import { McpServer } from "@modelcontextprotocol/sdk/server/mcp.js";
import { z } from "zod";
import { searchProducts, getProductByBarcode } from "./product-search.js";
import { requestSampling, createRecipeSuggestionRequest, getResponseText } from "../sampling/sampling-service.js";
import { logger } from '../transport/transports.js';
import { registerCategoryTools } from './category-tools.js';
import { registerNutritionTools } from './nutrition-tools.js';
import { registerInsightsTools } from './insights-tools.js';
import { registerPriceTools } from './price-tools.js';
export * from './types.js';
export * from './helpers.js';
/**
* Flexible product lookup - accepts either barcode or product name.
* If barcode provided, fetches directly. Otherwise, searches by name
* and retrieves the first matching product's full details.
*/
async function findProduct(nameOrBarcode: string): Promise<any> {
if (!nameOrBarcode?.trim()) return null;
const query = nameOrBarcode.trim();
const isBarcode = /^\d+$/.test(query);
if (isBarcode) {
try {
const product = await getProductByBarcode(query);
if (product) return { product };
} catch (error) {
logger.error(`Barcode lookup failed: ${error}`);
}
}
try {
const results = await searchProducts(query, 1, 1);
const barcode = results?.products?.[0]?.barcode;
if (barcode) {
const product = await getProductByBarcode(barcode);
return { product };
}
} catch (error) {
logger.error(`Search failed: ${error}`);
}
return null;
}
// Schema definitions for core tools
const searchSchema = { query: z.string(), page: z.number().default(1), pageSize: z.number().default(10) };
const barcodeSchema = { barcode: z.string() };
const productSchema = { nameOrBarcode: z.string() };
const compareSchema = { nameOrBarcode1: z.string(), nameOrBarcode2: z.string() };
/**
* Register all MCP tools with the server
*/
export function registerTools(server: McpServer): void {
server.registerTool('searchProducts', {
description: 'Search products by name, brand, or category',
inputSchema: searchSchema
}, async ({ query, page, pageSize }) => {
try {
const results = await searchProducts(query, page ?? 1, pageSize ?? 10);
return { content: [{ type: 'text' as const, text: JSON.stringify(results) }] };
} catch (error: any) {
return { content: [{ type: 'text' as const, text: `Error: ${error.message}` }], isError: true };
}
});
server.registerTool('getProductByBarcode', {
description: 'Get product details by barcode (EAN/UPC)',
inputSchema: barcodeSchema
}, async ({ barcode }) => {
try {
const product = await getProductByBarcode(barcode);
return { content: [{ type: 'text' as const, text: JSON.stringify(product) }] };
} catch (error: any) {
return { content: [{ type: 'text' as const, text: `Error: ${error.message}` }], isError: true };
}
});
/**
* AI-powered analysis via MCP sampling.
* Sends product data to connected LLM for nutritional insights.
* Falls back to basic info if sampling unavailable.
*/
server.registerTool('analyzeProduct', {
description: 'Get AI nutritional analysis of a product',
inputSchema: productSchema
}, async ({ nameOrBarcode }) => {
if (!nameOrBarcode?.trim()) {
return { content: [{ type: 'text' as const, text: 'Provide a product name or barcode.' }], isError: true };
}
const productData = await findProduct(nameOrBarcode);
if (!productData?.product) {
return { content: [{ type: 'text' as const, text: `"${nameOrBarcode}" not found. Try searchProducts first.` }], isError: true };
}
const p = productData.product;
try {
const res = await requestSampling(server, {
messages: [{ role: "user" as const, content: { type: "text" as const, text: `Analyze:\n\n${JSON.stringify(p, null, 2)}` } }],
systemPrompt: "Nutritional expert. Provide: 1) Overview 2) Nutrition 3) Ingredients 4) Allergens 5) Health 6) Recommendations",
modelPreferences: { hints: [{ name: "claude-3" }], intelligencePriority: 0.9 },
includeContext: "thisServer" as const,
temperature: 0.3,
maxTokens: 1500
});
const header = `# ${p.product_name || "Product"} (${p.brands || "Unknown"})\n`;
return { content: [{ type: 'text' as const, text: header + getResponseText(res) }] };
} catch {
return { content: [{ type: 'text' as const, text: `# ${p.product_name}\nBrand: ${p.brands}\nNutri-Score: ${p.nutriscore_grade}\nIngredients: ${p.ingredients_text}` }] };
}
});
/**
* AI-powered comparison via MCP sampling.
* Fetches both products in parallel, then requests comparative analysis.
*/
server.registerTool('compareProducts', {
description: 'Compare two products using AI',
inputSchema: compareSchema
}, async ({ nameOrBarcode1, nameOrBarcode2 }) => {
if (!nameOrBarcode1?.trim() || !nameOrBarcode2?.trim()) {
return { content: [{ type: 'text' as const, text: 'Provide both products.' }], isError: true };
}
const [d1, d2] = await Promise.all([findProduct(nameOrBarcode1), findProduct(nameOrBarcode2)]);
if (!d1?.product) return { content: [{ type: 'text' as const, text: `"${nameOrBarcode1}" not found.` }], isError: true };
if (!d2?.product) return { content: [{ type: 'text' as const, text: `"${nameOrBarcode2}" not found.` }], isError: true };
try {
const res = await requestSampling(server, {
messages: [{ role: "user" as const, content: { type: "text" as const, text: `Compare:\n\n1:\n${JSON.stringify(d1.product)}\n\n2:\n${JSON.stringify(d2.product)}` } }],
systemPrompt: "Compare: 1) Overview 2) Nutrition 3) Ingredients 4) Health 5) Recommendation",
modelPreferences: { hints: [{ name: "claude-3" }], intelligencePriority: 0.9 },
includeContext: "thisServer" as const,
temperature: 0.2,
maxTokens: 2000
});
return { content: [{ type: 'text' as const, text: `# ${d1.product.product_name} vs ${d2.product.product_name}\n\n${getResponseText(res)}` }] };
} catch {
return { content: [{ type: 'text' as const, text: `# ${d1.product.product_name} vs ${d2.product.product_name}\nNutri-Score: ${d1.product.nutriscore_grade} vs ${d2.product.nutriscore_grade}` }] };
}
});
/**
* AI-powered recipe generation via MCP sampling.
*/
server.registerTool('suggestRecipes', {
description: 'Get AI recipe suggestions using a product',
inputSchema: productSchema
}, async ({ nameOrBarcode }) => {
if (!nameOrBarcode?.trim()) {
return { content: [{ type: 'text' as const, text: 'Provide a product.' }], isError: true };
}
const productData = await findProduct(nameOrBarcode);
if (!productData?.product) {
return { content: [{ type: 'text' as const, text: `"${nameOrBarcode}" not found.` }], isError: true };
}
try {
const res = await requestSampling(server, createRecipeSuggestionRequest(productData.product));
return { content: [{ type: 'text' as const, text: `# Recipes using ${productData.product.product_name}\n\n${getResponseText(res)}` }] };
} catch (e: any) {
return { content: [{ type: 'text' as const, text: `Recipe generation failed: ${e.message}` }], isError: true };
}
});
logger.info("Core OpenFoodFacts tools registered");
registerCategoryTools(server);
registerNutritionTools(server);
registerInsightsTools(server);
registerPriceTools(server);
registerPriceTools(server);
logger.info("All OpenFoodFacts MCP tools registered successfully");
}
