refactor(llm): optimize Gemini config profiles for extraction and creativity

Introduces `get_extraction_config` and `get_creative_config` to standardize Gemini API calls.

* Defines explicit config profiles with appropriate `temperature` and `thinking_level` for Gemini 3 Flash.
* Extraction tasks use minimal thinking and temp=0.0 to reduce latency and token usage.
* Creative tasks use low thinking, temp=0.4, and top_p=0.8 to balance naturalness and safety.
* Applies these helpers across products, routines, and skincare endpoints.
* Also updates default model to `gemini-3-flash-preview`.

backend/test_query.py

@@ -0,0 +1,25 @@
+from datetime import date, timedelta
+
+from sqlmodel import select
+
+from db import get_session
+from innercontext.models import Routine, RoutineStep
+
+
+def run():
+    session = next(get_session())
+    ref_date = date.today()
+    cutoff = ref_date - timedelta(days=7)
+
+    recent_usage = session.exec(
+        select(RoutineStep.product_id)
+        .join(Routine, Routine.id == RoutineStep.routine_id)
+        .where(Routine.routine_date >= cutoff)
+        .where(Routine.routine_date <= ref_date)
+    ).all()
+
+    print("Found:", len(recent_usage))
+
+
+if __name__ == "__main__":
+    run()