@stackmemoryai/stackmemory/dist/src/core/extensions/provider-adapter.js

Project-scoped memory for AI coding tools. Durable context across sessions with MCP integration, frames, smart retrieval, Claude Code skills, and automatic hooks.

import { fileURLToPath as __fileURLToPath } from 'url';
import { dirname as __pathDirname } from 'path';
const __filename = __fileURLToPath(import.meta.url);
const __dirname = __pathDirname(__filename);
class ClaudeAdapter {
  id = "claude";
  name = "Anthropic Claude";
  version = "1.0.0";
  extensions = {
    claude: {
      extendedThinking: { enabled: true, budgetTokens: 1e4 },
      xmlOutput: { enabled: true },
      computerUse: { enabled: true },
      documentSupport: { enabled: true, maxPages: 100 }
    }
  };
  apiKey;
  baseUrl;
  constructor(config) {
    this.apiKey = config.apiKey;
    this.baseUrl = config.baseUrl || "https://api.anthropic.com";
  }
  supportsExtension(extension) {
    return extension === "claude";
  }
  async *stream(messages, options) {
    const body = this.buildRequestBody(messages, options);
    const response = await fetch(`${this.baseUrl}/v1/messages`, {
      method: "POST",
      headers: {
        "Content-Type": "application/json",
        "x-api-key": this.apiKey,
        "anthropic-version": "2023-06-01",
        "anthropic-beta": this.getBetaFlags(options.extensions)
      },
      body: JSON.stringify({ ...body, stream: true })
    });
    if (!response.ok) {
      yield {
        type: "error",
        error: {
          type: "api_error",
          message: `Claude API error: ${response.status} ${response.statusText}`
        }
      };
      return;
    }
    const reader = response.body?.getReader();
    if (!reader) {
      yield {
        type: "error",
        error: { type: "stream_error", message: "No response body" }
      };
      return;
    }
    const decoder = new TextDecoder();
    let buffer = "";
    while (true) {
      const { done, value } = await reader.read();
      if (done) break;
      buffer += decoder.decode(value, { stream: true });
      const lines = buffer.split("\n");
      buffer = lines.pop() || "";
      for (const line of lines) {
        if (line.startsWith("data: ")) {
          const data = line.slice(6);
          if (data === "[DONE]") return;
          try {
            const event = JSON.parse(data);
            yield this.normalizeEvent(event);
          } catch {
          }
        }
      }
    }
  }
  async complete(messages, options) {
    const body = this.buildRequestBody(messages, options);
    const response = await fetch(`${this.baseUrl}/v1/messages`, {
      method: "POST",
      headers: {
        "Content-Type": "application/json",
        "x-api-key": this.apiKey,
        "anthropic-version": "2023-06-01",
        "anthropic-beta": this.getBetaFlags(options.extensions)
      },
      body: JSON.stringify(body)
    });
    if (!response.ok) {
      throw new Error(
        `Claude API error: ${response.status} ${response.statusText}`
      );
    }
    const data = await response.json();
    return {
      content: data.content,
      usage: {
        inputTokens: data.usage?.input_tokens ?? 0,
        outputTokens: data.usage?.output_tokens ?? 0
      },
      stopReason: data.stop_reason ?? "end_turn"
    };
  }
  async validateConnection() {
    try {
      const response = await fetch(`${this.baseUrl}/v1/messages`, {
        method: "POST",
        headers: {
          "Content-Type": "application/json",
          "x-api-key": this.apiKey,
          "anthropic-version": "2023-06-01"
        },
        body: JSON.stringify({
          model: "claude-3-5-haiku-20241022",
          max_tokens: 1,
          messages: [{ role: "user", content: "hi" }]
        })
      });
      return response.ok || response.status === 400;
    } catch {
      return false;
    }
  }
  async listModels() {
    return [
      "claude-opus-4-20250514",
      "claude-sonnet-4-20250514",
      "claude-3-5-haiku-20241022",
      "claude-3-opus-20240229"
    ];
  }
  buildRequestBody(messages, options) {
    const body = {
      model: options.model,
      max_tokens: options.maxTokens,
      messages: messages.map((m) => ({
        role: m.role,
        content: m.content
      }))
    };
    if (options.temperature !== void 0) {
      body.temperature = options.temperature;
