# Knowledge Retrieval Node Configuration # This file defines the metadata for the Knowledge Retrieval workflow node # The corresponding Python implementation is in: pkg/workflow/nodes/knowledge_retrieval.py name: knowledge_retrieval label: en_US: Knowledge Retrieval zh_Hans: 知识库检索 category: process icon: Search color: '#8b5cf6' description: en_US: Retrieve relevant information from knowledge bases zh_Hans: 从知识库中检索相关信息 inputs: - name: query type: string label: en_US: Query zh_Hans: 查询 description: en_US: Query text to search for zh_Hans: 要搜索的查询文本 outputs: - name: results type: array label: en_US: Results zh_Hans: 结果 description: en_US: Retrieved documents/chunks zh_Hans: 检索到的文档/块 - name: context type: string label: en_US: Context zh_Hans: 上下文 description: en_US: Concatenated text from all results zh_Hans: 所有结果的连接文本 - name: scores type: array label: en_US: Scores zh_Hans: 分数 description: en_US: Similarity scores for each result zh_Hans: 每个结果的相似度分数 config: - name: knowledge_bases type: knowledge-base-multi-selector required: true default: [] label: en_US: Knowledge Bases zh_Hans: 知识库 description: en_US: Select knowledge bases to search zh_Hans: 选择要搜索的知识库 - name: top_k type: integer default: 5 label: en_US: Top K Results zh_Hans: 返回数量 (Top K) description: en_US: Number of top results to retrieve zh_Hans: 返回的最相关结果数量 - name: similarity_threshold type: number default: 0.5 label: en_US: Similarity Threshold zh_Hans: 相似度阈值 description: en_US: Minimum similarity score for results zh_Hans: 结果的最小相似度分数 - name: retrieval_mode type: select default: vector options: - vector - hybrid - keyword label: en_US: Retrieval Mode zh_Hans: 检索模式 description: en_US: Method used for retrieving documents zh_Hans: 用于检索文档的方法 - name: rerank_enabled type: boolean default: false label: en_US: Enable Reranking zh_Hans: 启用重排序 description: en_US: Use a reranking model to improve result relevance zh_Hans: 使用重排序模型提高结果相关性 - name: rerank_model type: rerank-model-selector default: "" label: en_US: Rerank Model zh_Hans: 重排序模型 description: en_US: Model to use for reranking results zh_Hans: 用于结果重排序的模型 show_if: rerank_enabled: true