Add gemini support

src/api/gemini.ts

@@ -0,0 +1,57 @@
+import { Anthropic } from "@anthropic-ai/sdk"
+import { FunctionCallingMode, GoogleGenerativeAI } from "@google/generative-ai"
+import { ApiHandler, ApiHandlerMessageResponse } from "."
+import { ApiHandlerOptions, geminiDefaultModelId, GeminiModelId, geminiModels, ModelInfo } from "../shared/api"
+import {
+	convertAnthropicMessageToGemini,
+	convertAnthropicToolToGemini,
+	convertGeminiResponseToAnthropic,
+} from "../utils/gemini-format"
+
+export class GeminiHandler implements ApiHandler {
+	private options: ApiHandlerOptions
+	private client: GoogleGenerativeAI
+
+	constructor(options: ApiHandlerOptions) {
+		if (!options.geminiApiKey) {
+			throw new Error("API key is required for Google Gemini")
+		}
+		this.options = options
+		this.client = new GoogleGenerativeAI(options.geminiApiKey)
+	}
+
+	async createMessage(
+		systemPrompt: string,
+		messages: Anthropic.Messages.MessageParam[],
+		tools: Anthropic.Messages.Tool[]
+	): Promise<ApiHandlerMessageResponse> {
+		const model = this.client.getGenerativeModel({
+			model: this.getModel().id,
+			systemInstruction: systemPrompt,
+			tools: [{ functionDeclarations: tools.map(convertAnthropicToolToGemini) }],
+			toolConfig: {
+				functionCallingConfig: {
+					mode: FunctionCallingMode.AUTO,
+				},
+			},
+		})
+		const result = await model.generateContent({
+			contents: messages.map(convertAnthropicMessageToGemini),
+			generationConfig: {
+				maxOutputTokens: this.getModel().info.maxTokens,
+			},
+		})
+		const message = convertGeminiResponseToAnthropic(result.response)
+
+		return { message }
+	}
+
+	getModel(): { id: GeminiModelId; info: ModelInfo } {
+		const modelId = this.options.apiModelId
+		if (modelId && modelId in geminiModels) {
+			const id = modelId as GeminiModelId
+			return { id, info: geminiModels[id] }
+		}
+		return { id: geminiDefaultModelId, info: geminiModels[geminiDefaultModelId] }
+	}
+}

src/api/index.ts

@@ -6,6 +6,7 @@ import { OpenRouterHandler } from "./openrouter"
 import { VertexHandler } from "./vertex"
 import { OpenAiHandler } from "./openai"
 import { OllamaHandler } from "./ollama"
+import { GeminiHandler } from "./gemini"
 
 export interface ApiHandlerMessageResponse {
 	message: Anthropic.Messages.Message
@@ -37,6 +38,8 @@ export function buildApiHandler(configuration: ApiConfiguration): ApiHandler {
 			return new OpenAiHandler(options)
 		case "ollama":
 			return new OllamaHandler(options)
+		case "gemini":
+			return new GeminiHandler(options)
 		default:
 			return new AnthropicHandler(options)
 	}

src/providers/ClaudeDevProvider.ts

@@ -18,7 +18,14 @@ https://github.com/microsoft/vscode-webview-ui-toolkit-samples/blob/main/default
 https://github.com/KumarVariable/vscode-extension-sidebar-html/blob/master/src/customSidebarViewProvider.ts
 */
 