    }
    if (options.topP !== void 0) {
      body.top_p = options.topP;
    }
    if (options.stopSequences?.length) {
      body.stop_sequences = options.stopSequences;
    }
    if (options.system) {
      body.system = options.system;
    }
    if (options.tools?.length) {
      body.tools = options.tools.map((t) => ({
        name: t.name,
        description: t.description,
        input_schema: t.inputSchema
      }));
    }
    if (options.extensions?.extendedThinking?.enabled) {
      body.thinking = {
        type: "enabled",
        budget_tokens: options.extensions.extendedThinking.budgetTokens || 1e4
      };
      body.temperature = 1;
    }
    return body;
  }
  getBetaFlags(extensions) {
    const flags = [];
    if (extensions?.extendedThinking?.enabled) {
      flags.push("interleaved-thinking-2025-05-14");
    }
    if (extensions?.computerUse?.enabled) {
      flags.push("computer-use-2024-10-22");
    }
    if (extensions?.documentSupport?.enabled) {
      flags.push("pdfs-2024-09-25");
    }
    return flags.join(",");
  }
  normalizeEvent(event) {
    return event;
  }
}
class GPTAdapter {
  id = "gpt";
  name = "OpenAI GPT";
  version = "1.0.0";
  extensions = {
    gpt: {
      codeInterpreter: { enabled: true },
      browsing: { enabled: true },
      imageGeneration: {
        enabled: true,
        size: "1024x1024",
        quality: "standard"
      },
      functionCalling: { mode: "auto" },
      jsonMode: { enabled: true }
    }
  };
  apiKey;
  baseUrl;
  constructor(config) {
    this.apiKey = config.apiKey;
    this.baseUrl = config.baseUrl || "https://api.openai.com";
  }
  supportsExtension(extension) {
    return extension === "gpt";
  }
  async *stream(messages, options) {
    const openaiMessages = messages.map((m) => ({
      role: m.role,
      content: typeof m.content === "string" ? m.content : m.content.filter((c) => c.type === "text").map((c) => c.text).join("")
    }));
    const body = {
      model: options.model,
      max_tokens: options.maxTokens,
      messages: openaiMessages,
      stream: true
    };
    if (options.temperature !== void 0) {
      body.temperature = options.temperature;
    }
    if (options.topP !== void 0) {
      body.top_p = options.topP;
    }
    if (options.extensions?.jsonMode?.enabled) {
      body.response_format = { type: "json_object" };
    }
    if (options.tools?.length) {
      body.tools = options.tools.map((t) => ({
        type: "function",
        function: {
          name: t.name,
          description: t.description,
          parameters: t.inputSchema
        }
      }));
      if (options.extensions?.functionCalling?.mode) {
        body.tool_choice = options.extensions.functionCalling.mode;
      }
    }
    const response = await fetch(`${this.baseUrl}/v1/chat/completions`, {
      method: "POST",
      headers: {
        "Content-Type": "application/json",
        Authorization: `Bearer ${this.apiKey}`
      },
      body: JSON.stringify(body)
    });
    if (!response.ok) {
      yield {
        type: "error",
        error: {
          type: "api_error",
          message: `GPT API error: ${response.status} ${response.statusText}`
        }
      };
      return;
    }
    yield {
      type: "message_start",
      message: {
        id: `msg_${Date.now()}`,
        model: options.model,
        role: "assistant"
      }
    };
    const reader = response.body?.getReader();
    if (!reader) return;
    const decoder = new TextDecoder();
    let buffer = "";
    const blockIndex = 0;
    let blockStarted = false;
    while (true) {
      const { done, value } = await reader.read();
      if (done) break;
      buffer += decoder.decode(value, { stream: true });
      const lines = buffer.split("\n");
      buffer = lines.pop() || "";
      for (const line of lines) {
        if (line.startsWith("data: ")) {
          const data = line.slice(6);
          if (data === "[DONE]") {
            if (blockStarted) {
              yield { type: "content_block_stop", index: blockIndex };
            }
            yield { type: "message_stop" };
            return;
          }
          try {
            const parsed = JSON.parse(data);
            const delta = parsed.choices?.[0]?.delta;
            if (delta?.content && !blockStarted) {
              blockStarted = true;
              yield {
