache dir for transformer models
• 6039e57 Fix Cosmosdb doDelete method to parameterize the query
• b61ac6a Prevent excessive char allocation via malformed pdf
• 3cccfdf Properly handle conversationId in VectorStoreChatMemoryAdvisor filter
• 01386e2 fix: vector store filter expression converter key handling
• 455f97a Apply integration test suffix to JdbcChatMemoryRepositorySchemaInitializerPos...
• 4747a3c Fix non-determinism in streamingWithTokenUsage in BedrockConverse tests
• a21e988 Updated README for CPU arch note
• Additional commits viewable in compare view

Sourced from org.springframework.ai:spring-ai-client-chat's releases.

Spring AI 1.1.6 Release Notes

🎯 Highlights

This release includes 1 new features, 5 bug fixes, 2 documentation improvements, 5 other improvements.

⏪ Breaking Changes

• Chat memory advisors now require an explicit conversation ID to be supplied. This is a behavioral change that affects how chat memory is scoped and managed. Applications relying on implicit conversation IDs must be updated to supply an explicit ID. 13cde41

⚠️ Upgrading Notes

• PromptChatMemoryAdvisor is now deprecated and chat memory advisors require an explicit conversation ID. Update your code to: (1) replace PromptChatMemoryAdvisor with the recommended alternative, and (2) ensure an explicit conversation ID is supplied when using any chat memory advisor. 917f62e

⭐ New Features

• MCP auto-configuration now includes the missing @​ConditionalOnMissingBean check, allowing users to provide their own bean definitions and override the auto-configured MCP beans as expected in Spring Boot auto-configuration patterns. d4025e5

🪲 Bug Fixes

• Fixed an issue where the MilvusVectorStore's doDelete method incorrectly escaped strings in the ID list, which could cause deletion operations to fail or behave unexpectedly. 20206a4
• Fixed the ChatClientAdvisorTests test suite to supply an explicit conversation ID, aligning tests with the new requirement for explicit conversation IDs in chat memory advisors. 704e5c6
• Fixed the MistralAiChatModelObservationIT integration test to ensure observation functionality works correctly with the MistralAI chat model provider. a89145d
• Corrects an issue where configured options were not being properly included in MistralAI API requests, ensuring all user-defined settings are correctly passed through. 7bcf32a
• Resolves a regression in how AssistantMessage.ToolCall.id is handled when using the Ollama integration, restoring correct tool call identification behavior. bb9d65e

📓 Documentation

• Corrected typographical errors in the MCP (Model Context Protocol) documentation for improved readability and accuracy. a1ad7f2
• Fixed broken xref anchors in the documentation to restore proper cross-reference navigation between documentation sections. f03c104

🔨 Dependency Upgrades

• Updated MCP SDK from version 0.17.0 to 0.18.2 and MCP annotations from 0.8.0 to 0.9.0, bringing in the latest MCP protocol improvements and bug fixes. #5961

🔩 Build Updates

• Updated the project build to use JDK 17.0.19, ensuring compatibility and incorporating the latest Java 17 patch release for the build environment. 27281e6
• Reorganizes the project structure by relocating Spring AI starter modules to a dedicated starters/ directory for better maintainability and clarity. 22f8676

🙏 Contributors

Thanks to all contributors who made this release possible:

• Christian Tzolov (@​tzolov)
• Emile Plas (@​emileplas)
• Ilayaperumal Gopinathan (@​ilayaperumalg)
• Soby Chacko (@​sobychacko)
• Sébastien Deleuze (@​sdeleuze)

Spring AI 1.1.5 Release Notes

🎯 Highlights

This release includes 9 bug fixes, 3 documentation improvements, 11 other improvements.

⚠️ Upgrading Notes

• The Pixtral 12B model has been removed and Pixtral Large is deprecated. Update your model configuration to use the currently recommended Pixtral models to avoid issues in future releases. 447d2a4

... (truncated)

• 10cd1f4 Release version 1.1.6
• 20206a4 Fix MilvusVectorStore doDelete idList string esacping
• 704e5c6 Fix ChatClientAdvisorTests to supply explicit conversation ID
• 917f62e Deprecate PromptChatMemoryAdvisor
• 13cde41 Require explicit conversation ID in chat memory advisors
• d4025e5 Add missing ConditionalOnMissingBean check on MCP autoconfig
• 974183e Bump MCP dependencies: sdk 0.17.0 → 0.18.2, annotations 0.8.0 → 0.9.0 (#5961)
• a89145d Fix MistralAiChatModelObservationIT
• 27281e6 Switch to JDK 17.0.19 to build the project
• a1ad7f2 Fix typos in MCP documentation
• Additional commits viewable in compare view

You can trigger Dependabot actions by commenting on this PR:

• @dependabot rebase will rebase this PR
• @dependabot recreate will recreate this PR, overwriting any edits that have been made to it
• @dependabot show <dependency name> ignore conditions will show all of the ignore conditions of the specified dependency
• @dependabot ignore <dependency name> major version will close this group update PR and stop Dependabot creating any more for the specific dependency's major version (unless you unignore this specific dependency's major version or upgrade to it yourself)
• @dependabot ignore <dependency name> minor version will close this group update PR and stop Dependabot creating any more for the specific dependency's minor version (unless you unignore this specific dependency's minor version or upgrade to it yourself)
• @dependabot ignore <dependency name> will close this group update PR and stop Dependabot creating any more for the specific dependency (unless you unignore this specific dependency or upgrade to it yourself)
• @dependabot unignore <dependency name> will remove all of the ignore conditions of the specified dependency
• @dependabot unignore <dependency name> <ignore condition> will remove the ignore condition of the specified dependency and ignore conditions
  You can disable automated security fix PRs for this repo from the Security Alerts page.