feat: add sources to retrieve news and divide the IA reflexions in 2 steps to limit the number of news

backend/internal/scheduler/scheduler.go

@@ -3,7 +3,6 @@ package scheduler
 import (
 	"context"
 	"fmt"
-	"strconv"
 
 	"github.com/robfig/cron/v3"
 	"github.com/tradarr/backend/internal/ai"
@@ -16,7 +15,7 @@ type Scheduler struct {
 	registry *scraper.Registry
 	pipeline *ai.Pipeline
 	repo     *models.Repository
-	entryID  cron.EntryID
+	entryIDs []cron.EntryID
 }
 
 func New(registry *scraper.Registry, pipeline *ai.Pipeline, repo *models.Repository) *Scheduler {
@@ -29,19 +28,10 @@ func New(registry *scraper.Registry, pipeline *ai.Pipeline, repo *models.Reposit
 }
 
 func (s *Scheduler) Start() error {
-	interval, err := s.getInterval()
-	if err != nil {
+	if err := s.loadSchedule(); err != nil {
 		return err
 	}
-
-	spec := fmt.Sprintf("@every %dm", interval)
-	s.entryID, err = s.cron.AddFunc(spec, s.run)
-	if err != nil {
-		return fmt.Errorf("add cron: %w", err)
-	}
-
 	s.cron.Start()
-	fmt.Printf("scheduler started, running every %d minutes\n", interval)
 	return nil
 }
 
@@ -50,39 +40,46 @@ func (s *Scheduler) Stop() {
 }
 
 func (s *Scheduler) Reload() error {
-	s.cron.Remove(s.entryID)
-	interval, err := s.getInterval()
-	if err != nil {
-		return err
+	for _, id := range s.entryIDs {
+		s.cron.Remove(id)
 	}
-	spec := fmt.Sprintf("@every %dm", interval)
-	s.entryID, err = s.cron.AddFunc(spec, s.run)
-	return err
+	s.entryIDs = nil
+	return s.loadSchedule()
+}
+
+func (s *Scheduler) loadSchedule() error {
+	slots, err := s.repo.ListScheduleSlots()
+	if err != nil {
+		return fmt.Errorf("load schedule: %w", err)
+	}
+	if len(slots) == 0 {
+		fmt.Println("scheduler: no schedule configured, scraping disabled")
+		return nil
+	}
+
+	for _, slot := range slots {
+		// Format cron: "minute hour * * day_of_week"
+		spec := fmt.Sprintf("%d %d * * %d", slot.Minute, slot.Hour, slot.DayOfWeek)
+		id, err := s.cron.AddFunc(spec, s.run)
+		if err != nil {
+			fmt.Printf("scheduler: invalid cron spec %q: %v\n", spec, err)
+			continue
+		}
+		s.entryIDs = append(s.entryIDs, id)
+	}
+
+	fmt.Printf("scheduler: %d time slots loaded\n", len(s.entryIDs))
+	return nil
 }
 
 func (s *Scheduler) run() {
-	fmt.Println("scheduler: running scraping cycle")
+	fmt.Println("scheduler: starting scraping cycle")
 	if err := s.registry.RunAll(); err != nil {
 		fmt.Printf("scheduler scrape error: %v\n", err)
 		return
 	}
-	fmt.Println("scheduler: running AI summaries")
+	fmt.Println("scheduler: starting AI summaries")
 	if err := s.pipeline.GenerateForAll(context.Background()); err != nil {
 		fmt.Printf("scheduler summary error: %v\n", err)
 	}
 }
-
-func (s *Scheduler) getInterval() (int, error) {
-	v, err := s.repo.GetSetting("scrape_interval_minutes")
-	if err != nil {
-		return 60, nil
-	}
-	if v == "" {
-		return 60, nil
-	}
-	n, err := strconv.Atoi(v)
-	if err != nil || n < 1 {
-		return 60, nil
-	}
-	return n, nil
-}