agenticSystem/mcp-server/tools/media/analyze_image.js

import { z } from "zod";
import axios from "axios";
import fs from "fs";
import path from "path";
import { withAuth } from "../../auth/index.js";
import { handleToolError } from "../helpers/errorHandler.js";
import { withAuthParams } from "../helpers/authSchema.js";

const GEMINI_ENDPOINT = "https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash:generateContent";
const CHAT_UPLOADS_DIR = "/opt/acai/chat-uploads";
const DEFAULT_PROMPT = "Describe esta imagen detalladamente, mencionando elementos visuales, texto, layout y proposito aparente.";

/**
 * Detecta el mime type a partir de la extension del fichero o del primer byte (magic number).
 */
function detectMimeType(filename, buffer) {
    const ext = (filename || "").toLowerCase().split('.').pop();
    const byExt = {
        jpg: "image/jpeg",
        jpeg: "image/jpeg",
        png: "image/png",
        webp: "image/webp",
        gif: "image/gif",
        bmp: "image/bmp",
        heic: "image/heic",
        heif: "image/heif",
    };
    if (byExt[ext]) return byExt[ext];

    // Magic numbers fallback
    if (buffer && buffer.length >= 4) {
        if (buffer[0] === 0xff && buffer[1] === 0xd8 && buffer[2] === 0xff) return "image/jpeg";
        if (buffer[0] === 0x89 && buffer[1] === 0x50 && buffer[2] === 0x4e && buffer[3] === 0x47) return "image/png";
        if (buffer[0] === 0x47 && buffer[1] === 0x49 && buffer[2] === 0x46) return "image/gif";
        if (buffer.length >= 12 && buffer.slice(8, 12).toString() === "WEBP") return "image/webp";
    }
    return "image/jpeg";
}

/**
 * Resuelve una URL de chat-preview a una ruta local segura dentro de CHAT_UPLOADS_DIR.
 * Acepta `/api/chat-preview?file=xxx` o variantes con host.
 */
function resolveChatPreviewPath(imageUrl) {
    let qs;
    try {
        // Permite tanto absolutas como relativas
        const u = imageUrl.startsWith("http")
            ? new URL(imageUrl)
            : new URL(imageUrl, "http://placeholder.local");
        if (!u.pathname.startsWith("/api/chat-preview")) return null;
        qs = u.searchParams;
    } catch {
        return null;
    }

    const fileParam = qs.get("file");
    if (!fileParam) return null;

    // Sanitizar: evitar traversal — solo nombre base permitido
    const safeName = path.basename(fileParam);
    if (!safeName || safeName === "." || safeName === "..") return null;

    return path.join(CHAT_UPLOADS_DIR, safeName);
}

/**
 * Carga la imagen como { mimeType, base64 } desde URL publica o chat-preview local.
 */
async function loadImage(imageUrl) {
    // Caso 1: chat-preview local
    const localPath = resolveChatPreviewPath(imageUrl);
    if (localPath) {
        if (!fs.existsSync(localPath)) {
            throw new Error(`Local chat upload not found: ${path.basename(localPath)}`);
        }
        const buffer = fs.readFileSync(localPath);
        return {
            mimeType: detectMimeType(localPath, buffer),
            base64: buffer.toString("base64"),
        };
    }

    // Caso 2: URL publica http(s)
    if (imageUrl.startsWith("http://") || imageUrl.startsWith("https://")) {
        const response = await axios.get(imageUrl, {
            responseType: "arraybuffer",
            timeout: 30000,
            maxContentLength: 20 * 1024 * 1024, // 20MB max
        });
        const buffer = Buffer.from(response.data, "binary");
        const headerMime = response.headers?.["content-type"]?.split(";")[0]?.trim();
        const mimeType = headerMime && headerMime.startsWith("image/")
            ? headerMime
            : detectMimeType(imageUrl.split("?")[0], buffer);
        return {
            mimeType,
            base64: buffer.toString("base64"),
        };
    }

    throw new Error("Unsupported image_url. Use http(s):// or /api/chat-preview?file=...");
}

export function registerAnalyzeImageTool(server) {
    server.tool(
        "analyze_image",
        "Analiza una imagen usando Gemini Vision. Util cuando el usuario adjunta una imagen, despues de un screenshot de Playwright, o para describir cualquier imagen accesible via URL. Devuelve descripcion text del contenido visual.",
        withAuthParams({
            image_url: z.string().describe("URL de la imagen. Acepta URL publica http(s):// o ruta relativa /api/chat-preview?file=..."),
            prompt: z.string().optional().describe("Que quieres saber de la imagen. Default: descripcion detallada."),
        }),
        { readOnlyHint: true, destructiveHint: false },
        withAuth(async ({ image_url, prompt }) => {
            try {
                const apiKey = process.env.NANO_BANANA_API_KEY;
                if (!apiKey) {
                    return {
                        content: [{
                            type: "text",
                            text: JSON.stringify({
                                success: false,
                                error: "NANO_BANANA_API_KEY no esta configurada en el entorno del MCP server.",
                            }, null, 2),
                        }],
                        isError: true,
                    };
                }

                // 1) Cargar imagen (local o remota) -> base64 + mime
                let image;
                try {
                    image = await loadImage(image_url);
                } catch (loadErr) {
                    return {
                        content: [{
                            type: "text",
                            text: JSON.stringify({
                                success: false,
                                error: `No se pudo cargar la imagen: ${loadErr.message}`,
                            }, null, 2),
                        }],
                        isError: true,
                    };
                }

                // 2) Llamar a Gemini Vision
                const finalPrompt = (prompt && prompt.trim()) || DEFAULT_PROMPT;
                const payload = {
                    contents: [{
                        parts: [
                            { inline_data: { mime_type: image.mimeType, data: image.base64 } },
                            { text: finalPrompt },
                        ],
                    }],
                };

                const geminiResp = await axios.post(GEMINI_ENDPOINT, payload, {
                    headers: {
                        "x-goog-api-key": apiKey,
                        "Content-Type": "application/json",
                    },
                    timeout: 60000,
                    maxBodyLength: 30 * 1024 * 1024,
                });

                const description = geminiResp.data?.candidates?.[0]?.content?.parts?.[0]?.text;
                if (!description) {
                    return {
                        content: [{
                            type: "text",
                            text: JSON.stringify({
                                success: false,
                                error: "Gemini no devolvio descripcion.",
                                raw: geminiResp.data,
                            }, null, 2),
                        }],
                        isError: true,
                    };
                }

                return {
                    content: [{
                        type: "text",
                        text: description,
                    }],
                };
            } catch (error) {
                // Mejorar error si es respuesta de Gemini
                if (error.response?.data) {
                    return handleToolError(
                        new Error(`Gemini API error: ${JSON.stringify(error.response.data).slice(0, 500)}`),
                        "analyze_image",
                        { image_url, status: error.response.status }
                    );
                }
                return handleToolError(error, "analyze_image", { image_url });
            }
        })
    );
}