Merge branch 'main' into litellm-provider-config, revert modelCache logic changes

Author: slytechnical

.changeset/curly-plants-pull.md

@@ -0,0 +1,11 @@
+---
+"roo-cline": patch
+---
+
+New models for the Chutes provider:
+
+- Qwen/Qwen3-235B-A22B
+- Qwen/Qwen3-32B
+- Qwen/Qwen3-30B-A3B
+- Qwen/Qwen3-14B
+- Qwen/Qwen3-8B

.changeset/seven-kids-return.md

@@ -0,0 +1,10 @@
+---
+"roo-cline": minor
+---
+
+Adds refresh models button for Unbound provider
+Adds a button above model picker to refresh models based on the current API Key.
+
+1. Clicking the refresh button saves the API Key and calls /models endpoint using that.
+2. Gets the new models and updates the current model if it is invalid for the given API Key.
+3. The refresh button also flushes existing Unbound models and refetches them.

.gitignore

@@ -36,3 +36,7 @@ logs
 
 # Vite development
 .vite-port
+
+# IntelliJ and Qodo plugin folders
+.idea/
+.qodo/

.vscodeignore

@@ -68,3 +68,7 @@ assets/docs/**
 
 # Include .env file for telemetry
 !.env
+
+# Ignore IntelliJ and Qodo plugin folders
+.idea/**
+.qodo/**

esbuild.js

@@ -34,10 +34,17 @@ const copyWasmFiles = {
 
 			// tiktoken WASM file
 			fs.copyFileSync(
-				path.join(nodeModulesDir, "tiktoken", "tiktoken_bg.wasm"),
+				path.join(nodeModulesDir, "tiktoken", "lite", "tiktoken_bg.wasm"),
 				path.join(distDir, "tiktoken_bg.wasm"),
 			)
 
+			// Also copy to the workers directory
+			fs.mkdirSync(path.join(distDir, "workers"), { recursive: true })
+			fs.copyFileSync(
+				path.join(nodeModulesDir, "tiktoken", "lite", "tiktoken_bg.wasm"),
+				path.join(distDir, "workers", "tiktoken_bg.wasm"),
+			)
+
 			// Main tree-sitter WASM file
 			fs.copyFileSync(
 				path.join(nodeModulesDir, "web-tree-sitter", "tree-sitter.wasm"),

src/api/providers/fetchers/modelCache.ts

@@ -6,6 +6,7 @@ import NodeCache from "node-cache"
 import { ContextProxy } from "../../../core/config/ContextProxy"
 import { getCacheDirectoryPath } from "../../../shared/storagePathManager"
 import { RouterName, ModelRecord } from "../../../shared/api"
+import { fileExistsAtPath } from "../../../utils/fs"
 
 import { getOpenRouterModels } from "./openrouter"
 import { getRequestyModels } from "./requesty"
@@ -21,6 +22,14 @@ async function writeModels(router: RouterName, data: ModelRecord) {
 	await fs.writeFile(path.join(cacheDir, filename), JSON.stringify(data))
 }
 
+async function readModels(router: RouterName): Promise<ModelRecord | undefined> {
+	const filename = `${router}_models.json`
+	const cacheDir = await getCacheDirectoryPath(ContextProxy.instance.globalStorageUri.fsPath)
+	const filePath = path.join(cacheDir, filename)
+	const exists = await fileExistsAtPath(filePath)
+	return exists ? JSON.parse(await fs.readFile(filePath, "utf8")) : undefined
+}
+
 /**
  * Get models from the cache or fetch them from the provider and cache them.
  * There are two caches:
@@ -37,37 +46,34 @@ export const getModels = async (
 	apiKey: string | undefined = undefined,
 	baseUrl: string | undefined = undefined,
 ): Promise<ModelRecord> => {
-	// If this call is meant for a refresh (indicated by apiKey/baseUrl for specific routers),
-	// the memory cache should have been flushed by the caller (e.g., webviewMessageHandler).
-	// Otherwise, for general calls, check memory cache first.
-	const modelsFromMemory = memoryCache.get<ModelRecord>(router)
-	if (modelsFromMemory) {
-		return modelsFromMemory
+	let models = memoryCache.get<ModelRecord>(router)
+	if (models) {
+		return models
 	}
 
