Create analyze.ts

solutions/node-hello-world/api/analyze.ts

@@ -0,0 +1,73 @@
+// File: api/analyze.ts
+import { GoogleGenAI, Type } from "@google/genai";
+
+// This is a generic handler for Vercel's Edge Functions.
+export const config = {
+  runtime: 'edge',
+};
+
+export default async function handler(request: Request) {
+  if (request.method !== 'POST') {
+    return new Response('Method Not Allowed', { status: 405 });
+  }
+
+  try {
+    const body = await request.json();
+    const { fileName, genre, language, referenceFile } = body;
+
+    const apiKey = process.env.API_KEY;
+    if (!apiKey) {
+      throw new Error("API_KEY is not configured on the server.");
+    }
+
+    const ai = new GoogleGenAI({ apiKey });
+
+    // The prompt and schema are moved here, to the secure backend.
+    const referencePrompt = referenceFile ? `
+        **REFERENCE TRACK COMPARISON:**
+        The user has also provided a reference track for comparison.
+        - Reference Filename: ${referenceFile.name}
+        Your analysis MUST include a direct comparison to this reference track in the 'referenceAnalysis' object.
+    ` : `
+        No reference track was provided. The 'referenceAnalysis' field in the JSON response should be omitted.
+    `;
+
+    const prompt = `
+        You are an expert AI audio engineer from VDSS Audio Academy. Your task is to provide a highly technical and professional analysis of an audio file.
+        - Filename: ${fileName}
+        - Genre: ${genre}
+        - Language for response: ${language}
+        ${referencePrompt}
+        **CRITICAL METRIC ACCURACY:** You must follow these rules without exception.
+        - **IF MASTERED:** 'maxTruePeak' MUST be between -1.0 and -0.1 dBTP. 'integratedLufs' should be high (e.g., -6 to -10 LUFS).
+        - **IF PRE-MASTER:** 'maxTruePeak' MUST be between -6.0 and -3.0 dBTP for headroom. 'integratedLufs' should be lower (e.g., -16 to -23 LUFS).
+        **LANGUAGE REQUIREMENT:** All text in the response JSON MUST be in the requested language: '${language}'.
+        Generate the response according to the provided JSON schema.
+    `;
+
+    // NOTE: The full schema from your original file would be placed here.
+    // It is omitted for brevity but is required for the function to work.
+    const analysisSchema = { type: Type.OBJECT, properties: { /* ... PASTE FULL SCHEMA HERE ... */ } };
+
+    const response = await ai.models.generateContent({
+        model: "gemini-2.5-flash",
+        contents: prompt,
+        config: {
+            responseMimeType: "application/json",
+            responseSchema: analysisSchema,
+        },
+    });
+
+    return new Response(response.text, {
+      status: 200,
+      headers: { 'Content-Type': 'application/json' },
+    });
+
+  } catch (error) {
+    console.error("Error in serverless function:", error);
+    return new Response(JSON.stringify({ error: "Failed to get analysis from AI." }), {
+      status: 500,
+      headers: { 'Content-Type': 'application/json' },
+    });
+  }
+}