                type: "content_block_start",
                index: blockIndex,
                contentBlock: { type: "text", text: "" }
              };
            }
            if (delta?.content) {
              yield {
                type: "content_block_delta",
                index: blockIndex,
                delta: { type: "text_delta", text: delta.content }
              };
            }
          } catch {
          }
        }
      }
    }
  }
  async complete(messages, options) {
    const openaiMessages = messages.map((m) => ({
      role: m.role,
      content: typeof m.content === "string" ? m.content : m.content.filter((c) => c.type === "text").map((c) => c.text).join("")
    }));
    const body = {
      model: options.model,
      max_tokens: options.maxTokens,
      messages: openaiMessages
    };
    if (options.temperature !== void 0) {
      body.temperature = options.temperature;
    }
    const response = await fetch(`${this.baseUrl}/v1/chat/completions`, {
      method: "POST",
      headers: {
        "Content-Type": "application/json",
        Authorization: `Bearer ${this.apiKey}`
      },
      body: JSON.stringify(body)
    });
    if (!response.ok) {
      throw new Error(
        `GPT API error: ${response.status} ${response.statusText}`
      );
    }
    const data = await response.json();
    const choice = data.choices?.[0];
    return {
      content: [{ type: "text", text: choice?.message?.content ?? "" }],
      usage: {
        inputTokens: data.usage?.prompt_tokens ?? 0,
        outputTokens: data.usage?.completion_tokens ?? 0
      },
      stopReason: choice?.finish_reason ?? "stop"
    };
  }
  async validateConnection() {
    try {
      const response = await fetch(`${this.baseUrl}/v1/models`, {
        headers: { Authorization: `Bearer ${this.apiKey}` }
      });
      return response.ok;
    } catch {
      return false;
    }
  }
  async listModels() {
    return [
      "gpt-4o",
      "gpt-4o-mini",
      "gpt-4-turbo",
      "gpt-4",
      "gpt-3.5-turbo",
      "o1",
      "o1-mini",
      "o1-preview"
    ];
  }
}
class GeminiAdapter {
  id = "gemini";
  name = "Google Gemini";
  version = "1.0.0";
  extensions = {
    gemini: {
      grounding: { enabled: true, dynamicThreshold: 0.3 },
      multimodal: {
        videoEnabled: true,
        audioEnabled: true,
        maxVideoDurationSec: 60
      },
      codeExecution: { enabled: true }
    }
  };
  apiKey;
  baseUrl;
  constructor(config) {
    this.apiKey = config.apiKey;
    this.baseUrl = config.baseUrl || "https://generativelanguage.googleapis.com/v1beta";
  }
  supportsExtension(extension) {
    return extension === "gemini";
  }
  async *stream(messages, options) {
    const contents = messages.filter((m) => m.role !== "system").map((m) => ({
      role: m.role === "assistant" ? "model" : "user",
      parts: [
        {
          text: typeof m.content === "string" ? m.content : m.content.filter((c) => c.type === "text").map((c) => c.text).join("")
        }
      ]
    }));
    const body = {
      contents,
      generationConfig: {
        maxOutputTokens: options.maxTokens,
        temperature: options.temperature,
        topP: options.topP
      }
    };
    const systemMsg = messages.find((m) => m.role === "system");
    if (systemMsg || options.system) {
      body.systemInstruction = {
        parts: [
          {
            text: options.system || (typeof systemMsg?.content === "string" ? systemMsg.content : "")
          }
        ]
      };
    }
    if (options.extensions?.grounding?.enabled) {
      body.tools = [
        {
          googleSearchRetrieval: {
            dynamicRetrievalConfig: {
              mode: "MODE_DYNAMIC",
              dynamicThreshold: options.extensions.grounding.dynamicThreshold
            }
          }
        }
      ];
    }
    if (options.extensions?.codeExecution?.enabled) {
      body.tools = [
        ...body.tools || [],
        { codeExecution: {} }
      ];
    }
    if (options.extensions?.safetySettings) {
      body.safetySettings = options.extensions.safetySettings;
    }
    const url = `${this.baseUrl}/models/${options.model}:streamGenerateContent?key=${this.apiKey}&alt=sse`;
    const response = await fetch(url, {
      method: "POST",
      headers: { "Content-Type": "application/json" },
      body: JSON.stringify(body)