-type SecretKey = "apiKey" | "openRouterApiKey" | "awsAccessKey" | "awsSecretKey" | "awsSessionToken" | "openAiApiKey"
+type SecretKey =
+	| "apiKey"
+	| "openRouterApiKey"
+	| "awsAccessKey"
+	| "awsSecretKey"
+	| "awsSessionToken"
+	| "openAiApiKey"
+	| "geminiApiKey"
 type GlobalStateKey =
 	| "apiProvider"
 	| "apiModelId"
@@ -329,6 +336,7 @@ export class ClaudeDevProvider implements vscode.WebviewViewProvider {
 								ollamaModelId,
 								ollamaBaseUrl,
 								anthropicBaseUrl,
+								geminiApiKey,
 							} = message.apiConfiguration
 							await this.updateGlobalState("apiProvider", apiProvider)
 							await this.updateGlobalState("apiModelId", apiModelId)
@@ -346,6 +354,7 @@ export class ClaudeDevProvider implements vscode.WebviewViewProvider {
 							await this.updateGlobalState("ollamaModelId", ollamaModelId)
 							await this.updateGlobalState("ollamaBaseUrl", ollamaBaseUrl)
 							await this.updateGlobalState("anthropicBaseUrl", anthropicBaseUrl)
+							await this.storeSecret("geminiApiKey", geminiApiKey)
 							this.claudeDev?.updateApi(message.apiConfiguration)
 						}
 						await this.postStateToWebview()
@@ -667,6 +676,7 @@ export class ClaudeDevProvider implements vscode.WebviewViewProvider {
 			ollamaModelId,
 			ollamaBaseUrl,
 			anthropicBaseUrl,
+			geminiApiKey,
 			lastShownAnnouncementId,
 			customInstructions,
 			alwaysAllowReadOnly,
@@ -688,6 +698,7 @@ export class ClaudeDevProvider implements vscode.WebviewViewProvider {
 			this.getGlobalState("ollamaModelId") as Promise<string | undefined>,
 			this.getGlobalState("ollamaBaseUrl") as Promise<string | undefined>,
 			this.getGlobalState("anthropicBaseUrl") as Promise<string | undefined>,
+			this.getSecret("geminiApiKey") as Promise<string | undefined>,
 			this.getGlobalState("lastShownAnnouncementId") as Promise<string | undefined>,
 			this.getGlobalState("customInstructions") as Promise<string | undefined>,
 			this.getGlobalState("alwaysAllowReadOnly") as Promise<boolean | undefined>,
@@ -726,6 +737,7 @@ export class ClaudeDevProvider implements vscode.WebviewViewProvider {
 				ollamaModelId,
 				ollamaBaseUrl,
 				anthropicBaseUrl,
+				geminiApiKey,
 			},
 			lastShownAnnouncementId,
 			customInstructions,
@@ -804,6 +816,7 @@ export class ClaudeDevProvider implements vscode.WebviewViewProvider {
 			"awsSecretKey",
 			"awsSessionToken",
 			"openAiApiKey",
+			"geminiApiKey",
 		]
 		for (const key of secretKeys) {
 			await this.storeSecret(key, undefined)

src/shared/api.ts

@@ -1,4 +1,4 @@
-export type ApiProvider = "anthropic" | "openrouter" | "bedrock" | "vertex" | "openai" | "ollama"
+export type ApiProvider = "anthropic" | "openrouter" | "bedrock" | "vertex" | "openai" | "ollama" | "gemini"
 
 export interface ApiHandlerOptions {
 	apiModelId?: string
@@ -16,6 +16,7 @@ export interface ApiHandlerOptions {
 	openAiModelId?: string
 	ollamaModelId?: string
 	ollamaBaseUrl?: string
+	geminiApiKey?: string
 }
 
 export type ApiConfiguration = ApiHandlerOptions & {
@@ -305,3 +306,26 @@ export const openAiModelInfoSaneDefaults: ModelInfo = {
 	inputPrice: 0,
 	outputPrice: 0,
 }
+
+// Gemini
+// https://ai.google.dev/gemini-api/docs/models/gemini
+export type GeminiModelId = keyof typeof geminiModels
+export const geminiDefaultModelId: GeminiModelId = "gemini-1.5-flash-latest"
+export const geminiModels = {
+	"gemini-1.5-flash-latest": {
+		maxTokens: 8192,
+		contextWindow: 1_048_576,
+		supportsImages: true,
+		supportsPromptCache: false,
+		inputPrice: 0,
+		outputPrice: 0,
+	},
+	"gemini-1.5-pro-latest": {
+		maxTokens: 8192,
+		contextWindow: 2_097_152,
+		supportsImages: true,
+		supportsPromptCache: false,
+		inputPrice: 0,
+		outputPrice: 0,
+	},
+} as const satisfies Record<string, ModelInfo>

