Separate current reference lookups from publication date restrictions

services/search/core.py

@@ -720,7 +720,7 @@ def searxng_search_results(query: str, count: int = 10, time_filter: str = None)
     # first instead of spending the full tool deadline retrying generic search
     # providers; the returned official URL lets the agent proceed to PDF tools.
     scholarly_title = _scholarly_title_from_query(provider_query)
-    if scholarly_title:
+    if scholarly_title and not time_filter:
         direct_results = [result for result in _direct_scholarly_title_results(scholarly_title, count)
                           if _result_matches_site_scope(provider_query, result)]
         if direct_results:

services/search/providers.py

@@ -220,14 +220,12 @@ def searxng_search_api(query: str, count: Optional[int] = None, categories: str
     if is_news and categories == "general":
         params["categories"] = "news"
         if time_filter in ("day", "week", "month", "year"):
-            # 'day' is too sparse on most SearXNG news engines — widen to a week
-            # so there's enough volume; the news category already biases recent.
-            params["time_range"] = "week" if time_filter in ("day", "week") else time_filter
+            params["time_range"] = time_filter
     else:
         params["categories"] = categories
         # Freshness and source category are independent: current manuals,
         # comparisons and documentation still belong in general search.
-        if not is_software_release_query and time_filter in ("day", "week", "month", "year"):
+        if time_filter in ("day", "week", "month", "year"):
             params["time_range"] = time_filter
         # Route general queries to engines that aren't blocked (default general
         # set returns 0 on this instance — see _GENERAL_ENGINES).