Commit Graph

23 Commits (014f95f1394a7bfd1c7bbbdcdc19057a80dcaca5)

Author SHA1 Message Date
lyingbug 2be7bd4063
fix(agent): search bound knowledge bases before answering, and pass suggested-question origins (#3400)
* fix(agent): search bound knowledge bases before answering, and pass suggested-question origins

A user picked the suggested question "为什么在比较图形时需要谨慎处理?"
and the agent answered in round 1 with no tool call, claiming it had
searched the knowledge base and found nothing, then explained misleading
statistical charts. The question had been generated from "Corners Don't
Look Like That: Regarding Screenspace Ambient Occlusion" in a bound base,
where "comparing graphs" means SSAO falloff curves.

Two things let that happen: the grounding prompt said "Stable general
explanations need no lookup", and nothing told the agent where the
question came from - the base in question had no profile, only two
recent titles in the directory, and the click sent the text alone.

Prompt
- With knowledge tools available, search bound bases before answering a
  question about a topic, even one that reads like general knowledge;
  directory profiles and recent titles cannot prove a base lacks it.
- The general-explanation exemption applies only without bound bases.
- Say you searched or found nothing only when a search tool ran.

Question origin
- Suggested questions from FAQ/document chunks also return knowledge_id.
- Picking one on the home or chat page sends question_origin
  {knowledge_base_id, knowledge_id} with that question only.
- The service keeps the origin only when its base is inside the turn's
  resolved scope and the document belongs to that base; it never
  changes the scope (unlike knowledge_base_ids, which narrows it).
- The engine renders it in runtime_context as <question_origin> with the
  base and document handles and a note to search it before answering.

* fix(agent): keep question origins across file/tag scopes and bind them to the send

Review follow-up for the suggested-question origin.

- Scope check: the origin was kept only when its base was in the resolved
  KB list, which is empty for file-only and tag-only scopes even though
  their bases are searched. It is now checked against SearchTargets after
  they are built; the document is kept only when a target for that base
  covers it (unfiltered base, or a document scope listing it).
  resolveQuestionOriginInfo re-checks SearchTargets, looks up the name of
  a base that is not in kbInfos, and the renderer omits an unknown name.
- Frontend: the origin travelled through module-level "pending" state, so
  a send that returned early (agent not ready, upload in progress,
  steering) left it behind and a second pick overwrote it. It now rides
  with the one send: triggerSend(text, options) -> createSession ->
  send-msg -> sendMsg / changeFirstQuery.
- RetrieveKBOnlyWhenMentioned: suggestions are still generated from the
  agent's bases, but picking one retrieved nothing. The origin now
  selects its base, as an @mention would, when the agent may read it.
- API sync: question_origin in docs/api/chat.md and swagger;
  QuestionOrigin on the Go SDK's KnowledgeQARequest/AgentQARequest and
  knowledge_id on SuggestedQuestion.
2026-09-18 17:00:47 +08:00
wizardchen e71c2dce2e fix(api): address resource_urls review follow-ups
Harden public URL mode: stop logging signed URLs at INFO, copy messages
before rewriting history responses, reuse winning grants on insert races,
sync the Go client resource_urls query param, and add knowledge-search
coverage.
2026-08-05 14:42:51 +08:00
langcaiye 1708d6895a fix(search): apply tag filters to knowledge search 2026-07-06 23:30:33 +08:00
nullkey 8f36ea2229 feat(sdk): typed errors, enums, and KB model config for the CLI
Adds the client-library surface the agent-first CLI builds on:
- named server error-code constants + typed *APIError from non-2xx responses,
- MessageSearchMode / AgentMode / KBSelectionMode and model type/source
  enumerators so the CLI validates flags against the server vocabulary,
- SetKBModelConfig for binding embedding/chat models to a knowledge base.
2026-07-06 22:53:01 +08:00
wizardchen 1109e85a57 fix(cli): align stream error handling after output-mode merge
Follow up on #1740/#1741 by typing terminal SSE errors in the SDK,
classifying them as server.error in the CLI/MCP, marking Projector done
on error frames, attaching session_id to MCP stream failures, and
migrating the e2e chat step to the bounded JSON envelope.
2026-07-03 00:30:34 +08:00
coconut-yc 238297e2c5 fix(client): harden chat and agent SSE reliability 2026-07-03 00:13:41 +08:00
nullkey fc694d5d21 fix(sdk): silence stdout logging via opt-in slog
The client SDK emitted 17 fmt.Printf / fmt.Println trace events directly to
stdout from KnowledgeQAStream and SearchKnowledge. Any consumer of the SDK —
the new CLI in particular — saw these lines interleaved with its own
output, breaking JSON envelope contracts and polluting human-mode tty.

Replace with a package-level slog logger. Default writes to io.Discard so
existing callers see no output change; set WEKNORA_SDK_DEBUG=1 to route
events to stderr for development.

