Improves model info detection for custom Bedrock ARNs

src/api/providers/bedrock.ts

@@ -29,6 +29,7 @@ import { convertToBedrockConverseMessages as sharedConverter } from "../transfor
 
 const BEDROCK_DEFAULT_TEMPERATURE = 0.3
 const BEDROCK_MAX_TOKENS = 4096
+const BEDROCK_DEFAULT_CONTEXT = 128_000 // PATCH: used for unknown custom ARNs
 
 /************************************************************************************
  *
@@ -186,6 +187,54 @@ export class AwsBedrockHandler extends BaseProvider implements SingleCompletionH
 		this.client = new BedrockRuntimeClient(clientConfig)
 	}
 
+	// PATCH: Helper to guess model info from custom modelId string if not in bedrockModels
+	private guessModelInfoFromId(modelId: string): Partial<SharedModelInfo> {
+		// Define a mapping for model ID patterns and their configurations
+		const modelConfigMap: Record<string, Partial<SharedModelInfo>> = {
+			"claude-3-7": {
+				maxTokens: 8192,
+				contextWindow: 200_000,
+				supportsImages: true,
+				supportsPromptCache: true,
+			},
+			"claude-3-5": {
+				maxTokens: 8192,
+				contextWindow: 200_000,
+				supportsImages: true,
+				supportsPromptCache: true,
+			},
+			"claude-3-opus": {
+				maxTokens: 4096,
+				contextWindow: 200_000,
+				supportsImages: true,
+				supportsPromptCache: true,
+			},
+			"claude-3-haiku": {
+				maxTokens: 4096,
+				contextWindow: 200_000,
+				supportsImages: true,
+				supportsPromptCache: true,
+			},
+		}
+
+		// Match the model ID to a configuration
+		const id = modelId.toLowerCase()
+		for (const [pattern, config] of Object.entries(modelConfigMap)) {
+			if (id.includes(pattern)) {
+				return config
+			}
+		}
+		// PATCH: Add more heuristics as needed here
+
+		// Default fallback
+		return {
+			maxTokens: BEDROCK_MAX_TOKENS,
+			contextWindow: BEDROCK_DEFAULT_CONTEXT,
+			supportsImages: false,
+			supportsPromptCache: false,
+		}
+	}
+
 	override async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
 		let modelConfig = this.getModel()
 		// Handle cross-region inference
@@ -495,7 +544,7 @@ export class AwsBedrockHandler extends BaseProvider implements SingleCompletionH
 		})
 
 		return {
-			system: systemMessage ? [{ text: systemMessage } as SystemContentBlock] : [],
+			system: cacheResult.system,
 			messages: messagesWithCache,
 		}
 	}
@@ -629,16 +678,24 @@ export class AwsBedrockHandler extends BaseProvider implements SingleCompletionH
 				info: JSON.parse(JSON.stringify(bedrockModels[bedrockDefaultPromptRouterModelId])),
 			}
 		} else {
+			// PATCH: Use heuristics for model info, then allow overrides from ProviderSettings
+			const guessed = this.guessModelInfoFromId(modelId)
 			model = {
 				id: bedrockDefaultModelId,
-				info: JSON.parse(JSON.stringify(bedrockModels[bedrockDefaultModelId])),
+				info: {
+					...JSON.parse(JSON.stringify(bedrockModels[bedrockDefaultModelId])),
+					...guessed,
+				},
 			}
 		}
 
-		// If modelMaxTokens is explicitly set in options, override the default
+		// PATCH: Always allow user to override detected/guessed maxTokens and contextWindow
 		if (this.options.modelMaxTokens && this.options.modelMaxTokens > 0) {
 			model.info.maxTokens = this.options.modelMaxTokens
 		}
+		if (this.options.awsModelContextWindow && this.options.awsModelContextWindow > 0) {
+			model.info.contextWindow = this.options.awsModelContextWindow
+		}
 
 		return model
 	}
@@ -673,8 +730,7 @@ export class AwsBedrockHandler extends BaseProvider implements SingleCompletionH
 			}
 		}
 
