我想要一个可以在 Azure 函数中运行的解决方案，所有内容都在内存中，并且不依赖于 servlet 或门户 API。于是我写了一篇：
package my.package;
import java.io.ByteArrayInputStream;
import java.io.ByteArrayOutputStream;
import java.io.File;
import java.io.IOException;
import java.io.InputStream;
import java.io.OutputStream;
import java.io.UnsupportedEncodingException;
import java.nio.charset.StandardCharsets;
import java.util.List;
import java.util.Map;
import org.apache.commons.fileupload.FileItem;
import org.apache.commons.fileupload.FileItemFactory;
import org.apache.commons.fileupload.FileItemHeaders;
import org.apache.commons.fileupload.FileUpload;
import org.apache.commons.fileupload.FileUploadException;
import org.apache.commons.fileupload.ParameterParser;
import org.apache.commons.fileupload.UploadContext;
public class MultipartParser {
private static class SimpleContext implements UploadContext {
private final byte[] request;
private final String contentType;
private SimpleContext(byte[] requestBody, String contentTypeHeader) {
this.request = requestBody;
this.contentType = contentTypeHeader;
}
@Override
public long contentLength() {
return request.length;
}
@Override
public String getCharacterEncoding() {
// The 'Content-Type' header may look like:
// multipart/form-data; charset=UTF-8; boundary="xxxx"
// in which case we can extract the charset, otherwise,
// just default to UTF-8.
ParameterParser parser = new ParameterParser();
parser.setLowerCaseNames(true);
String charset = parser.parse(contentType, ';').get("charset");
return charset != null ? charset : "UTF-8";
}
@Override
public int getContentLength() {
return request.length;
}
@Override
public String getContentType() {
return contentType;
}
@Override
public InputStream getInputStream() throws IOException {
return new ByteArrayInputStream(request);
}
}
/**
* A form field which stores the field or file data completely in
* memory. Will be limited by the maximum size of
* a byte array (about 2GB).
*/
private static class MemoryFileItem implements FileItem {
private String fieldName;
private String fileName;
private String contentType;
private boolean isFormField;
private FileItemHeaders headers;
private final ByteArrayOutputStream os = new ByteArrayOutputStream();
public MemoryFileItem(String fieldName, String contentType, boolean isFormField, String fileName) {
this.fieldName = fieldName;
this.contentType = contentType;
this.isFormField = isFormField;
this.fileName = fileName;
}
@Override
public void delete() {
}
/**
* Not cached, so only call once.
*/
@Override
public byte[] get() {
return os.toByteArray();
}
@Override
public String getContentType() {
return contentType;
}
@Override
public String getFieldName() {
return fieldName;
}
@Override
public InputStream getInputStream() throws IOException {
return new ByteArrayInputStream(get());
}
@Override
public String getName() {
return fileName;
}
@Override
public OutputStream getOutputStream() throws IOException {
return os;
}
@Override
public long getSize() {
return os.size();
}
@Override
public String getString() {
return new String(get(), StandardCharsets.UTF_8);
}
@Override
public String getString(String encoding) throws UnsupportedEncodingException {
return new String(get(), encoding);
}
@Override
public boolean isFormField() {
return isFormField;
}
@Override
public boolean isInMemory() {
return true;
}
@Override
public void setFieldName(String name) {
fieldName = name;
}
@Override
public void setFormField(boolean state) {
isFormField = state;
}
@Override
public void write(File file) throws Exception {
}
@Override
public FileItemHeaders getHeaders() {
return headers;
}
@Override
public void setHeaders(FileItemHeaders headers) {
this.headers = headers;
}
}
private static class MemoryFileItemFactory implements FileItemFactory {
@Override
public FileItem createItem(String fieldName, String contentType, boolean isFormField, String fileName) {
return new MemoryFileItem(fieldName, contentType, isFormField, fileName);
}
}
/**
* Assumes the request body really is multipart/form-data.
* Given the binary request body and the Content-Type header value,
* attempts to parse fields into a map from field name to list
* of FileItem objects.
*
* Everything is stored in memory and an individual item will only be limited
* by the maximum size of a byte array (about 2GB). It is recommended that the
* user sets a limit on maximum upload request size. Doing this will obviously
* differ by environment.
*
* Example:
* <code>
* var fields = MultipartParser.parseRequest(requestBody, contentTypeHeader);
* String firstName = fields.get("firstname").get(0).getString();
* byte[] profilePic = fields.get("picture").get(0).get();
* </code>
*
* @param requestBody The binary request body
* @param contentTypeHeader The string value of the Content-Type header.
* @return a map, with each entry having one or more values for that named field.
* @throws FileUploadException
*/
public static Map<String, List<FileItem>> parseRequest(byte[] requestBody, String contentTypeHeader)
throws FileUploadException {
FileUpload fileUpload = new FileUpload(new MemoryFileItemFactory());
return fileUpload.parseParameterMap(new SimpleContext(requestBody, contentTypeHeader));
}
}
以及与 Azure 函数一起使用的示例：
public class Function {
@FunctionName("doSomethingWithBinaryFile")
public HttpResponseMessage run(
@HttpTrigger(
name = "req",
methods = {HttpMethod.POST},
authLevel = AuthorizationLevel.ANONYMOUS)
HttpRequestMessage<byte[]> request,
final ExecutionContext context) throws IOException, FileUploadException {
Logger logger = context.getLogger();
byte[] body = request.getBody();
String contentTypeHeader = request.getHeaders().get("content-type");
byte[] fileBytes = MultipartParser.parseRequest(body, contentTypeHeader).get("file").get(0).get();
return request.createResponseBuilder(HttpStatus.OK).body(
transformBytesForExample(fileBytes)).build();
}
}