refactor: solidify startup UX and engine-aware preprocessing

- Slim core deps: move ML stack to optional extras (paddle/tables/figures/scientific/full)
- Lazy settings proxy with config search paths (env var, cwd, user dir)
- New commands: init, demo, setup [basic|full], first-run guard on run/watch
- Engine-aware preprocessing: Paddle gets original image (fixes dark mode 0.83->0.95)
- Results table shows Skipped count; lazy run-dir creation
- kg_ocr marked experimental with extra, Docker defaults with OCR_PIPELINE_CONFIG
- 25/25 tests, ruff clean

src/ocr_pipeline/watch/watcher.py

@@ -113,7 +113,15 @@ class Watcher:
                         failed += 1
 
         if self.run_id is not None:
-            self.db.complete_run(self.run_id, {"total": processed + failed, "successful": processed, "failed": failed, "chunks": 0})
+            self.db.complete_run(
+                self.run_id,
+                {
+                    "total": processed + failed,
+                    "successful": processed,
+                    "failed": failed,
+                    "chunks": 0,
+                },
+            )
         logger.info("watcher_stopped", run_id=self.run_id, processed=processed, failed=failed)
 
     def _process_file(self, path: Path) -> None: