FlowiseAI/Flowise

[FEATURE] Custom Cohere ReRank API endpoints

Open

#4,086 建立於 2025年2月26日

在 GitHub 查看
 (3 留言) (0 反應) (0 負責人)TypeScript (52,823 star) (24,335 fork)batch import
enhancementhelp wanted

描述

Cohere allows users to create custom fine-tuned rerank models, but the current rerank node doesn't allow us to set such a model. Perhaps it could be added as an option under additional parameters? I believe @gururise did the implementation for 3.5, is this something you could look into?

貢獻者指南

技術棧
typescriptnodejsrest api
領域
backendapi
議題類型
feature
難度面向新貢獻者的預計實作難度，1 表示很小改動，5 表示專家級工作。
3
預計時間有經驗貢獻者完成調查、實作、測試並準備 pull request 的粗略時間範圍。
half day
活動狀態議題目前的可參與程度：新鮮、活躍、陳舊、阻塞或等待維護者輸入。
fresh
清晰度議題是否清楚說明預期改動、驗收標準和下一步。
mostly clear
前置要求
Basic knowledge of TypeScriptFamiliarity with REST API integrationUnderstanding of Cohere API
新手友善度1-100 的估計分數，表示該議題對首次貢獻者的友善程度。
45
研究方向
First, locate the rerank node implementation in the Flowise codebase (likely under packages/components/nodes/rerank). Review the current Cohere rerank API call and identify where additional parameters are passed. Look at how the custom model field could be added, possibly in the node's configuration form (e.g., a text input for model name). Check the Cohere API documentation for fine tuned rerank model endpoints. Also examine any previous PR by @gururise for reference on how they implemented similar parameter additions. Then implement the change to accept an optional model parameter and pass it to the API request.