OMERO MATLAB language bindings ============================== See |DevelopingOmeroClients| and |OmeroModel|, for an introduction to **Object**. .. highlight:: matlab Installing the OMERO.matlab toolbox ----------------------------------- - Download the latest released version from the :downloads:`download page <>`. - Unzip the directory anywhere on your system. - In MATLAB, move to the newly unzipped directory and run ``loadOmero;``. - The MATLAB files are now on your path, and the necessary jars are on your java classpath. You can change directories and still have access to OMERO. Once OMERO.matlab is installed, the typical workflow is: #. :ref:`connection_init` #. :ref:`connection_keepalive` #. :ref:`unsecure_client` (optional) #. Do some work (load objects, work with them, upload to the server…) #. :ref:`connection_close` #. :ref:`unload` (optional) As a quickstart example, the following lines create a secure connection to a server, read a series of images and close the connection. :: client = loadOmero(servername, port); session = client.createSession(user, password); client.enableKeepAlive(60); images = getImages(session, ids); client.closeSession(); Examples of usage of the OMERO.matlab toolbox are provided in the :sourcedir:`training examples ` directory. Configuring the OMERO.matlab connection --------------------------------------- .. _connection_init: Creating a connection ^^^^^^^^^^^^^^^^^^^^^ As described under |OmeroClients|, there are several ways to configure your connection to an OMERO server. OMERO.matlab comes with a few conveniences for making this work. If you run ``client = loadOmero();`` (i.e. loadOmero with an output argument), then OMERO.matlab will try to configure the ``omero.client`` object for you. First, it checks the :envvar:`ICE_CONFIG` environment variable. If set, it will let the ``omero.client`` constructor initialize itself. Otherwise, it looks for the file :file:`ice.config` in the current directory. The OMERO.matlab toolbox comes with a default :file:`ice.config` file pointing at ``localhost``. To use this configuration file, you should replace ``localhost`` by your server address. Alternatively, you can pass the same parameters to ``loadOmero;`` that you would pass to ``omero.client``:: >> omero_client_1 = loadOmero('localhost'); >> omero_client_2 = omero.client('localhost'); Or, if you want a session created directly, the following are equivalent: :: >> [client1, session1] = loadOmero('localhost'); >> client2 = loadOmero('localhost'); >> session2 = client2.createSession() .. _connection_keepalive: Keeping your session alive ^^^^^^^^^^^^^^^^^^^^^^^^^^ For executing any long running task, you will need a background thread which keeps your session alive. If you are familiar with MATLAB ``Timers`` you can use :source:`omeroKeepAlive.m ` directly or modify it to your liking. :: >> [c,s] = loadOmero; >> t = omeroKeepAlive(c); % Create a 60-second timer and starts it >> … >> delete(t); % Disable the keep-alive Alternatively, you can use the Java-based ``enableKeepAlive`` method, but it is not configurable from within MATLAB:: c.enableKeepAlive(60); % Call session.keepAlive() every 60 seconds c.closeSession(); % Close session to end the keep-alive .. _unsecure_client: Creating an unencrypted session ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ Once a session has been created, if you want to speed up the data transfer, you can create and use an unencrypted session as:: unsecureClient = client.createClient(false); sessionUnencrypted = unsecureClient.getSession(); .. _connection_close: Closing your connection ^^^^^^^^^^^^^^^^^^^^^^^ When you are done with OMERO, it is critical that you close your connection to save resources:: client.closeSession(); clear client1; clear session1; If you created an unencrypted session, you will need to close the unsecure session as well:: client.closeSession(); unsecureClient.closeSession(); .. _unload: Unloading OMERO ^^^^^^^^^^^^^^^ Then if you would like, you can unload OMERO as well:: unloadOmero(); You may see the following warning when unloading OMERO:: >> unloadOmero() Warning: Objects of omero/client class exist - not clearing java > In javaclasspath>doclear at 377 In javaclasspath>local_javapath at 194 In javaclasspath at 105 In javarmpath at 48 In