RED-7375: integrate Table Extractor

* fix filenames
This commit is contained in:
Kilian Schuettler 2023-09-06 14:20:38 +02:00
parent 49bcc43716
commit 74ac3c0c34
3 changed files with 5 additions and 3 deletions

View File

@ -1,5 +1,7 @@
package com.iqser.red.service.persistence.management.v1.processor.model; package com.iqser.red.service.persistence.management.v1.processor.model;
import com.iqser.red.service.persistence.service.v1.api.shared.model.dossiertemplate.dossier.file.FileType;
import lombok.AllArgsConstructor; import lombok.AllArgsConstructor;
import lombok.Builder; import lombok.Builder;
import lombok.Data; import lombok.Data;
@ -11,7 +13,7 @@ import lombok.NoArgsConstructor;
@NoArgsConstructor @NoArgsConstructor
public class TableExtractorRequest { public class TableExtractorRequest {
public static final String TABLE_EXTRACTOR_FILE_EXTENSION = "EXTRACTED_TABLES.json.gz"; public static final String TABLE_EXTRACTOR_FILE_EXTENSION = FileType.EXTRACTED_TABLES.name() + FileType.EXTRACTED_TABLES.getExtension() + ".gz";
public static final String TARGET_FILE_EXTENSION = "ORIGIN.pdf.gz"; public static final String TARGET_FILE_EXTENSION = "ORIGIN.pdf.gz";

View File

@ -176,7 +176,7 @@ public class FileStatusService {
return; return;
} }
if (settings.isTableExtractorEnabled() && !fileManagementStorageService.objectExists(dossierId, fileId, FileType.INVISIBLE_TABLES)) { if (settings.isTableExtractorEnabled() && !fileManagementStorageService.objectExists(dossierId, fileId, FileType.EXTRACTED_TABLES)) {
log.info("Add file: {} from dossier {} to Table Extractor queue", fileId, dossierId); log.info("Add file: {} from dossier {} to Table Extractor queue", fileId, dossierId);
addToTableExtractorQueue(dossierId, fileId); addToTableExtractorQueue(dossierId, fileId);
return; return;

View File

@ -17,7 +17,7 @@ public enum FileType {
TEXT_HIGHLIGHTS(".json"), TEXT_HIGHLIGHTS(".json"),
FIGURE(".json"), FIGURE(".json"),
TABLES(".json"), TABLES(".json"),
INVISIBLE_TABLES(".json"), EXTRACTED_TABLES(".json"),
COMPONENTS(".json"), COMPONENTS(".json"),
// document is split into 4 files, all should be overridden/deleted at the same time // document is split into 4 files, all should be overridden/deleted at the same time
DOCUMENT_TEXT(".json"), DOCUMENT_TEXT(".json"),