feat(llm): add thinking parameter support for Ollama API calls

Add optional thinking parameter to the call_ollama function and pass it from the request. Also enhance timezone handling in prompt generation to support configurable timezone preferences.

backend/llm.py

@@ -54,31 +54,36 @@ def _extract_message(response) -> tuple[str, str]:
     return content, thinking
 
 
-async def call_ollama(prompt: str, *, tag: str = "default", temperature: float = 0.7) -> dict:
+async def call_ollama(prompt: str, *, tag: str = "default", temperature: float = 0.7, thinking: str = None) -> dict:
     """
     调用 Ollama API 并返回 content 和 thinking。
     """
     start = time.perf_counter()
     start_dt = datetime.now()
     logger.info(
-        "[LLM][%s] request model=%s host=%s prompt_chars=%d temp=%.2f",
+        "[LLM][%s] request model=%s host=%s prompt_chars=%d temp=%.2f thinking=%s",
         tag,
         OLLAMA_MODEL,
         OLLAMA_HOST,
         len(prompt),
         temperature,
+        thinking,
     )
 
     try:
-        response = await client.chat(
-            model=OLLAMA_MODEL,
-            messages=[{'role': 'user', 'content': prompt}],
-            stream=False,
-            options={
+        kwargs = {
+            "model": OLLAMA_MODEL,
+            "messages": [{'role': 'user', 'content': prompt}],
+            "stream": False,
+            "options": {
                 'temperature': temperature,
                 'repeat_penalty': 1.1,
             },
-        )
+        }
+        if thinking:
+            kwargs["thinking"] = thinking
+
+        response = await client.chat(**kwargs)
     except Exception:
         elapsed_ms = (time.perf_counter() - start) * 1000
         end_dt = datetime.now()