unloadOmero at 75 =============================================================== While unloading OMERO, found java objects left in workspace. Please remove with 'clear ' and then run 'unloadOmero' again. Printing all objects... =============================================================== Name Size Bytes Class Attributes c 1x1 omero.client Closing session(s) for 1 found client(s): c This means that there is still an OMERO.matlab object in your workspace. If not listed, use ``whos`` to find such objects, and ``clear`` to remove them. After that, run ``unloadOmero()`` again:: >> clear c >> unloadOmero() .. warning:: You should also unload OMERO before installing a new version of OMERO.matlab or calling ``loadOmero`` again. If you need to create another session without unloading/loading OMERO again, use the ``omero.client`` object directly:: >> [c,s] = loadOmero(arg1,arg2); >> c = omero.client(arg3,arg4); >> s = c.createSession(); Reading data ------------ The ``IContainer`` service provides methods to load the data management hierarchy in OMERO -- projects, datasets… A list of examples follows indicating how to load projects, datasets, screens… - **Projects** The projects owned by the user currently logged in can be retrieved using the :source:`getProjects ` function:: projects = getProjects(session) If the project identifiers are known, they can be specified as:: projects = getProjects(session, ids) If the projects contain datasets, the datasets will automatically be loaded:: for j = 1 : numel(projects) datasetsList = projects(j).linkedDatasetList; for i = 0:datasetsList.size()-1, d = datasetsList.get(i); end end If the datasets contain images, the images will automatically be loaded:: imageList = projects(1).linkedDatasetList.get(0).linkedImageList; To avoid loading the whole graph (projects, datasets, images), pass `false` as a second optional argument. Only datasets will be loaded:: unloadedProjects = getProjects(session, ids, false) To return the orphaned datasets as well as the projects, you can query the second output argument of :source:`getProjects `:: [projects, orphanedDatasets] = getProjects(session) [unloadedProjects, unloadedOrphanedDatasets] = getProjects(session, [], false) - **Datasets** The datasets owned by the user currently logged in can be retrieved using the :source:`getDatasets ` function:: datasets = getDatasets(session) If the dataset identifiers are known, they can be specified as:: datasets = getDatasets(session, ids) If the datasets contain images, the images will automatically be loaded:: imageList = datasets(1).linkedImageList; To avoid loading the images, pass `false` as a second optional argument:: unloadedDatasets = getDatasets(session, ids, false) - **Images** All the images owned by the user currently logged in can be retrieved using the :source:`getImages ` function:: images = getImages(session) If the image identifiers are known, they can be specified as:: images = getImages(session, ids) All the images contained in a subset of datasets of known identifiers ``datasetsIds`` can be returned using:: datasetImages = getImages(session, 'dataset', datasetsIds) All the images contained in all the datasets under a subset of projects of known identifiers ``projectIds`` can be returned using:: projectImages = getImages(session, 'project', projectIds) The ``Image``-``Pixels`` model implies you need to use the ``Pixels`` objects to access valuable data about the ``Image``:: pixels = image.getPrimaryPixels(); sizeZ = pixels.getSizeZ().getValue(); % The number of z-sections. sizeT = pixels.getSizeT().getValue(); % The number of timepoints. sizeC = pixels.getSizeC().getValue(); % The number of channels. sizeX = pixels.getSizeX().getValue(); % The number of pixels along the X-axis. sizeY = pixels.getSizeY().getValue(); % The number of pixels along the Y-axis. - **Screens** The screens owned by the user currently logged in can be retrieved using the :source:`getScreens ` function:: screens = getScreens(session) If the screen identifiers are known, they can be specified as:: screens = getScreens(session, ids) Note that the wells are not loaded. The plate objects can be accessed using:: for j = 1 : numel(screens), platesList = screens(j).linkedPlateList; for i = 0:platesList.size()-1, plate = platesList.get(i); plateAcquisitionList = plate.copyPlateAcquisitions(); for k = 0:plateAcquisitionList.size()-1, pa = plateAcquisitionList.get(i); end end To return the orphaned plates as well as the screens, you can query the second output argument of :source:`getScreens `:: [screens, orphanedPlates] = getScreens(session) [unloadedScreens, unloadedOrphanedPlates] = getScreens(session, [], false) - **Plates** The plates owned by the user currently logged in can be retrieved using the :source:`getPlates ` function:: plates = getPlates(session) If the plate identifiers are known, they can be specified as:: plates = getPlates(session, ids) - **Wells** Given a plate identifier, the wells can be loaded using the ``findAllByQuery`` method:: wellList = session.getQueryService().findAllByQuery( ['select well from Well as well '... 'left outer join fetch well.plate as pt '... 'left outer join fetch well.wellSamples as ws '... 'left outer join fetch ws.plateAcquisition as pa '... 'left outer join fetch ws.image as img '... 'left outer join fetch img.pixels as pix '... 'left outer join fetch pix.pixelsType as pt '... 'where well.plate.id = ', num2str(plateId)], []); for j = 0:wellList.size()-1, well = wellList.get(j); wellsSampleList = well.copyWellSamples(); well.getId().getValue() for i = 0:wellsSampleList.size()-1, ws = wellsSampleList.get(i); ws.getId().getValue() pa = ws.getPlateAcquisition(); end end Raw data access --------------- You can retrieve data, plane by plane or retrieve a stack. - **Plane** The plane of an input image at coordinates (z, c, t) can be retrieved using the :source:`getPlane ` function:: plane = getPlane(session, image, z, c, t); Alternatively, the image identifier can be passed to the function:: plane = getPlane(session, imageID, z, c, t); - **Tile** The tile of an input image at coordinates (z, c, t) originated at (x, y) and of dimensions (w, h) can be retrieved using the :source:`getTile ` function:: tile = getTile(session, image, z, c, t, x, y, w, h); Alternatively, the image identifier can be passed to the function:: tile = getTile(session, imageID, z, c, t, x, y, w, h); - **Stack** The stack of an input image at coordinates (c, t) can be retrieved using the :source:`getStack ` function:: stack = getStack(session, image, c, t); Alternatively, the image identifier can be passed to the function:: stack = getStack(session, imageID, c, t); All the methods described above will internally initialize a raw pixels store to retrieve the pixels data and close this store at the end of the call. This is inefficient when multiple planes/tiles/stacks need to be retrieved. For each function, it is possible to initialize a pixels store and pass this store directly to the pixel retrieval function, e.g.:: [store, pixels] = getRawPixelsStore(session, image); for z = 0 : sizeZ - 1 for c = 0 : sizeC - 1 for t = 0 : sizeT - 1 plane = getPlane(pixels, store, z, c, t); end end end store.close(); - **Hypercube** This is useful when you need the ``Pixels`` intensity. :: % Create the store to load the stack. No access via the gateway store = session.createRawPixelsStore(); % Indicate the pixels set you are working on store.setPixelsId(pixelsId, false); % Offset values in each dimension XYZCT offset = java.util.ArrayList; offset.add(java.lang.Integer(0)); offset.add(java.lang.Integer(0)); offset.add(java.lang.Integer(0)); offset.add(java.lang.Integer(0)); offset.add(java.lang.Integer(0)); size = java.util.ArrayList; size.add(java.lang.Integer(sizeX)); size.add(java.lang.Integer(sizeY)); size.add(java.lang.Integer(sizeZ)); size.add(java.lang.Integer(sizeC)); size.add(java.lang.Integer(sizeT)); % Indicate the step in each direction, % step = 1, will return values at index 0, 1, 2. % step = 2, values at index 0, 2, 4… step = java.util.ArrayList; step.add(java.lang.Integer(1)); step.add(java.lang.Integer(1)); step.add(java.lang.Integer(1)); step.add(java.lang.Integer(1)); step.add(java.lang.Integer(1)); % Retrieve the data store.getHypercube(offset, size, step); % Close the store store.close(); .. seealso:: :source:`RawDataAccess.m ` Example script showing methods to retrieve the pixel data from an image Annotations ----------- - **Reading annotations by ID** If the identifier of the annotation of a given type is known, the annotation can be retrieved from the server