Add interval-aware polling and a 'cycle' command for scheduling

- poll_due_sources(): polls only sources whose last successful poll is older
  than their poll_interval_minutes (or never polled), finally giving that
  config field meaning.
- classify gains only_unclassified to spend the LLM solely on new (heuristic)
  articles, so a frequent scheduled run stays cheap.
- 'cycle' command runs poll-due -> classify-new -> rebuild-today's-brief, with
  each step non-fatal so a down model endpoint or empty day never aborts it.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

goodnews/llm.py

@@ -200,8 +200,11 @@ def classify_articles(
     limit: int,
     include_rejected: bool = False,
     dry_run: bool = False,
+    only_unclassified: bool = False,
 ) -> list[tuple[int, dict]]:
-    rows = _classification_candidates(conn, limit=limit, include_rejected=include_rejected)
+    rows = _classification_candidates(
+        conn, limit=limit, include_rejected=include_rejected, only_unclassified=only_unclassified
+    )
     results = []
     for row in rows:
         try:
@@ -297,8 +300,16 @@ def _classification_candidates(
     conn: sqlite3.Connection,
     limit: int,
     include_rejected: bool,
+    only_unclassified: bool = False,
 ) -> list[sqlite3.Row]:
-    where = "" if include_rejected else "WHERE s.accepted = 1 OR s.constructive_score >= 4"
+    filters = []
+    if not include_rejected:
+        filters.append("(s.accepted = 1 OR s.constructive_score >= 4)")
+    if only_unclassified:
+        # Articles still carrying the fast heuristic score, i.e. not yet judged
+        # by the model. Lets a scheduled cycle only spend the LLM on new items.
+        filters.append("s.model_name LIKE 'heuristic-%'")
+    where = ("WHERE " + " AND ".join(filters)) if filters else ""
     return conn.execute(
         f"""
         SELECT