feat(ollama): move AI panel to shui with server and model set in .env;

backend/src/controllers/ollama.controller.js

@@ -1,5 +1,10 @@
-const OLLAMA_BASE_URL = "http://jin.neu.edu:11434";
-const OLLAMA_MODEL = "qwen3.6:27b";
+// Ollama server + model come from .env, so switching needs no code change:
+//   OLLAMA_BASE_URL=http://shui.neu.edu:11434
+//   OLLAMA_MODEL=qwen3.8-careful:latest
+const OLLAMA_BASE_URL = (
+  process.env.OLLAMA_BASE_URL || "http://shui.neu.edu:11434"
+).replace(/\/+$/, "");
+const OLLAMA_MODEL = process.env.OLLAMA_MODEL || "qwen3.8-careful:latest";
 
 const proxyChat = async (req, res) => {
   console.log("🟣 [Ollama] proxyChat hit — model:", OLLAMA_MODEL);

src/components/User/Dashboard/DatasetOrganizer/LLMPanel.tsx

@@ -66,7 +66,7 @@ const llmProviders: Record<string, LLMProvider> = {
     name: "Ollama (NeuroJSON Server)",
     baseUrl: "",
     models: [
-      { id: "qwen3.6:27b", name: "Qwen 3.6 27B" },
+      { id: "qwen3.8-careful:latest", name: "Qwen 3.8 Careful" },
       // { id: "qwen3-coder-next:latest", name: "Qwen 3 Coder Next" },
       // { id: "qwen3-coder-careful:latest", name: "Qwen 3 Coder Careful" },
       // { id: "qwen3.5:9b", name: "Qwen 3.5 9B" },
@@ -1461,7 +1461,7 @@ const LLMPanel: React.FC<LLMPanelProps> = ({
 
           {provider === "ollama" && (
             <Typography variant="caption" color="text.secondary" sx={{ mt: -1, mb: 2, display: "block" }}>
-              Using qwen3.6:27b on NeuroJSON server
+              Using Qwen 3.8 Careful on NeuroJSON server
             </Typography>
           )}
 

src/services/ollama.service.ts

@@ -15,6 +15,9 @@ export const OllamaService = {
     // const temperature = getQwenTemperature(model);
     const response = await fetch(`${API_URL}/ollama/bidsify`, {
       method: "POST",
+      // Send the login cookie: /ollama/bidsify requires login, and locally the
+      // API (localhost:5000) is a different origin from the page.
+      credentials: "include",
       headers: { "Content-Type": "application/json" },
       body: JSON.stringify({
         model,
@@ -29,19 +32,23 @@ export const OllamaService = {
     const data = await response.json();
 
     if (!response.ok) {
-      throw new Error(data.error || "Failed to call Ollama");
+      throw new Error(data.error || data.message || "Failed to call Ollama");
     }
 
     return data;
   },
 
   getTags: async (): Promise<any> => {
-    const response = await fetch(`${API_URL}/ollama/tags`);
+    const response = await fetch(`${API_URL}/ollama/tags`, {
+      credentials: "include",
+    });
 
     const data = await response.json();
 
     if (!response.ok) {
-      throw new Error(data.error || "Failed to fetch Ollama models");
+      throw new Error(
+        data.error || data.message || "Failed to fetch Ollama models"
+      );
     }
 
     return data;