fix: added

server.js

@@ -6,6 +6,7 @@ import {
   Qwen3_5ForConditionalGeneration,
   RawImage,
 } from "@huggingface/transformers";
+import crypto from "crypto";
 
 const app = express();
 const upload = multer({ storage: multer.memoryStorage() });
@@ -16,8 +17,24 @@ let model = null;
 let processor = null;
 let inferenceQueue = Promise.resolve();
 
+function log(level, event, meta = {}) {
+  const payload = {
+    ts: new Date().toISOString(),
+    level,
+    event,
+    ...meta,
+  };
+  const line = JSON.stringify(payload);
+  if (level === "error") {
+    console.error(line);
+  } else {
+    console.log(line);
+  }
+}
+
 async function loadModel() {
-  console.log(`Loading model ${MODEL_ID}...`);
+  const start = Date.now();
+  log("info", "model_load_started", { model_id: MODEL_ID });
   processor = await AutoProcessor.from_pretrained(MODEL_ID);
   model = await Qwen3_5ForConditionalGeneration.from_pretrained(MODEL_ID, {
     dtype: {
@@ -26,7 +43,10 @@ async function loadModel() {
       decoder_model_merged: "q4",
     },
   });
-  console.log("Model loaded successfully.");
+  log("info", "model_load_completed", {
+    model_id: MODEL_ID,
+    duration_ms: Date.now() - start,
+  });
 }
 
 async function runInference(imageBuffer, prompt, maxTokens) {
@@ -144,57 +164,133 @@ const swaggerDoc = {
 
 app.use("/docs", swaggerUi.serve, swaggerUi.setup(swaggerDoc));
 
-app.get("/", (_req, res) => {
+app.use((req, res, next) => {
+  const requestId = crypto.randomUUID();
+  const start = Date.now();
+  req.requestId = requestId;
+  log("info", "request_started", {
+    request_id: requestId,
+    method: req.method,
+    path: req.originalUrl,
+    ip: req.ip,
+  });
+  res.on("finish", () => {
+    log("info", "request_finished", {
+      request_id: requestId,
+      method: req.method,
+      path: req.originalUrl,
+      status_code: res.statusCode,
+      duration_ms: Date.now() - start,
+    });
+  });
+  next();
+});
+
+app.get("/", (req, res) => {
+  log("info", "root_status", { request_id: req.requestId });
   res.json({ status: "ok", model: MODEL_ID });
 });
 
-app.get("/health", (_req, res) => {
+app.get("/health", (req, res) => {
+  log("info", "health_checked", {
+    request_id: req.requestId,
+    model_loaded: model !== null && processor !== null,
+  });
   res.json({ status: "healthy", model_loaded: model !== null });
 });
 
 app.post("/inference", upload.single("file"), async (req, res) => {
+  const prompt = req.body.prompt || "Describe this image in detail.";
+  const maxTokens = parseInt(req.body.max_tokens) || 256;
+  log("info", "inference_request_received", {
+    request_id: req.requestId,
+    has_file: Boolean(req.file),
+    file_size_bytes: req.file?.size ?? 0,
+    prompt_chars: prompt.length,
+    max_tokens: maxTokens,
+  });
+
   if (!model || !processor) {
+    log("error", "inference_model_unavailable", { request_id: req.requestId });
     return res.status(503).json({ detail: "Model not loaded yet." });
   }
   if (!req.file) {
+    log("error", "inference_validation_failed", {
+      request_id: req.requestId,
+      reason: "missing_file",
+    });
     return res.status(400).json({ detail: "No image file provided." });
   }
 
-  const prompt = req.body.prompt || "Describe this image in detail.";
-  const maxTokens = parseInt(req.body.max_tokens) || 256;
-
   try {
+    const start = Date.now();
     const response = await queueInference(req.file.buffer, prompt, maxTokens);
+    log("info", "inference_completed", {
+      request_id: req.requestId,
+      duration_ms: Date.now() - start,
+      response_chars: response?.length ?? 0,
+    });
     res.json({ response });
   } catch (err) {
-    console.error(err);
+    log("error", "inference_failed", {
+      request_id: req.requestId,
+      error: err.message,
+      stack: err.stack,
+    });
     res.status(500).json({ detail: "Inference failed.", error: err.message });
   }
 });
 
 app.post("/inference/base64", express.urlencoded({ extended: true, limit: "50mb" }), async (req, res) => {
+  const prompt = req.body.prompt || "Describe this image in detail.";
+  const maxTokens = parseInt(req.body.max_tokens) || 256;
+  log("info", "inference_base64_request_received", {
+    request_id: req.requestId,
+    has_image_base64: Boolean(req.body.image_base64),
+    image_base64_chars: req.body.image_base64?.length ?? 0,
+    prompt_chars: prompt.length,
+    max_tokens: maxTokens,
+  });
+
   if (!model || !processor) {
+    log("error", "inference_base64_model_unavailable", { request_id: req.requestId });
     return res.status(503).json({ detail: "Model not loaded yet." });
   }
   if (!req.body.image_base64) {
+    log("error", "inference_base64_validation_failed", {
+      request_id: req.requestId,
+      reason: "missing_image_base64",
+    });
     return res.status(400).json({ detail: "No base64 image provided." });
   }
 
-  const prompt = req.body.prompt || "Describe this image in detail.";
-  const maxTokens = parseInt(req.body.max_tokens) || 256;
-
   try {
+    const start = Date.now();
     const imageBuffer = Buffer.from(req.body.image_base64, "base64");
     const response = await queueInference(imageBuffer, prompt, maxTokens);
+    log("info", "inference_base64_completed", {
+      request_id: req.requestId,
+      duration_ms: Date.now() - start,
+      image_bytes: imageBuffer.length,
+      response_chars: response?.length ?? 0,
+    });
     res.json({ response });
   } catch (err) {
-    console.error(err);
+    log("error", "inference_base64_failed", {
+      request_id: req.requestId,
+      error: err.message,
+      stack: err.stack,
+    });
     res.status(500).json({ detail: "Inference failed.", error: err.message });
   }
 });
 
 loadModel().then(() => {
   app.listen(PORT, "0.0.0.0", () => {
-    console.log(`Server running on http://0.0.0.0:${PORT}`);
+    log("info", "server_started", {
+      host: "0.0.0.0",
+      port: PORT,
+      model_id: MODEL_ID,
+    });
   });
 });