Add React frontend and Sinbad2IA LLM integration.

Introduce a full Vite/React UI for exams, auth, materials, images, generation, and export.
Adapt backend for Sinbad2IA chat API, bcrypt passwords, CORS on port 5173, and schema migrations.

backend/app/services/llm.py

@@ -5,44 +5,69 @@ from app.core.errors import LLMUnavailableError
 
 
 class LLMClient:
+    """Cliente para el API de chat de Sinbad2IA (Ollama-compatible en UJA)."""
+
     def __init__(self, settings: Settings) -> None:
         self.settings = settings
 
-    async def generate(self, prompt: str) -> str:
-        if not self.settings.llm_api_key:
-            raise LLMUnavailableError("LLM_API_KEY is not configured")
+    def _chat_url(self) -> str:
+        base = self.settings.llm_base_url.rstrip("/")
+        if base.endswith("/api/chat"):
+            return base
+        return f"{base}/api/chat"
 
-        url = f"{self.settings.llm_base_url.rstrip('/')}/chat/completions"
+    async def generate(self, prompt: str) -> str:
         payload = {
             "model": self.settings.llm_model,
             "messages": [
                 {
-                    "role": "system",
-                    "content": "You generate safe, valid JSON exam questions for Moodle imports.",
+                    "role": "user",
+                    "content": (
+                        "Genera preguntas de examen en formato JSON válido para importar en Moodle. "
+                        "Responde únicamente con el JSON solicitado, sin texto adicional ni bloques markdown.\n\n"
+                        f"{prompt}"
+                    ),
                 },
-                {"role": "user", "content": prompt},
             ],
-            "temperature": 0.2,
-            "response_format": {"type": "json_object"},
-        }
-        headers = {
-            "Authorization": f"Bearer {self.settings.llm_api_key}",
-            "Content-Type": "application/json",
+            "stream": False,
         }
 
+        headers = {"Content-Type": "application/json"}
+        if self.settings.llm_api_key:
+            headers["Authorization"] = f"Bearer {self.settings.llm_api_key}"
+
         try:
             async with httpx.AsyncClient(timeout=self.settings.llm_timeout_seconds) as client:
-                response = await client.post(url, json=payload, headers=headers)
+                response = await client.post(self._chat_url(), json=payload, headers=headers)
                 response.raise_for_status()
         except httpx.HTTPError as exc:
             raise LLMUnavailableError("LLM request failed") from exc
 
-        data = response.json()
-        try:
-            content = data["choices"][0]["message"]["content"]
-        except (KeyError, IndexError, TypeError) as exc:
-            raise LLMUnavailableError("LLM response did not include message content") from exc
-
-        if not isinstance(content, str) or not content.strip():
+        content = _extract_assistant_content(response.json())
+        if not content.strip():
             raise LLMUnavailableError("LLM returned empty content")
         return content
+
+
+def _extract_assistant_content(data: object) -> str:
+    """Soporta respuesta Sinbad2IA/Ollama (`message.content`) y OpenAI (`choices`)."""
+    if not isinstance(data, dict):
+        raise LLMUnavailableError("LLM response is not a JSON object")
+
+    message = data.get("message")
+    if isinstance(message, dict):
+        content = message.get("content")
+        if isinstance(content, str) and content.strip():
+            return content
+
+    choices = data.get("choices")
+    if isinstance(choices, list) and choices:
+        first = choices[0]
+        if isinstance(first, dict):
+            msg = first.get("message")
+            if isinstance(msg, dict):
+                content = msg.get("content")
+                if isinstance(content, str) and content.strip():
+                    return content
+
+    raise LLMUnavailableError("LLM response did not include message content")