Francisco Dorado
Francisco Dorado Software Architect at sngular.com in Seville. Specialised in backend technologies based in the Java ecosystem. Currently working on Microservices using Spring Framework and AWS Cloud technologies

Spring upload and download files to S3 with localstack

Spring upload and download files to S3 with localstack

Summary

In this post we are going to create an example of REST Controllers for upload and download files in AWS S3 using LocalStack.

The controllers

The upload controller will use a MultipartFile to receive the file. The max filename size could be configured by properties.

1
2
3
4
5

@PostMapping(value = “/upload”, produces = “application/json”)
public ResponseEntity<String> upload(@RequestParam(“file”) MultipartFile file) {
    String key = storageService.upload(file);
    return new ResponseEntity<>(key, HttpStatus.OK);
}

The download controller will return a Resource file.

1
2
3
4
5
6
7
8
9
10
11

@GetMapping(“/download”)
public ResponseEntity<Resource> download(String id) {
    DownloadedResource downloadedResource = storageService.download(id);
        
    return ResponseEntity.ok()
		.header(HttpHeaders.CONTENT_DISPOSITION, “attachment; filename=“ 
			+ downloadedResource.getFileName())
          	.contentLength(downloadedResource.getContentLength())
		.contentType(MediaType.APPLICATION_OCTET_STREAM)
          	.body(new InputStreamResource(downloadedResource.getInputStream()));
}

The Storage Service

One storage service will have to provide the implementation for the next interface:

1
2
3
4
5
6

public interface StorageService {
    
    String upload(MultipartFile multipartFile);
    
    DownloadedResource download(String id);
}

In our example, we provide a S3StorageService with this implementation and uses AmazonS3 client to upload and upload the file.

  • The upload put the object in a bucket in S3 using the InputStream of MultipartFile and add some extra object metadata (content length, content type and file extension)

    1

      amazonS3.putObject(bucketName, key, multipartFile.getInputStream(), extractObjectMetadata(multipartFile));

  • The download create a DownloadedResource object with all the information stored in S3.

Configuration for S3 in Localstack

First we should create a docker-compose.yml file where indicates the localstack service with S3 enabled.

1
2
3
4
5
6
7
8
9
10
11

services:
  localstack:
    image: localstack/localstack
    ports:
      - “4572:4572”
    environment:
      - SERVICES=s3
      - DOCKER_HOST=unix:///var/run/docker.sock
    volumes:
      - “${TMPDIR:-/tmp/localstack}:/tmp/localstack”
      - “/var/run/docker.sock:/var/run/docker.sock”

Then we have to create the AmazonS3 client like that:

1
2
3
4
5
6
7
8
9
10
11
12

@Value(“${aws.s3.endpoint-url}”)
private String endpointUrl;
    
@Bean
AmazonS3 amazonS3() {
    EndpointConfiguration endpointConfiguration = new EndpointConfiguration(endpointUrl,
                Regions.US_EAST_1.getName());
                
    return AmazonS3ClientBuilder.standard()
                                .withEndpointConfiguration(endpointConfiguration)
                                .withPathStyleAccessEnabled(true).build();
}

The value for aws.s3.endpoint-url will be defined in a property file with the value http://localhost:4572

References

[1] Link to the project in Github

comments powered by Disqus