chore: setup boilerplate for particle processor worker

This sets up an example of listening to firestore data and reacting. It
also sets up our first worker in the kubernetes cluster.

go/cmd/particleprocessorworker/main.go

@@ -0,0 +1,67 @@
+package main
+
+import (
+	"context"
+	"fmt"
+	"log"
+	"log/slog"
+
+	"github.com/flowy-live/llink/internal/utils"
+	"google.golang.org/grpc/codes"
+	"google.golang.org/grpc/status"
+
+	"cloud.google.com/go/firestore"
+)
+
+func createClient(ctx context.Context) *firestore.Client {
+	projectId := utils.MustGetEnv("GCP_PROJECT")
+
+	client, err := firestore.NewClient(ctx, projectId)
+	if err != nil {
+		log.Fatalf("Failed to create client: %v", err)
+	}
+	return client
+}
+
+// The purpose of the particle processor worker is to listen for new particles
+// across all streams and perform side effects such as
+//   - generate transcript if the particle is of type media
+//   - send mobile notifications if a client is offline
+//   - update the parent stream's `last_child_created_at`
+func main() {
+	ctx := context.Background()
+	client := createClient(ctx)
+	defer client.Close()
+
+	it := client.CollectionGroup("children").Snapshots(ctx)
+	var initialLoad = true
+	for {
+		snap, err := it.Next()
+		if e := status.Code(err); e == codes.DeadlineExceeded || e == codes.Canceled {
+			panic(fmt.Errorf("error: %w", err))
+		}
+		if err != nil {
+			slog.Error("error in processing snapshot", "error", err)
+		}
+
+		if snap != nil {
+			if initialLoad {
+				slog.Info("initial load", "changeCount", len(snap.Changes))
+			} else {
+				for _, change := range snap.Changes {
+					switch change.Kind {
+					case firestore.DocumentAdded:
+						slog.Info("document added: ")
+					case firestore.DocumentModified:
+						slog.Info("document modified")
+					case firestore.DocumentRemoved:
+						slog.Info("document removed")
+					}
+					slog.Info("received document snapshot", "data", change.Doc.Data())
+				}
+			}
+		}
+
+		initialLoad = false
+	}
+}