-	let fetchedModels: ModelRecord
 	try {
 		switch (router) {
 			case "openrouter":
-				fetchedModels = await getOpenRouterModels()
+				models = await getOpenRouterModels()
 				break
 			case "requesty":
-				// Assuming getRequestyModels will throw if apiKey is needed and not provided or invalid.
-				fetchedModels = await getRequestyModels(apiKey)
+				// Requesty models endpoint requires an API key for per-user custom policies
+				models = await getRequestyModels(apiKey)
 				break
 			case "glama":
-				fetchedModels = await getGlamaModels()
+				models = await getGlamaModels()
 				break
 			case "unbound":
-				fetchedModels = await getUnboundModels()
+				// Unbound models endpoint requires an API key to fetch application specific models
+				models = await getUnboundModels(apiKey)
 				break
 			case "litellm":
 				if (!baseUrl || !apiKey) {
 					// This case should ideally be handled by the caller if baseUrl is strictly required.
 					// However, for robustness, if called without baseUrl for litellm, it would fail in getLiteLLMModels or here.
 					throw new Error("Base URL and api key are required for LiteLLM models.")
 				}
-				fetchedModels = await getLiteLLMModels(apiKey || "", baseUrl)
+				models = await getLiteLLMModels(apiKey || "", baseUrl)
 				break
 			default:
 				// Ensures router is exhaustively checked if RouterName is a strict union
@@ -76,11 +82,18 @@ export const getModels = async (
 		}
 
 		// Cache the fetched models (even if empty, to signify a successful fetch with no models)
-		memoryCache.set(router, fetchedModels)
-		await writeModels(router, fetchedModels).catch((err) =>
+		memoryCache.set(router, models)
+		await writeModels(router, models).catch((err) =>
 			console.error(`[getModels] Error writing ${router} models to file cache:`, err),
 		)
-		return fetchedModels
+
+		try {
+			models = await readModels(router)
+			// console.log(`[getModels] read ${router} models from file cache`)
+		} catch (error) {
+			console.error(`[getModels] error reading ${router} models from file cache`, error)
+		}
+		return models || {}
 	} catch (error) {
 		// Log the error and re-throw it so the caller can handle it (e.g., show a UI message).
 		console.error(`[getModels] Failed to fetch models for ${router}:`, error)
@@ -89,10 +102,6 @@ export const getModels = async (
 	}
 }
 
-/**
- * Flush models memory cache for a specific router
- * @param router - The router to flush models for.
- */
 export const flushModels = async (router: RouterName) => {
 	memoryCache.del(router)
 }

src/api/providers/fetchers/unbound.ts

@@ -2,11 +2,17 @@ import axios from "axios"
 
 import { ModelInfo } from "../../../shared/api"
 
-export async function getUnboundModels(): Promise<Record<string, ModelInfo>> {
+export async function getUnboundModels(apiKey?: string | null): Promise<Record<string, ModelInfo>> {
 	const models: Record<string, ModelInfo> = {}
 
 	try {
-		const response = await axios.get("https://api.getunbound.ai/models")
+		const headers: Record<string, string> = {}
+
+		if (apiKey) {
+			headers["Authorization"] = `Bearer ${apiKey}`
+		}
+
+		const response = await axios.get("https://api.getunbound.ai/models", { headers })
 
 		if (response.data) {
 			const rawModels: Record<string, any> = response.data
@@ -40,6 +46,7 @@ export async function getUnboundModels(): Promise<Record<string, ModelInfo>> {
 		}
 	} catch (error) {
 		console.error(`Error fetching Unbound models: ${JSON.stringify(error, Object.getOwnPropertyNames(error), 2)}`)
+		throw new Error(`Failed to fetch Unbound models: ${error instanceof Error ? error.message : "Unknown error"}`)
 	}
 
 	return models

src/core/task/Task.ts

@@ -350,7 +350,7 @@ export class Task extends EventEmitter<ClineEvents> {
 
 			await this.providerRef.deref()?.updateTaskHistory(historyItem)
 		} catch (error) {
-			console.error("Failed to save cline messages:", error)
+			console.error("Failed to save Roo messages:", error)
 		}
 	}
 
@@ -372,7 +372,7 @@ export class Task extends EventEmitter<ClineEvents> {
 		// simply removes the reference to this instance, but the instance is
 		// still alive until this promise resolves or rejects.)
 		if (this.abort) {
-			throw new Error(`[Cline#ask] task ${this.taskId}.${this.instanceId} aborted`)
+			throw new Error(`[RooCode#ask] task ${this.taskId}.${this.instanceId} aborted`)
 		}
 
 		let askTs: number
@@ -492,7 +492,7 @@ export class Task extends EventEmitter<ClineEvents> {
 		} = {},
 	): Promise<undefined> {
 		if (this.abort) {
-			throw new Error(`[Cline#say] task ${this.taskId}.${this.instanceId} aborted`)
+			throw new Error(`[RooCode#say] task ${this.taskId}.${this.instanceId} aborted`)
 		}
 
