feat(scraper): add checkpointing and richer page extraction

Add resumable checkpoint support so long scrapes can recover from
interruptions instead of restarting from scratch.

- introduce autosave/load/clear checkpoint flow in `.cache/scrape-state.json`, including SIGINT/SIGTERM save-on-exit handling
- expand parsing/model output to capture legacy and portable infobox fields, primary image URLs, effects, recipes, raw tables, and improved category extraction
- skip infobox tables during recipe parsing to avoid false recipe matches
- add cache log event type, ignore cache/output artifacts, and document new autosave tuning options in READMEfeat(scraper): add checkpointing and richer page extraction

Add resumable checkpoint support so long scrapes can recover from
interruptions instead of restarting from scratch.

- introduce autosave/load/clear checkpoint flow in `.cache/scrape-state.json`, including SIGINT/SIGTERM save-on-exit handling
- expand parsing/model output to capture legacy and portable infobox fields, primary image URLs, effects, recipes, raw tables, and improved category extraction
- skip infobox tables during recipe parsing to avoid false recipe matches
- add cache log event type, ignore cache/output artifacts, and document new autosave tuning options in README

internal/app/run.go

@@ -4,6 +4,8 @@ import (
 	"encoding/json"
 	"math/rand"
 	"os"
+	"os/signal"
+	"syscall"
 	"time"
 
 	"scrappr/internal/logx"
@@ -20,6 +22,7 @@ func Run() error {
 
 	cfg := scraper.DefaultConfig()
 	s := scraper.New(cfg)
+	installSignalCheckpoint(s)
 
 	dataset, err := s.Run()
 	if err != nil {
@@ -39,9 +42,27 @@ func Run() error {
 		outputPath,
 	)
 
+	if err := s.ClearCheckpoint(); err != nil {
+		logx.Eventf("warn", "failed to clear checkpoint: %v", err)
+	}
+
 	return nil
 }
 
+func installSignalCheckpoint(s *scraper.Scraper) {
+	signals := make(chan os.Signal, 1)
+	signal.Notify(signals, os.Interrupt, syscall.SIGTERM)
+
+	go func() {
+		sig := <-signals
+		logx.Eventf("warn", "received %s, saving checkpoint before exit", sig.String())
+		if err := s.SaveCheckpoint("signal"); err != nil {
+			logx.Eventf("error", "failed to save checkpoint on signal: %v", err)
+		}
+		os.Exit(130)
+	}()
+}
+
 func writeDataset(outputPath string, dataset model.Dataset) error {
 	file, err := os.Create(outputPath)
 	if err != nil {