Mongo and GridFS Java Driver with MATLAB
1 view (last 30 days)
Show older comments
Marc Youcef
on 6 Jun 2019
Commented: Marc Youcef
on 19 Jun 2019
I am trying to upload and download pictures from MATLAB to MongoDB using the Mongo Java Driver.
I am following below tutorial on GridFS java:
In MATLAB I did :
javaaddpath("C:\Users\****\Documents\jar_files\mongo-java-driver-3.11.0-beta3.jar")
import com.mongodb.Block;
import com.mongodb.client.MongoClients;
import com.mongodb.client.MongoClient;
import com.mongodb.client.MongoDatabase;
import com.mongodb.client.gridfs.*;
import com.mongodb.client.gridfs.model.*;
import org.bson.Document;
import org.bson.types.ObjectId;
import java.io.*;
import java.nio.file.Files;
import java.nio.charset.StandardCharsets;
import com.mongodb.client.model.Filters.eq;
mongoClient = MongoClients.create();
myDatabase = mongoClient.getDatabase("TestData");
gridFSFilesBucket = GridFSBuckets.create(myDatabase, "PicturesTest");
streamToUploadFrom = FileInputStream(File("C:\Users\****\Pictures\MyPicture.png"));
Until that point it works but then I can't do the following line:
fileId = gridFSBucket.uploadFromStream("mongodb-tutorial", streamToUploadFrom)
I get follwing error message ;
Undefined variable "gridFSBucket" or class "gridFSBucket.uploadFromStream".
I do not get why it is not able to find gridFSBucket function in the java driver.
I tried to play with the jar files it worked once but is not working anymore (do not know what I did exactly)...
Any help is much appreciated.
Thanks.
0 Comments
Accepted Answer
Debasish Samal
on 7 Jun 2019
The error occurs because there is no such object "gridFSBucket".
You need to create an object "gridFSBucket" of type GridFSBucket. Just add this line to your code:
gridFSBucket = GridFSBuckets.create(myDatabase);
Then add this:
fileId = gridFSBucket.uploadFromStream("mongodb-tutorial", streamToUploadFrom);
More Answers (1)
Marc Youcef
on 7 Jun 2019
Edited: Marc Youcef
on 7 Jun 2019
2 Comments
Martijn
on 12 Jun 2019
Hello Marc,
While your code is correct, it can be slightly optimized, there is no need to actually copy the streams, ImageIO.read should be able to read from the download stream directly:
%% Setup
% Load the driver
javaaddpath mongo-java-driver-3.9.1.jar
%% Upload
% Connect to the server and database
mongoClient = com.mongodb.client.MongoClients.create();
myDatabase = mongoClient.getDatabase("TestDb");
% Get GridFS Bucket
gridFSFilesBucket = com.mongodb.client.gridfs.GridFSBuckets.create(myDatabase, "PicturesTest");
% Open local file
streamToUploadFrom = java.io.FileInputStream(java.io.File(fullfile(pwd,"spy.png")));
% Upload to the server
fileId = gridFSFilesBucket.uploadFromStream("mongodb-tutorial", streamToUploadFrom);
% Close the local file
streamToUploadFrom.close();
%% Download
% Connect to the server and database
mongoClient = com.mongodb.client.MongoClients.create('mongodb://192.168.99.105:27017');
myDatabase = mongoClient.getDatabase("TestDb");
% Get GridFS Bucket
gridFSFilesBucket = com.mongodb.client.gridfs.GridFSBuckets.create(myDatabase, "PicturesTest");
% Open the download stream
downloadStream = gridFSFilesBucket.openDownloadStream('mongodb-tutorial');
% Download from the stream as image
javaImage = javax.imageio.ImageIO.read(downloadStream);
% Close the download stream
downloadStream.close
% Process the Java image into a MATLAB image
% Determine the image height and width
nrows = javaImage.getHeight; ncols = javaImage.getWidth;
% Convert to MATLAB image
% In Java these images are typically stored in a row first BGR interleaved
% format, so that is BGR value of top-left pixel followed by BGR value
% of the pixel to the right of it, etc. until the end of the line, then BGR
% value of the first pixel on the second line, followed by BGR value of
% pixel to the right of it, etc.
% In MATLAB this is different, we actually store by column instead of row
% first and the different RGB channels are actually in the last dimension
% instead of the first.
% First reshape the Java image into a format where we have access to the
% separate dimensions: BGR x width x height
data = reshape(typecast(javaImage.getData.getDataStorage, 'uint8'), [], ncols, nrows);
% Re-order those to get MATLAB format: height x width x BGR
matImg = permute(data,[3 2 1]);
% Re-order BGR into RGB
matImg2=matImg(:,:,[3 2 1]);
% Display Image
imagesc(matImg2);
Regards,
Martijn
See Also
Categories
Find more on Data Import and Analysis in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!