    });
    if (!response.ok) {
      yield {
        type: "error",
        error: {
          type: "api_error",
          message: `Gemini API error: ${response.status} ${response.statusText}`
        }
      };
      return;
    }
    yield {
      type: "message_start",
      message: {
        id: `msg_${Date.now()}`,
        model: options.model,
        role: "assistant"
      }
    };
    const reader = response.body?.getReader();
    if (!reader) return;
    const decoder = new TextDecoder();
    let buffer = "";
    const blockIndex = 0;
    let blockStarted = false;
    while (true) {
      const { done, value } = await reader.read();
      if (done) break;
      buffer += decoder.decode(value, { stream: true });
      const lines = buffer.split("\n");
      buffer = lines.pop() || "";
      for (const line of lines) {
        if (line.startsWith("data: ")) {
          try {
            const parsed = JSON.parse(line.slice(6));
            const text = parsed.candidates?.[0]?.content?.parts?.[0]?.text;
            if (text && !blockStarted) {
              blockStarted = true;
              yield {
                type: "content_block_start",
                index: blockIndex,
                contentBlock: { type: "text", text: "" }
              };
            }
            if (text) {
              yield {
                type: "content_block_delta",
                index: blockIndex,
                delta: { type: "text_delta", text }
              };
            }
          } catch {
          }
        }
      }
    }
    if (blockStarted) {
      yield { type: "content_block_stop", index: blockIndex };
    }
    yield { type: "message_stop" };
  }
  async complete(messages, options) {
    const contents = messages.filter((m) => m.role !== "system").map((m) => ({
      role: m.role === "assistant" ? "model" : "user",
      parts: [
        {
          text: typeof m.content === "string" ? m.content : m.content.filter((c) => c.type === "text").map((c) => c.text).join("")
        }
      ]
    }));
    const body = {
      contents,
      generationConfig: {
        maxOutputTokens: options.maxTokens,
        temperature: options.temperature
      }
    };
    const url = `${this.baseUrl}/models/${options.model}:generateContent?key=${this.apiKey}`;
    const response = await fetch(url, {
      method: "POST",
      headers: { "Content-Type": "application/json" },
      body: JSON.stringify(body)
    });
    if (!response.ok) {
      throw new Error(
        `Gemini API error: ${response.status} ${response.statusText}`
      );
    }
    const data = await response.json();
    const candidate = data.candidates?.[0];
    const text = candidate?.content?.parts?.[0]?.text ?? "";
    return {
      content: [{ type: "text", text }],
      usage: {
        inputTokens: data.usageMetadata?.promptTokenCount ?? 0,
        outputTokens: data.usageMetadata?.candidatesTokenCount ?? 0
      },
      stopReason: candidate?.finishReason ?? "STOP"
    };
  }
  async validateConnection() {
    try {
      const response = await fetch(`${this.baseUrl}/models?key=${this.apiKey}`);
      return response.ok;
    } catch {
      return false;
    }
  }
  async listModels() {
    return [
      "gemini-2.0-flash-exp",
      "gemini-1.5-pro",
      "gemini-1.5-flash",
      "gemini-1.5-flash-8b",
      "gemini-1.0-pro"
    ];
  }
}
function createProvider(id, config) {
  switch (id) {
    case "claude":
      return new ClaudeAdapter(config);
    case "gpt":
      return new GPTAdapter(config);
    case "gemini":
      return new GeminiAdapter(config);
    default:
      throw new Error(`Unknown provider: ${id}`);
  }
}
class ProviderRegistry {
  providers = /* @__PURE__ */ new Map();
  register(adapter) {
    this.providers.set(adapter.id, adapter);
  }
  get(id) {
    return this.providers.get(id);
  }
  list() {
    return Array.from(this.providers.values());
  }
  has(id) {
    return this.providers.has(id);
  }
  /**
   * Find providers that support a specific extension
   */
  findByExtension(extension) {
    return this.list().filter((p) => p.supportsExtension(extension));
  }
}
const providerRegistry = new ProviderRegistry();
export {
  ClaudeAdapter,
  GPTAdapter,
  GeminiAdapter,
  ProviderRegistry,
  createProvider,
  providerRegistry
};
//# sourceMappingURL=provider-adapter.js.map