No public SetLogger() API: the trace events were dev-only and a setter
would expand the SDK surface for a one-off use case.
2026-05-12 13:20:42 +08:00
wizardchen 5b2ebd1f44 feat(channel): add channel support across various components and locales
- Introduced a new `channel` field in multiple request and message structures to track the source channel (e.g., "web", "api", "im").
- Updated frontend API calls and chat components to include the `channel` parameter, ensuring consistent channel tracking in user messages.
- Enhanced localization files to support channel labels in English, Korean, Russian, and Chinese.
- Added database migration scripts to incorporate the `channel` column in the messages table, facilitating the storage of channel information.
- Refactored related functions and components to accommodate the new channel feature, improving overall message context and tracking.
2026-03-25 22:08:29 +08:00
wizardchen bf1ae9c018 feat: enhance agent and user information structures
- Added support for image attachments in agent Q&A requests, enabling multimodal chat capabilities.
- Introduced a new UserInfo struct to represent user information in API responses, improving user data management.
- Updated existing request structures to include image handling, enhancing the overall functionality of the chat system.

These changes significantly improve the application's ability to handle multimodal interactions and manage user data effectively.
2026-03-12 15:08:49 +08:00
wizardchen daa9ef500c feat: add agent management functionality
- Introduced a new package for managing custom agents, including CRUD operations for agent creation, retrieval, updating, and deletion.
- Implemented API endpoints for listing agents and retrieving agent placeholders.
- Added data structures for agent configuration and requests, enhancing the overall agent management capabilities.
- Enhanced the client with methods to interact with the new agent management features, improving user experience in managing agents.

These changes significantly expand the application's functionality for handling custom agents, providing users with a comprehensive toolset for agent management.
2026-03-09 17:34:13 +08:00
wizardchen 72a3cbb0ba feat: 增加流式响应类型和工具调用支持
新增 LLMToolCall 和 FunctionCall 结构体支持工具调用,扩展 ResponseType 枚举添加多种响应类型,优化 StreamResponse 结构体增加会话和工具调用字段,并改进引用事件处理逻辑。
2025-12-31 15:36:20 +08:00
wizardchen c5228acbab refactor: 简化会话创建请求结构
移除会话策略和代理配置相关字段,仅保留标题和描述。会话现在作为独立的对话容器,配置由自定义代理在查询时动态提供。
2025-12-31 15:36:20 +08:00
wizardchen b0a9952785 feat: 添加禁用自动标题生成的选项 2025-12-31 15:36:20 +08:00
wizardchen 0d020ce30c feat: 添加思考模式配置支持
在SummaryConfig结构体中新增Thinking字段,支持启用思考模式功能。更新相关配置文件、类型定义和客户端代码以支持该功能。
2025-12-31 15:36:20 +08:00
wizardchen d6a1647be3 feat: 增强 KnowledgeQAStream Client参数 2025-12-31 15:36:20 +08:00
wizardchen 423cb3a2a3 feat: 增强知识库搜索功能,支持多知识库和指定文件ID搜索;优化SearchKnowledgeRequest结构,向后兼容单知识库ID参数 2025-12-24 20:47:08 +08:00
wizardchen 12497a2fea feat: 重构客户端结构体,增强块、消息、会话和租户模型,并更新API文档 2025-12-02 22:47:13 +08:00
wizardchen eb8bfddfd1 refactor: Update function signatures across multiple files to improve readability by adding context parameters and enhancing code structure 2025-11-28 16:17:52 +08:00
wizardchen c222993777 refactor: Enhance AgentQARequest structure with additional fields and improve session handling methods for clarity and functionality 2025-11-27 17:56:53 +08:00
wizardchen 601f73ec5f feat: Refactor agent configuration by removing deprecated parameters and enhancing conversation model management
- Removed the `thinkingModelId` and `rerankModelId` from agent configuration, streamlining the settings structure.
- Introduced a new `conversationModels` interface to manage summary and rerank model IDs.
- Updated related UI components and API logic to accommodate the changes in agent configuration.
- Enhanced localization for new conversation model settings in English, Russian, and Chinese.
2025-11-20 10:33:26 +08:00
wizardchen b9c1191dd2 feat: Introduce query expansion feature and enhance related configurations
- Added support for query expansion in conversation settings, allowing the system to generate additional queries when recall is low.
- Updated localization files in English, Russian, and Chinese to include new strings for the query expansion feature.
- Enhanced UI in AgentSettings.vue to allow users to toggle query expansion settings.
- Refactored related API and service logic to accommodate the new query expansion functionality in session management and chat processing.
2025-11-19 20:23:30 +08:00
wizardchen 4fa3adbf3b feat: Add agent configuration and cleanup scripts for database migrations 2025-11-05 23:18:44 +08:00
wizardchen 56eb2bce33 init commit 2025-08-05 15:08:07 +08:00