 		if (partial !== undefined) {
@@ -623,7 +623,7 @@ export class Task extends EventEmitter<ClineEvents> {
 		} catch (error) {
 			this.providerRef
 				.deref()
-				?.log(`Error failed to add reply from subtast into conversation of parent task, error: ${error}`)
+				?.log(`Error failed to add reply from subtask into conversation of parent task, error: ${error}`)
 
 			throw error
 		}
@@ -957,7 +957,7 @@ export class Task extends EventEmitter<ClineEvents> {
 		includeFileDetails: boolean = false,
 	): Promise<boolean> {
 		if (this.abort) {
-			throw new Error(`[Cline#recursivelyMakeClineRequests] task ${this.taskId}.${this.instanceId} aborted`)
+			throw new Error(`[RooCode#recursivelyMakeRooRequests] task ${this.taskId}.${this.instanceId} aborted`)
 		}
 
 		if (this.consecutiveMistakeCount >= this.consecutiveMistakeLimit) {
@@ -1253,7 +1253,7 @@ export class Task extends EventEmitter<ClineEvents> {
 
 			// Need to call here in case the stream was aborted.
 			if (this.abort || this.abandoned) {
-				throw new Error(`[Cline#recursivelyMakeClineRequests] task ${this.taskId}.${this.instanceId} aborted`)
+				throw new Error(`[RooCode#recursivelyMakeRooRequests] task ${this.taskId}.${this.instanceId} aborted`)
 			}
 
 			this.didCompleteReadingStream = true

src/shared/api.ts

@@ -1533,6 +1533,11 @@ export type ChutesModelId =
 	| "deepseek-ai/DeepSeek-V3-Base"
 	| "deepseek-ai/DeepSeek-R1-Zero"
 	| "deepseek-ai/DeepSeek-V3-0324"
+	| "Qwen/Qwen3-235B-A22B"
+	| "Qwen/Qwen3-32B"
+	| "Qwen/Qwen3-30B-A3B"
+	| "Qwen/Qwen3-14B"
+	| "Qwen/Qwen3-8B"
 	| "microsoft/MAI-DS-R1-FP8"
 	| "tngtech/DeepSeek-R1T-Chimera"
 export const chutesDefaultModelId: ChutesModelId = "deepseek-ai/DeepSeek-R1"
@@ -1663,6 +1668,51 @@ export const chutesModels = {
 		outputPrice: 0,
 		description: "DeepSeek V3 (0324) model.",
 	},
+	"Qwen/Qwen3-235B-A22B": {
+		maxTokens: 32768,
+		contextWindow: 40960,
+		supportsImages: false,
+		supportsPromptCache: false,
+		inputPrice: 0,
+		outputPrice: 0,
+		description: "Qwen3 235B A22B model.",
+	},
+	"Qwen/Qwen3-32B": {
+		maxTokens: 32768,
+		contextWindow: 40960,
+		supportsImages: false,
+		supportsPromptCache: false,
+		inputPrice: 0,
+		outputPrice: 0,
+		description: "Qwen3 32B model.",
+	},
+	"Qwen/Qwen3-30B-A3B": {
+		maxTokens: 32768,
+		contextWindow: 40960,
+		supportsImages: false,
+		supportsPromptCache: false,
+		inputPrice: 0,
+		outputPrice: 0,
+		description: "Qwen3 30B A3B model.",
+	},
+	"Qwen/Qwen3-14B": {
+		maxTokens: 32768,
+		contextWindow: 40960,
+		supportsImages: false,
+		supportsPromptCache: false,
+		inputPrice: 0,
+		outputPrice: 0,
+		description: "Qwen3 14B model.",
+	},
+	"Qwen/Qwen3-8B": {
+		maxTokens: 32768,
+		contextWindow: 40960,
+		supportsImages: false,
+		supportsPromptCache: false,
+		inputPrice: 0,
+		outputPrice: 0,
+		description: "Qwen3 8B model.",
+	},
 	"microsoft/MAI-DS-R1-FP8": {
 		maxTokens: 32768,
 		contextWindow: 163840,

tsconfig.json

@@ -21,5 +21,5 @@
 		"useUnknownInCatchVariables": false
 	},
 	"include": ["src/**/*", "scripts/**/*", ".changeset/**/*"],
-	"exclude": ["node_modules", ".vscode-test", "webview-ui"]
+	"exclude": ["node_modules", ".vscode-test", "webview-ui", ".idea", ".qodo"]
 }