javier/Stirling-PDF
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

api docs

Browse Source
This commit is contained in:
Anthony Stirling
2023-05-08 15:07:53 +01:00
parent 632f29d1d6
commit 0f935b92cb
7 changed files with 197 additions and 37 deletions
Download Patch File Download Diff File Expand all files Collapse all files

27
src/main/java/stirling/software/SPDF/controller/api/other/ExtractImageScansController.java
View File

@@ -29,6 +29,8 @@ 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 stirling.software.SPDF.utils.PdfUtils;
import stirling.software.SPDF.utils.ProcessExecutor;
@@ -38,10 +40,27 @@ public class ExtractImageScansController {
private static final Logger logger = LoggerFactory.getLogger(ExtractImageScansController.class);
@PostMapping(consumes = "multipart/form-data", value = "/extract-image-scans")
public ResponseEntity<byte[]> extractImageScans(@RequestPart(required = true, value = "fileInput") MultipartFile inputFile,
@RequestParam(name = "angle_threshold", defaultValue = "5") int angleThreshold, @RequestParam(name = "tolerance", defaultValue = "20") int tolerance,
@RequestParam(name = "min_area", defaultValue = "8000") int minArea, @RequestParam(name = "min_contour_area", defaultValue = "500") int minContourArea,
@RequestParam(name = "border_size", defaultValue = "1") int borderSize) throws IOException, InterruptedException {
@Operation(summary = "Extract image scans from an input file",
description = "This endpoint extracts image scans from a given file based on certain parameters. Users can specify angle threshold, tolerance, minimum area, minimum contour area, and border size.")
public ResponseEntity<byte[]> extractImageScans(
@RequestPart(required = true, value = "fileInput")
@Parameter(description = "The input file containing image scans")
MultipartFile inputFile,
@RequestParam(name = "angle_threshold", defaultValue = "5")
@Parameter(description = "The angle threshold for the image scan extraction", example = "5")
int angleThreshold,
@RequestParam(name = "tolerance", defaultValue = "20")
@Parameter(description = "The tolerance for the image scan extraction", example = "20")
int tolerance,
@RequestParam(name = "min_area", defaultValue = "8000")
@Parameter(description = "The minimum area for the image scan extraction", example = "8000")
int minArea,
@RequestParam(name = "min_contour_area", defaultValue = "500")
@Parameter(description = "The minimum contour area for the image scan extraction", example = "500")
int minContourArea,
@RequestParam(name = "border_size", defaultValue = "1")
@Parameter(description = "The border size for the image scan extraction", example = "1")
int borderSize) throws IOException, InterruptedException {
String fileName = inputFile.getOriginalFilename();
String extension = fileName.substring(fileName.lastIndexOf(".") + 1);

14
src/main/java/stirling/software/SPDF/controller/api/other/ExtractImagesController.java
View File

@@ -26,15 +26,25 @@ 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 stirling.software.SPDF.utils.PdfUtils;
import io.swagger.v3.oas.annotations.media.Schema;
@RestController
public class ExtractImagesController {
private static final Logger logger = LoggerFactory.getLogger(ExtractImagesController.class);
@PostMapping(consumes = "multipart/form-data", value = "/extract-images")
public ResponseEntity<byte[]> extractImages(@RequestPart(required = true, value = "fileInput") MultipartFile file, @RequestParam("format") String format) throws IOException {
@Operation(summary = "Extract images from a PDF file",
description = "This endpoint extracts images from a given PDF file and returns them in a zip file. Users can specify the output image format.")
public ResponseEntity<byte[]> extractImages(
@RequestPart(required = true, value = "fileInput")
@Parameter(description = "The input PDF file containing images")
MultipartFile file,
@RequestParam("format")
@Parameter(description = "The output image format e.g., 'png', 'jpeg', or 'gif'", schema = @Schema(allowableValues = {"png", "jpeg", "gif"}))
String format) throws IOException {
System.out.println(System.currentTimeMillis() + "file=" + file.getName() + ", format=" + format);
PDDocument document = PDDocument.load(file.getBytes());

47
src/main/java/stirling/software/SPDF/controller/api/other/MetadataController.java
View File

