fix: render thinking channel regardless of tool_calls presence

Remove message.get('tool_calls') guard from thinking channel
rendering. Reasoning on assistant messages without tool_calls
(e.g. final answer after a tool chain) was silently dropped
from conversation history.

Ref: https://github.com/vllm-project/vllm/pull/45553

chat_template.jinja

@@ -231,7 +231,7 @@
     {#- Render reasoning/reasoning_content as thinking channel (tool-call turns only) -#}
     {%- set thinking_text = message.get('reasoning') or message.get('reasoning_content') -%}
     {%- set thinking_gate = (loop.index0 > ns_turn.last_user_idx) or preserve_thinking -%}
-    {%- if thinking_text and thinking_gate and message.get('tool_calls') -%}
+    {%- if thinking_text and thinking_gate -%}
         {{- '<|channel>thought\n' + thinking_text + '\n<channel|>' -}}
     {%- endif -%}