Compare commits

..
Author SHA1 Message Date
Anthony Stirling 8a375c5da4 revert 2025-07-11 23:01:18 +01:00
Anthony Stirling ae3ed72283 AOP Fixes 2025-07-11 22:53:16 +01:00
bbf5d5f6d4 Support multi-file async job results and ZIP extraction (#3922)
# Description of Changes

This PR introduces multi-file support for asynchronous jobs in the
Stirling PDF backend, enabling jobs to return and manage multiple result
files. Previously, job results were limited to a single file represented
by fileId, originalFileName, and contentType. This change replaces that
legacy structure with a new ResultFile abstraction and expands the
functionality throughout the core system.

ZIP File Support
If a job result is a ZIP file:
It is automatically unpacked using buffered streaming.
Each contained file is stored individually and recorded as a ResultFile.
The original ZIP is deleted after successful extraction.
If ZIP extraction fails, the job result is treated as a single file.


New and Updated API Endpoints

1. GET /api/v1/general/job/{jobId}/result

If the job has multiple files → returns a JSON metadata list.

If the job has a single file → streams the file directly.

Includes UTF-8-safe Content-Disposition headers for filename support.

2. GET /api/v1/general/job/{jobId}/result/files
New endpoint that returns:

```json
{
  "jobId": "123",
  "fileCount": 2,
  "files": [
    {
      "fileId": "abc",
      "fileName": "page1.pdf",
      "contentType": "application/pdf",
      "fileSize": 12345
    },
    ...
  ]
}
```


