feat: add in-memory vector search plugin with multilingual support - Implemented the in-memory vector search functionality in the new `vector-search-memory` plugin. - Added English and Chinese localization files for plugin configuration and descriptions. - Created a comprehensive configuration structure for embedding API integration, including API host, key, model, level, and similarity threshold. - Developed core functionalities for document management, including adding, updating, deleting, and searching documents based on cosine similarity. - Introduced synchronization capabilities to bulk index questions and answers from the registered syncer. - Embedded necessary metadata and logging for better traceability and debugging.
Apache Answer Official Plugins are built to enhance the feature of Answer.
Our plugin is under development and the interface definition of the plugin can be viewed at this link.
The Connector plugin helps us to implement third-party login functionality. For example: Google or GitHub OAuth login.
The Storage plugin helps us to upload files to third-party storage. For example: Aliyun OSS or AWS S3.
Using the Cache plugin allows you to store cached data in a different location. For example: Redis or Memcached.
Support using search plugin to speed up the search of question answers. For example: Elasticsearch or Meilisearch.
Using the third-party user system to manage users. For example: WeCom
The Notification plugin helps us to send messages to third-party notification systems. For example: Slack.
Add custom content on the sidebar navigation
Support for custom routing.
Support for extending the markdown editor's toolbar.
Support for customizing the reviewer.
Learn more about the plugin, please visit our docs.
Building the Answer docker image with plugins is easy, see here.