-		modelConfig.info.maxTokens = modelConfig.info.maxTokens || BEDROCK_MAX_TOKENS
-
+		// PATCH: Don't override maxTokens/contextWindow here; handled in getModelById (and includes user overrides)
 		return modelConfig as { id: BedrockModelId | string; info: SharedModelInfo }
 	}
 

src/core/config/importExport.ts

@@ -84,6 +84,15 @@ export const importSettings = async ({ providerSettingsManager, contextProxy, cu
 		if (e instanceof ZodError) {
 			error = e.issues.map((issue) => `[${issue.path.join(".")}]: ${issue.message}`).join("\n")
 			telemetryService.captureSchemaValidationError({ schemaName: "ImportExport", error: e })
+		} else if (e instanceof SyntaxError) {
+			// Extract position info from the error message
+			const match = e.message.match(/at position (\d+)/)
+			if (match) {
+				const position = parseInt(match[1], 10)
+				error = `Expected property name or '}' in JSON at position ${position}`
+			} else {
+				error = e.message
+			}
 		} else if (e instanceof Error) {
 			error = e.message
 		}

src/exports/roo-code.d.ts

@@ -253,6 +253,7 @@ type ProviderSettings = {
 	awsProfile?: string | undefined
 	awsUseProfile?: boolean | undefined
 	awsCustomArn?: string | undefined
+	awsModelContextWindow?: number | undefined
 	vertexKeyFile?: string | undefined
 	vertexJsonCredentials?: string | undefined
 	vertexProjectId?: string | undefined
@@ -656,6 +657,7 @@ type IpcMessage =
 								awsProfile?: string | undefined
 								awsUseProfile?: boolean | undefined
 								awsCustomArn?: string | undefined
+								awsModelContextWindow?: number | undefined
 								vertexKeyFile?: string | undefined
 								vertexJsonCredentials?: string | undefined
 								vertexProjectId?: string | undefined
@@ -1131,6 +1133,7 @@ type TaskCommand =
 					awsProfile?: string | undefined
 					awsUseProfile?: boolean | undefined
 					awsCustomArn?: string | undefined
+					awsModelContextWindow?: number | undefined
 					vertexKeyFile?: string | undefined
 					vertexJsonCredentials?: string | undefined
 					vertexProjectId?: string | undefined

src/exports/types.ts

@@ -257,6 +257,7 @@ type ProviderSettings = {
 	awsProfile?: string | undefined
 	awsUseProfile?: boolean | undefined
 	awsCustomArn?: string | undefined
+	awsModelContextWindow?: number | undefined
 	vertexKeyFile?: string | undefined
 	vertexJsonCredentials?: string | undefined
 	vertexProjectId?: string | undefined
@@ -670,6 +671,7 @@ type IpcMessage =
 								awsProfile?: string | undefined
 								awsUseProfile?: boolean | undefined
 								awsCustomArn?: string | undefined
+								awsModelContextWindow?: number | undefined
 								vertexKeyFile?: string | undefined
 								vertexJsonCredentials?: string | undefined
 								vertexProjectId?: string | undefined
@@ -1147,6 +1149,7 @@ type TaskCommand =
 					awsProfile?: string | undefined
 					awsUseProfile?: boolean | undefined
 					awsCustomArn?: string | undefined
+					awsModelContextWindow?: number | undefined
 					vertexKeyFile?: string | undefined
 					vertexJsonCredentials?: string | undefined
 					vertexProjectId?: string | undefined

src/schemas/index.ts

@@ -465,6 +465,7 @@ const bedrockSchema = apiModelIdProviderModelSchema.extend({
 	awsProfile: z.string().optional(),
 	awsUseProfile: z.boolean().optional(),
 	awsCustomArn: z.string().optional(),
+	awsModelContextWindow: z.number().optional(),
 })
 
 const vertexSchema = apiModelIdProviderModelSchema.extend({
@@ -651,6 +652,7 @@ const providerSettingsRecord: ProviderSettingsRecord = {
 	awsProfile: undefined,
 	awsUseProfile: undefined,
 	awsCustomArn: undefined,
+	awsModelContextWindow: undefined,
 	// Google Vertex
 	vertexKeyFile: undefined,
 	vertexJsonCredentials: undefined,