3. GET /api/v1/general/files/{fileId}/metadata
Returns metadata for a specific file:


4. GET /api/v1/general/files/{fileId}
Downloads a file by fileId, using metadata to determine filename and
content type.

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: pixeebot[bot] <104101892+pixeebot[bot]@users.noreply.github.com>
2025-07-11 13:15:55 +01:00
4 changed files with 14 additions and 87 deletions
@@ -43,6 +43,7 @@ public class AutoJobAspect {
// This aspect will run before any audit aspects due to @Order(0)
// Extract parameters from the request and annotation
boolean async = Boolean.parseBoolean(request.getParameter("async"));
log.debug("AutoJobAspect: Processing {} {} with async={}", request.getMethod(), request.getRequestURI(), async);
long timeout = autoJobPostMapping.timeout();
int retryCount = autoJobPostMapping.retryCount();
boolean trackProgress = autoJobPostMapping.trackProgress();
@@ -54,19 +55,8 @@ public class AutoJobAspect {
retryCount,
trackProgress);
// Copy and process arguments
// In a test environment, we might need to update the original objects for verification
boolean isTestEnvironment = false;
try {
isTestEnvironment = Class.forName("org.junit.jupiter.api.Test") != null;
} catch (ClassNotFoundException e) {
// Not in a test environment
}
Object[] args =
isTestEnvironment
? processArgsInPlace(joinPoint.getArgs(), async)
: copyAndProcessArgs(joinPoint.getArgs(), async);
// Process arguments in-place to avoid type mismatch issues
Object[] args = processArgsInPlace(joinPoint.getArgs(), async);
// Extract queueable and resourceWeight parameters and validate
boolean queueable = autoJobPostMapping.queueable();
@@ -229,79 +219,10 @@ public class AutoJobAspect {
resourceWeight);
}
/**
* Creates deep copies of arguments when needed to avoid mutating the original objects
* Particularly important for PDFFile objects that might be reused by Spring
*
* @param originalArgs The original arguments
* @param async Whether this is an async operation
* @return A new array with safely processed arguments
*/
private Object[] copyAndProcessArgs(Object[] originalArgs, boolean async) {
if (originalArgs == null || originalArgs.length == 0) {
return originalArgs;
}
Object[] processedArgs = new Object[originalArgs.length];
// Copy all arguments
for (int i = 0; i < originalArgs.length; i++) {
Object arg = originalArgs[i];
if (arg instanceof PDFFile pdfFile) {
// Create a copy of PDFFile to avoid mutating the original
// Using direct property access instead of reflection for better performance
PDFFile pdfFileCopy = new PDFFile();
pdfFileCopy.setFileId(pdfFile.getFileId());
pdfFileCopy.setFileInput(pdfFile.getFileInput());
// Case 1: fileId is provided but no fileInput
if (pdfFileCopy.getFileInput() == null && pdfFileCopy.getFileId() != null) {
try {
log.debug("Using fileId {} to get file content", pdfFileCopy.getFileId());
MultipartFile file = fileStorage.retrieveFile(pdfFileCopy.getFileId());
pdfFileCopy.setFileInput(file);
} catch (Exception e) {
throw new RuntimeException(
"Failed to resolve file by ID: " + pdfFileCopy.getFileId(), e);
}
}
// Case 2: For async requests, we need to make a copy of the MultipartFile
else if (async && pdfFileCopy.getFileInput() != null) {
try {
log.debug("Making persistent copy of uploaded file for async processing");
MultipartFile originalFile = pdfFileCopy.getFileInput();
String fileId = fileStorage.storeFile(originalFile);
// Store the fileId for later reference
pdfFileCopy.setFileId(fileId);
// Replace the original MultipartFile with our persistent copy
MultipartFile persistentFile = fileStorage.retrieveFile(fileId);
pdfFileCopy.setFileInput(persistentFile);
log.debug("Created persistent file copy with fileId: {}", fileId);
} catch (IOException e) {
throw new RuntimeException(
"Failed to create persistent copy of uploaded file", e);
}
}
processedArgs[i] = pdfFileCopy;
} else {
// For non-PDFFile objects, just pass the original reference
// If other classes need copy-on-write, add them here
processedArgs[i] = arg;
}
}
return processedArgs;
}
/**
* Processes arguments in-place for testing purposes This is similar to our original
* implementation before introducing copy-on-write It's only used in test environments to
* maintain test compatibility
* Processes arguments in-place to handle file resolution and async file persistence.
* This approach avoids type mismatch issues by modifying the original objects directly.
*
* @param originalArgs The original arguments
* @param async Whether this is an async operation
@@ -6,6 +6,8 @@ import java.util.Collections;
import java.util.List;
import java.util.concurrent.CopyOnWriteArrayList;
import com.fasterxml.jackson.annotation.JsonIgnore;
import lombok.AllArgsConstructor;
import lombok.Builder;
import lombok.Data;
@@ -28,6 +30,7 @@ public class JobResult {
private String error;
/** List of result files for jobs that produce files */
@JsonIgnore
private List<ResultFile> resultFiles;
/** Time when the job was created */
@@ -1,6 +1,5 @@
package stirling.software.common.service;
import io.github.pixee.security.ZipSecurity;
import java.io.ByteArrayInputStream;
import java.io.ByteArrayOutputStream;
import java.io.IOException;
@@ -21,6 +20,8 @@ import org.springframework.http.MediaType;
import org.springframework.stereotype.Service;
import org.springframework.web.multipart.MultipartFile;
import io.github.pixee.security.ZipSecurity;
import jakarta.annotation.PreDestroy;
import lombok.extern.slf4j.Slf4j;
@@ -361,7 +362,8 @@ public class TaskManager {
MultipartFile zipFile = fileStorage.retrieveFile(zipFileId);
try (ZipInputStream zipIn =
ZipSecurity.createHardenedInputStream(new ByteArrayInputStream(zipFile.getBytes()))) {
ZipSecurity.createHardenedInputStream(
new ByteArrayInputStream(zipFile.getBytes()))) {
ZipEntry entry;
while ((entry = zipIn.getNextEntry()) != null) {
if (!entry.isDirectory()) {
@@ -29,7 +29,8 @@ public class CleanUrlInterceptor implements HandlerInterceptor {
"type",
"principal",
"startDate",
"endDate");
"endDate",
"async");
@Override
public boolean preHandle(