using the generic :source:`getAnnotations ` function:: tagAnnotations = getAnnotations(session, 'tag', tagIds); Shortcut functions are available for the main object and annotation types, e.g. to retrieve tag annotations:: tagAnnotations = getTagAnnotations(session, tagIds); - **Reading annotations linked to an object** The annotations of a given type linked to a given object can be retrieved using the generic :source:`getObjectAnnotations ` function:: tagAnnotations = getObjectAnnotations(session, 'tag', 'image', imageIds); Shortcut functions are available for the main object and annotation types, e.g. to retrieve the tag annotations linked to images:: tagAnnotations = getImageTagAnnotations(session, imageIds); Annotations can be filtered by namespace. To include only annotations with a given namespace ``ns``, use the ``include`` parameter/key value:: tagAnnotations = getImageTagAnnotations(session, imageIds, 'include', ns); To exclude all annotations with a given namespace ``ns``, use the ``exclude`` parameter/key value:: tagAnnotations = getImageTagAnnotations(session, imageIds, 'exclude', ns); By default, only the annotations owned by the session owner are returned. To specify the owner of the annotations, use the ``owner`` paramter/key value pair. For instance to return all tag annotations owned by user 5:: tagAnnotations = getImageTagAnnotations(session, imageIds, 'owner', 5); To retrieve all annotations independently of their owner, use -1 as the owner identifier:: tagAnnotations = getImageTagAnnotations(session, imageIds, 'owner', -1); - **Reading file annotations** The content of a file annotation can be downloaded to local disk using the :source:`getFileAnnotationContent ` function. If the file annotation has been retrieved from the server as ``fileAnnotation``, then the content of its ``OriginalFile`` can be downloaded under ``target_file`` using:: getFileAnnotationContent(session, fileAnnotation, target_file); Alternatively, if only the identifier of the file annotation ``faId`` is known:: getFileAnnotationContent(session, faId, target_file); - **Writing and linking annotations** New annotations can be created using the corresponding ``write*Annotation`` function. For example, to create a new tag annotation with value ``tag_name``:: tagAnnotation = writeTagAnnotation(session, tag_name); To create a file annotations from the content of a ``local_file_path``:: fileAnnotation = writeFileAnnotation(session, local_file_path); Existing annotations can be linked to existing objects on the server using the :source:`linkAnnotation ` function. For example, to link a tag annotation and a file annotation to the image ``image_id``:: link1 = linkAnnotation(session, tagAnnotation, 'image', image_id); link2 = linkAnnotation(session, fileAnnotation, 'image', image_id); For existing file annotations, it is possible to replace the content of the original file without having to recreate a new file annotation using the :source:`updateFileAnnotation ` function. If the file annotation has been retrieved from the server as ``fileAnnotation``, then the content of its ``OriginalFile`` can be replaced by the content of ``local_file_path`` using:: updateFileAnnotation(session, fileAnnotation, local_file_path); .. seealso:: :source:`WriteData.m ` Example script showing methods to write, link and retrieve annotations Writing data ------------ - **Create a Dataset** and link it to an existing project. :: dataset = omero.model.DatasetI; dataset.setName(omero.rtypes.rstring(char('name dataset'))); dataset.setDescription(omero.rtypes.rstring(char('description dataset'))); % Link Dataset and Project link = omero.model.ProjectDatasetLinkI; link.setChild(dataset); link.setParent(omero.model.ProjectI(projectId, false)); session.getUpdateService().saveAndReturnObject(link); How to use OMERO tables ----------------------- - **Create a table**. In the following example, a table is created with 2 columns. :: name = char(java.util.UUID.randomUUID()); columns = javaArray('omero.grid.Column', 2) columns(1) = omero.grid.LongColumn('Uid', 'testLong', []); valuesString = javaArray('java.lang.String', 1); columns(2) = omero.grid.StringColumn('MyStringColumn', '', 64, valuesString); % Create a new table. table = session.sharedResources().newTable(1, name); % Initialize