user friendly errors

Author: soham-chitnis10

api/api.py

@@ -1,5 +1,6 @@
 import os
 import json
+import logging
 
 from dotenv import dotenv_values
 import torch
@@ -27,6 +28,8 @@
 else:
     TOKEN = dotenv_values()["HF_TOKEN"]
 
+logger = logging.getLogger(__name__)
+
 ml_models = {}
 
 
@@ -81,7 +84,7 @@ async def generate(
     """
     Generate an output stream based on a prompt, using a LLM (Llama 3.2 1B Instruct).
     """
-    print(safenudge)
+    logger.info("generate called: safenudge=%s", safenudge)
     if not safenudge:
         data = generate_output_stream(
             init_prompt=init_prompt,

client/src/api/stream.js

@@ -38,8 +38,15 @@ export async function* streamEndpoint(url, params, signal) {
   });
 
   if (!res.ok) {
-    const text = await res.text().catch(() => "");
-    throw new Error(`Request to ${url} failed: ${res.status} ${text}`);
+    const friendlyMessages = {
+      429: "Too many requests — please wait a moment before trying again.",
+      422: "Invalid request parameters.",
+      500: "Server error. Please try again.",
+      502: "Server is unavailable. Please try again shortly.",
+      503: "Server is unavailable. Please try again shortly.",
+    };
+    const msg = friendlyMessages[res.status] ?? `Request failed (${res.status}).`;
+    throw new Error(msg);
   }
   if (!res.body) {
     throw new Error("Streaming response body unavailable");