Merged in feature/textextraction (pull request #110)

Text Extraction * bases * go * splitting * structure * movetoasync * movetoasync * settinguptrigger * reorder * storevent * standardisepollingvalidation * unittests * fixlint * fixlint * awscfg * generatesample * followthrough * tests * clena * store * externalidcleanup * clientid * local * baseunittests * putobjecttests * tests
2025-04-02 18:50:03 +00:00
parent e6a4cb3990
commit 81e7223560
226 changed files with 17283 additions and 2744 deletions
@@ -0,0 +1,66 @@
+// **Listens For**: Body Containing Textract Result Document Metadata
+//
+// **Action**.
+//
+// Process textract output.
+//
+// An event with the document id is pushed to the QUERYSYNC queue.
+package main
+
+import (
+	"context"
+	"log/slog"
+	"os"
+
+	doctextprocessrunner "queryorchestration/api/docTextProcessRunner"
+	documenttext "queryorchestration/internal/document/text"
+	"queryorchestration/internal/server/runner"
+	"queryorchestration/internal/serviceconfig/aws"
+	"queryorchestration/internal/serviceconfig/objectstore"
+	"queryorchestration/internal/serviceconfig/queue/querysync"
+	"queryorchestration/internal/serviceconfig/textract"
+
+	_ "github.com/lib/pq"
+)
+
+type DocTextProcessConfig struct {
+	runner.BaseConfig[doctextprocessrunner.Body]
+	aws.AWSConfig
+	textract.TextractConfig
+	querysync.QuerySyncConfig
+	objectstore.ObjectStoreConfig
+}
+
+func main() {
+	ctx := context.Background()
+
+	cfg := &DocTextProcessConfig{}
+
+	cfg.ControllerFunc = func() runner.Controller[doctextprocessrunner.Body] {
+		text := documenttext.New(cfg)
+
+		return doctextprocessrunner.New(cfg.GetValidator(), &doctextprocessrunner.Services{
+			Text: text,
+		})
+	}
+
+	server, err := runner.New(ctx, cfg)
+	if err != nil {
+		slog.Error(err.Error())
+		os.Exit(1)
+	}
+
+	err = cfg.SetStoreClient(ctx)
+	if err != nil {
+		slog.Error(err.Error())
+		os.Exit(1)
+	}
+
+	err = cfg.SetTextractClient(ctx)
+	if err != nil {
+		slog.Error(err.Error())
+		os.Exit(1)
+	}
+
+	server.Listen(ctx)
+}