the table table.initialize(columns); % Add data to the table. data = javaArray('omero.grid.Column', 2); data(1) = omero.grid.LongColumn('Uid', 'test Long', [2]); valuesString = javaArray('java.lang.String', 1); valuesString(1) = java.lang.String('add'); data(2) = omero.grid.StringColumn('MyStringColumn', '', 64, valuesString); table.addData(data); file = table.getOriginalFile(); % if you need to interact with the table - **Read the contents of the table**. :: of = omero.model.OriginalFileI(file.getId(), false); tablePrx = session.sharedResources().openTable(of); % Read headers headers = tablePrx.getHeaders(); for i=1:size(headers, 1), headers(i).name; % name of the header % Do something end % Depending on the size of table, you may only want to read some blocks. cols = [0:size(headers, 1)-1]; % The number of columns you wish to read. rows = [0:tablePrx.getNumberOfRows()-1]; % The number of rows you wish to read. data = tablePrx.slice(cols, rows); % Read the data. c = data.columns; for i=1:size(c), column = c(i); % Do something end tablePrx.close(); % Important to close when done. ROIs ---- To learn about the model, see the :model_doc:`developers guide to the ROI model `. Note that annotations can be linked to ROI. - **Creating ROI** This example creates a ROI with two shapes, a rectangle and an ellipse, and attaches it to an image:: % First create a rectangular shape. rectangle = createRectangle(0, 0, 10, 20); % Indicate on which plane to attach the shape setShapeCoordinates(rectangle, 0, 0, 0); % First create an ellipse shape. ellipse = createEllipse(0, 0, 10, 20); % Indicate on which plane to attach the shape setShapeCoordinates(ellipse, 0, 0, 0); % Create the roi. roi = omero.model.RoiI; % Attach the shapes to the roi, several shapes can be added. roi.addShape(rectangle); roi.addShape(ellipse); % Link the roi and the image roi.setImage(omero.model.ImageI(imageId, false)); % Save iUpdate = session.getUpdateService(); roi = iUpdate.saveAndReturnObject(roi); % Check that the shape has been added. numShapes = roi.sizeOfShapes; for ns = 1:numShapes shape = roi.getShape(ns-1); end .. seealso:: :sourcedir:`ROI utility functions ` OMERO.matlab functions for creating and managing Shape and ROI objects. - **Retrieving ROIs linked to an image** :: service = session.getRoiService(); roiResult = service.findByImage(imageId, []); rois = roiResult.rois; n = rois.size; shapeType = ''; for thisROI = 1:n roi = rois.get(thisROI-1); numShapes = roi.sizeOfShapes; for ns = 1:numShapes shape = roi.getShape(ns-1); if (isa(shape, 'omero.model.Rect')) rectangle = shape; rectangle.getX().getValue() elseif (isa(shape, 'omero.model.Ellipse')) ellipse = shape; ellipse.getCx().getValue() elseif (isa(shape, 'omero.model.Point')) point = shape; point.getX().getValue(); elseif (isa(shape, 'omero.model.Line')) line = shape; line.getX1().getValue(); end end end - **Removing a shape from ROI** :: // Retrieve the roi linked to an image service = session.getRoiService(); roiResult = service.findByImage(imageId, []); n = rois.size; for thisROI = 1:n roi = rois.get(thisROI-1); numShapes = roi.sizeOfShapes; for ns = 1:numShapes shape = roi.getShape(ns-1); % Remove the shape roi.removeShape(shape); end % Update the roi. roi = iUpdate.saveAndReturnObject(roi); end Deleting data ------------- It is possible to delete projects, datasets, images, ROIs… and objects linked to them depending on the specified options (see :doc:`/developers/Modules/Delete`). For example, images of known identifiers can be deleted from the server using the :source:`deleteImages ` function:: deleteImages(session, imageIds); .. seealso:: :source:`deleteProjects `, :source:`deleteDatasets `, :source:`deleteScreens `, :source:`deletePlates ` Utility functions to delete objects Rendering images ----------------- The :source:`RenderImages.m ` example script shows how to initialize the rendering engine and render an image. Creating Image -------------- The :source:`CreateImage.m ` example script shows how to create an image in OMERO. A similar approach can be applied when uploading an image. To upload individual planes onto the server, the data must be converted into a byte (int8) array first. If the ``Pixels`` object has been created, this conversion can done using the :source:`toByteArray ` function.