@@ -17,6 +17,8 @@ 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 stirling.software.SPDF.utils.PdfUtils;
@RestController
@@ -35,13 +37,44 @@ public class MetadataController {
}
@PostMapping(consumes = "multipart/form-data", value = "/update-metadata")
public ResponseEntity<byte[]> metadata(@RequestPart(required = true, value = "fileInput") MultipartFile pdfFile,
@RequestParam(value = "deleteAll", required = false, defaultValue = "false") Boolean deleteAll, @RequestParam(value = "author", required = false) String author,
@RequestParam(value = "creationDate", required = false) String creationDate, @RequestParam(value = "creator", required = false) String creator,
@RequestParam(value = "keywords", required = false) String keywords, @RequestParam(value = "modificationDate", required = false) String modificationDate,
@RequestParam(value = "producer", required = false) String producer, @RequestParam(value = "subject", required = false) String subject,
@RequestParam(value = "title", required = false) String title, @RequestParam(value = "trapped", required = false) String trapped,
@RequestParam Map<String, String> allRequestParams) throws IOException {
@Operation(summary = "Update metadata of a PDF file",
description = "This endpoint allows you to update the metadata of a given PDF file. You can add, modify, or delete standard and custom metadata fields.")
public ResponseEntity<byte[]> metadata(
@RequestPart(required = true, value = "fileInput")
@Parameter(description = "The input PDF file to update metadata")
MultipartFile pdfFile,
@RequestParam(value = "deleteAll", required = false, defaultValue = "false")
@Parameter(description = "Delete all metadata if set to true")
Boolean deleteAll,
@RequestParam(value = "author", required = false)
@Parameter(description = "The author of the document")
String author,
@RequestParam(value = "creationDate", required = false)
@Parameter(description = "The creation date of the document (format: yyyy/MM/dd HH:mm:ss)")
String creationDate,
@RequestParam(value = "creator", required = false)
@Parameter(description = "The creator of the document")
String creator,
@RequestParam(value = "keywords", required = false)
@Parameter(description = "The keywords for the document")
String keywords,
@RequestParam(value = "modificationDate", required = false)
@Parameter(description = "The modification date of the document (format: yyyy/MM/dd HH:mm:ss)")
String modificationDate,
@RequestParam(value = "producer", required = false)
@Parameter(description = "The producer of the document")
String producer,
@RequestParam(value = "subject", required = false)
@Parameter(description = "The subject of the document")
String subject,
@RequestParam(value = "title", required = false)
@Parameter(description = "The title of the document")
String title,
@RequestParam(value = "trapped", required = false)
@Parameter(description = "The trapped status of the document")
String trapped,
@RequestParam Map<String, String> allRequestParams)
throws IOException {
// Load the PDF file into a PDDocument
PDDocument document = PDDocument.load(pdfFile.getBytes());

40
src/main/java/stirling/software/SPDF/controller/api/other/OCRController.java
View File

@@ -24,6 +24,9 @@ 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.media.Schema;
import stirling.software.SPDF.utils.PdfUtils;
import stirling.software.SPDF.utils.ProcessExecutor;
@@ -43,13 +46,36 @@ public class OCRController {
}
@PostMapping(consumes = "multipart/form-data", value = "/ocr-pdf")
public ResponseEntity<byte[]> processPdfWithOCR(@RequestPart(required = true, value = "fileInput") MultipartFile inputFile,
@RequestParam("languages") List<String> selectedLanguages, @RequestParam(name = "sidecar", required = false) Boolean sidecar,
@RequestParam(name = "deskew", required = false) Boolean deskew, @RequestParam(name = "clean", required = false) Boolean clean,
@RequestParam(name = "clean-final", required = false) Boolean cleanFinal, @RequestParam(name = "ocrType", required = false) String ocrType,
@RequestParam(name = "ocrRenderType", required = false, defaultValue = "hocr") String ocrRenderType,
@RequestParam(name = "removeImagesAfter", required = false) Boolean removeImagesAfter)
throws IOException, InterruptedException {
@Operation(summary = "Process a PDF file with OCR",
description = "This endpoint processes a PDF file using OCR (Optical Character Recognition). Users can specify languages, sidecar, deskew, clean, cleanFinal, ocrType, ocrRenderType, and removeImagesAfter options.")
public ResponseEntity<byte[]> processPdfWithOCR(
@RequestPart(required = true, value = "fileInput")
@Parameter(description = "The input PDF file to be processed with OCR")
MultipartFile inputFile,
@RequestParam("languages")
@Parameter(description = "List of languages to use in OCR processing")
List<String> selectedLanguages,
@RequestParam(name = "sidecar", required = false)
@Parameter(description = "Include OCR text in a sidecar text file if set to true")
Boolean sidecar,
@RequestParam(name = "deskew", required = false)
@Parameter(description = "Deskew the input file if set to true")
Boolean deskew,
@RequestParam(name = "clean", required = false)
@Parameter(description = "Clean the input file if set to true")
Boolean clean,
@RequestParam(name = "clean-final", required = false)
@Parameter(description = "Clean the final output if set to true")
Boolean cleanFinal,
@RequestParam(name = "ocrType", required = false)
@Parameter(description = "Specify the OCR type, e.g., 'skip-text', 'force-ocr', or 'Normal'", schema = @Schema(allowableValues = {"skip-text", "force-ocr", "Normal"}))
String ocrType,
@RequestParam(name = "ocrRenderType", required = false, defaultValue = "hocr")
@Parameter(description = "Specify the OCR render type, either 'hocr' or 'sandwich'", schema = @Schema(allowableValues = {"hocr", "sandwich"}))
String ocrRenderType,
@RequestParam(name = "removeImagesAfter", required = false)
@Parameter(description = "Remove images from the output PDF if set to true")
Boolean removeImagesAfter) throws IOException, InterruptedException {
// --output-type pdfa
if (selectedLanguages == null || selectedLanguages.isEmpty()) {