src/utils/gemini-format.ts

@@ -0,0 +1,137 @@
+import { Anthropic } from "@anthropic-ai/sdk"
+import { Content, EnhancedGenerateContentResponse, FunctionDeclaration, Part, SchemaType } from "@google/generative-ai"
+
+export function convertAnthropicContentToGemini(
+	content:
+		| string
+		| Array<
+				| Anthropic.Messages.TextBlockParam
+				| Anthropic.Messages.ImageBlockParam
+				| Anthropic.Messages.ToolUseBlockParam
+				| Anthropic.Messages.ToolResultBlockParam
+		  >
+): Part[] {
+	if (typeof content === "string") {
+		return [{ text: content }]
+	}
+	return content.map((block) => {
+		switch (block.type) {
+			case "text":
+				return { text: block.text }
+			case "image":
+				if (block.source.type !== "base64") {
+					throw new Error("Unsupported image source type")
+				}
+				return {
+					inlineData: {
+						data: block.source.data,
+						mimeType: block.source.media_type,
+					},
+				}
+			case "tool_use":
+				return {
+					functionCall: {
+						name: block.name,
+						args: block.input,
+					},
+				} as Part
+			case "tool_result":
+				return {
+					functionResponse: {
+						name: block.tool_use_id,
+						response: {
+							content: block.content,
+						},
+					},
+				}
+			default:
+				throw new Error(`Unsupported content block type: ${(block as any).type}`)
+		}
+	})
+}
+
+export function convertAnthropicMessageToGemini(message: Anthropic.Messages.MessageParam): Content {
+	return {
+		role: message.role === "assistant" ? "model" : message.role,
+		parts: convertAnthropicContentToGemini(message.content),
+	}
+}
+
+export function convertAnthropicToolToGemini(tool: Anthropic.Messages.Tool): FunctionDeclaration {
+	return {
+		name: tool.name,
+		description: tool.description || "",
+		parameters: {
+			type: SchemaType.OBJECT,
+			properties: Object.fromEntries(
+				Object.entries(tool.input_schema.properties || {}).map(([key, value]) => [
+					key,
+					{
+						type: (value as any).type.toUpperCase(),
+						description: (value as any).description || "",
+					},
+				])
+			),
+			required: (tool.input_schema.required as string[]) || [],
+		},
+	}
+}
+
+export function convertGeminiResponseToAnthropic(
+	response: EnhancedGenerateContentResponse
+): Anthropic.Messages.Message {
+	const content: Anthropic.Messages.ContentBlock[] = []
+
+	// Add the main text response
+	const text = response.text()
+	if (text) {
+		content.push({ type: "text", text })
+	}
+
+	// Add function calls as tool_use blocks
+	const functionCalls = response.functionCalls()
+	if (functionCalls) {
+		functionCalls.forEach((call, index) => {
+			content.push({
+				type: "tool_use",
+				id: `tool_${index}`,
+				name: call.name,
+				input: call.args,
+			})
+		})
+	}
+
+	// Determine stop reason
+	let stop_reason: Anthropic.Messages.Message["stop_reason"] = null
+	const finishReason = response.candidates?.[0]?.finishReason
+	if (finishReason) {
+		switch (finishReason) {
+			case "STOP":
+				stop_reason = "end_turn"
+				break
+			case "MAX_TOKENS":
+				stop_reason = "max_tokens"
+				break
+			case "SAFETY":
+			case "RECITATION":
+			case "OTHER":
+				stop_reason = "stop_sequence"
+				break
+			// Add more cases if needed
+		}
+	}
+
+	return {
+		id: `msg_${Date.now()}`, // Generate a unique ID
+		type: "message",
+		role: "assistant",
+		content,
+		model: "",
+		stop_reason,
+		stop_sequence: null, // Gemini doesn't provide this information
+		usage: {
+			input_tokens: response.usageMetadata?.promptTokenCount ?? 0,
+			output_tokens: response.usageMetadata?.candidatesTokenCount ?? 0,
+		},
+	}
+}