Durability pass: tests, clearer diversity/classify behavior, Calm Filters foundation

- Add pytest suite (34 tests) covering scoring thresholds, dedup clustering +
  representative selection + time window, brief source/category diversity,
  avoid-term phrase matching, and text canonicalization/truncation.
- Rewrite _select_diverse with an explicit, tested contract (best-first, one
  per source, backfill, then inject a second category by evicting the
  lowest-ranked pick).
- classify_articles now returns attempted/succeeded/skipped (ClassifyReport) so
  silent model failures are visible in both the cycle and classify output.
- Fix clean_text truncation to stay within max_len (ellipsis no longer
  overshoots).
- New filters.py: canonical FilterPrefs shape (include/mute topics+flavors,
  avoid_terms, pauses) and pure word/phrase-boundary matching engine seeding
  Calm Filters. Not yet wired into the API.

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

goodnews/cli.py

@@ -144,20 +144,24 @@ def main() -> None:
     elif args.command == "classify":
         init_db(conn)
         client = llm_client_from_args(args)
-        results = classify_articles(
+        report = classify_articles(
             conn,
             client,
             limit=args.limit,
             include_rejected=args.include_rejected,
             dry_run=args.dry_run,
         )
-        for article_id, scores in results:
+        for article_id, scores in report.results:
             accepted = "yes" if scores["accepted"] else "no"
             print(
                 f"[{article_id}] accepted={accepted} {scores['topic']}/{scores['flavor']} "
                 f"reason={scores['reason_code']}"
             )
             print(f"  {scores['reason_text']}")
+        print(
+            f"classify: attempted={report.attempted} succeeded={report.succeeded} "
+            f"skipped={report.skipped}"
+        )
         if args.dry_run:
             print("Dry run only; database was not updated.")
     elif args.command == "cycle":
@@ -294,7 +298,7 @@ def _run_cycle_locked(conn: sqlite3.Connection, args: argparse.Namespace) -> Non
             print(f"  classify {done}/{total} (article {article_id})", flush=True)
 
         try:
-            results = classify_articles(
+            report = classify_articles(
                 conn,
                 client,
                 limit=args.classify_limit,
@@ -302,7 +306,11 @@ def _run_cycle_locked(conn: sqlite3.Connection, args: argparse.Namespace) -> Non
                 only_unclassified=True,
                 progress=_progress,
             )
-            print(f"classify: {len(results)} new article(s) scored by {client.model}", flush=True)
+            print(
+                f"classify: attempted={report.attempted} succeeded={report.succeeded} "
+                f"skipped={report.skipped} (model {client.model})",
+                flush=True,
+            )
         except Exception as exc:  # endpoint down, timeout, etc. — keep going
             print(f"classify: skipped ({exc})", flush=True)