Track Spotify origin (liked/playlist/top); make top tracks opt-in; clean Ctrl+C

- Tag each Spotify track with its origin (liked / from_playlist / from_top,
  combinable) so genuinely-saved tracks are distinguishable from ones only
  surfaced by the top-tracks endpoint.
- Top tracks are no longer added by default (they inflated "missing" with
  unowned listening stats). --include-top pulls them as full objects so they
  carry an ISRC for real ID matching.
- On --spotify-full, prune Spotify rows that are neither liked nor in a
  playlist (top-only / orphaned), cleaning an existing DB in one pass.
- Cached-ingest fallback honours include_top and prunes too.
- In-place SQLite migration adds the new columns (no DB rebuild needed).
- Ctrl+C exits cleanly (130) with a message instead of a traceback;
  incremental progress is already committed, so re-run to continue.
- Surface origin breakdown in the fetch log, match summary, and GUI.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Gui.py

@@ -49,6 +49,12 @@ def main():
         c4.metric("Have (text)", s["present_text"])
         coverage = (s["has_mbid"] / s["total"] * 100) if s["total"] else 0
         c5.metric("MBID coverage", f"{coverage:.0f}%")
+        if source == "spotify":
+            st.caption(
+                f"Origins — liked: {s.get('liked', 0)} · "
+                f"in a playlist: {s.get('from_playlist', 0)} · "
+                f"top tracks: {s.get('from_top', 0)}"
+            )
 
     local = summ.get("local", {})
     st.caption(
@@ -80,7 +86,7 @@ def main():
         return
 
     df = pd.DataFrame(rows)[
-        ["status", "source", "artist", "title", "album", "matched_by",
+        ["status", "source", "origin", "artist", "title", "album", "matched_by",
          "isrc", "mbid", "play_count", "spotify_id"]
     ]
     st.write(f"**{len(df)}** tracks")