130 lines
5.5 KiB
Java
130 lines
5.5 KiB
Java
package stirling.software.SPDF.controller.api;
|
|
|
|
import java.io.ByteArrayOutputStream;
|
|
import java.io.IOException;
|
|
import java.io.InputStream;
|
|
import java.nio.file.Files;
|
|
import java.nio.file.Path;
|
|
import java.util.ArrayList;
|
|
import java.util.List;
|
|
import java.util.stream.Collectors;
|
|
import java.util.zip.ZipEntry;
|
|
import java.util.zip.ZipOutputStream;
|
|
|
|
import org.apache.pdfbox.pdmodel.PDDocument;
|
|
import org.apache.pdfbox.pdmodel.PDPage;
|
|
import org.slf4j.Logger;
|
|
import org.slf4j.LoggerFactory;
|
|
import org.springframework.http.MediaType;
|
|
import org.springframework.http.ResponseEntity;
|
|
import org.springframework.web.bind.annotation.PostMapping;
|
|
import org.springframework.web.bind.annotation.RequestParam;
|
|
import org.springframework.web.bind.annotation.RequestPart;
|
|
import org.springframework.web.bind.annotation.RestController;
|
|
import org.springframework.web.multipart.MultipartFile;
|
|
|
|
import io.swagger.v3.oas.annotations.Operation;
|
|
import io.swagger.v3.oas.annotations.Parameter;
|
|
import io.swagger.v3.oas.annotations.tags.Tag;
|
|
import stirling.software.SPDF.utils.GeneralUtils;
|
|
import stirling.software.SPDF.utils.WebResponseUtils;
|
|
|
|
@RestController
|
|
@Tag(name = "General", description = "General APIs")
|
|
public class SplitPDFController {
|
|
|
|
private static final Logger logger = LoggerFactory.getLogger(SplitPDFController.class);
|
|
|
|
@PostMapping(consumes = "multipart/form-data", value = "/split-pages")
|
|
@Operation(summary = "Split a PDF file into separate documents",
|
|
description = "This endpoint splits a given PDF file into separate documents based on the specified page numbers or ranges. Users can specify pages using individual numbers, ranges, or 'all' for every page. Input:PDF Output:PDF Type:SIMO")
|
|
public ResponseEntity<byte[]> splitPdf(
|
|
@RequestPart(required = true, value = "fileInput")
|
|
@Parameter(description = "The input PDF file to be split")
|
|
MultipartFile file,
|
|
@RequestParam("pages")
|
|
@Parameter(description = "The pages to be included in separate documents. Specify individual page numbers (e.g., '1,3,5'), ranges (e.g., '1-3,5-7'), or 'all' for every page.")
|
|
String pages) throws IOException {
|
|
// parse user input
|
|
|
|
// open the pdf document
|
|
InputStream inputStream = file.getInputStream();
|
|
PDDocument document = PDDocument.load(inputStream);
|
|
|
|
List<Integer> pageNumbers = new ArrayList<>();
|
|
pages = pages.replaceAll("\\s+", ""); // remove whitespaces
|
|
if (pages.toLowerCase().equals("all")) {
|
|
for (int i = 0; i < document.getNumberOfPages(); i++) {
|
|
pageNumbers.add(i);
|
|
}
|
|
} else {
|
|
String[] splitPoints = pages.split(",");
|
|
for (String splitPoint : splitPoints) {
|
|
List<Integer> orderedPages = GeneralUtils.parsePageList(new String[] {splitPoint}, document.getNumberOfPages());
|
|
pageNumbers.addAll(orderedPages);
|
|
}
|
|
// Add the last page as a split point
|
|
pageNumbers.add(document.getNumberOfPages() - 1);
|
|
}
|
|
|
|
logger.info("Splitting PDF into pages: {}", pageNumbers.stream().map(String::valueOf).collect(Collectors.joining(",")));
|
|
|
|
// split the document
|
|
List<ByteArrayOutputStream> splitDocumentsBoas = new ArrayList<>();
|
|
int previousPageNumber = 0;
|
|
for (int splitPoint : pageNumbers) {
|
|
try (PDDocument splitDocument = new PDDocument()) {
|
|
for (int i = previousPageNumber; i <= splitPoint; i++) {
|
|
PDPage page = document.getPage(i);
|
|
splitDocument.addPage(page);
|
|
logger.debug("Adding page {} to split document", i);
|
|
}
|
|
previousPageNumber = splitPoint + 1;
|
|
|
|
ByteArrayOutputStream baos = new ByteArrayOutputStream();
|
|
splitDocument.save(baos);
|
|
|
|
splitDocumentsBoas.add(baos);
|
|
} catch (Exception e) {
|
|
logger.error("Failed splitting documents and saving them", e);
|
|
throw e;
|
|
}
|
|
}
|
|
|
|
|
|
// closing the original document
|
|
document.close();
|
|
|
|
Path zipFile = Files.createTempFile("split_documents", ".zip");
|
|
|
|
String filename = file.getOriginalFilename().replaceFirst("[.][^.]+$", "");
|
|
try (ZipOutputStream zipOut = new ZipOutputStream(Files.newOutputStream(zipFile))) {
|
|
// loop through the split documents and write them to the zip file
|
|
for (int i = 0; i < splitDocumentsBoas.size(); i++) {
|
|
String fileName = filename + "_" + (i + 1) + ".pdf";
|
|
ByteArrayOutputStream baos = splitDocumentsBoas.get(i);
|
|
byte[] pdf = baos.toByteArray();
|
|
|
|
// Add PDF file to the zip
|
|
ZipEntry pdfEntry = new ZipEntry(fileName);
|
|
zipOut.putNextEntry(pdfEntry);
|
|
zipOut.write(pdf);
|
|
zipOut.closeEntry();
|
|
|
|
logger.info("Wrote split document {} to zip file", fileName);
|
|
}
|
|
} catch (Exception e) {
|
|
logger.error("Failed writing to zip", e);
|
|
throw e;
|
|
}
|
|
|
|
logger.info("Successfully created zip file with split documents: {}", zipFile.toString());
|
|
byte[] data = Files.readAllBytes(zipFile);
|
|
Files.delete(zipFile);
|
|
|
|
// return the Resource in the response
|
|
return WebResponseUtils.bytesToWebResponse(data, filename + ".zip", MediaType.APPLICATION_OCTET_STREAM);
|
|
|
|
}
|
|
|
|
}
|