ome.api

Interface IContainer

All Superinterfaces:

ServiceInterface

All Known Implementing Classes:

PojosImpl

public interface IContainer
extends ServiceInterface

Provides methods for dealing with the core "Pojos" of OME. Included are: Projects, Datasets, Images.

Read API

The names of the methods correlate to how the function operates:

• load: start at container objects and work down toward the leaves, returning hierarchy (Proejct->Dataset->Image
• find: start at leave objects and work up to containers, returning hierarchy
• get: retrieves only leaves in the hierarchy (currently only Images)

Options Mechanism

The options are used to add some constraints to the generic method e.g. load hierarchy trees images for a given user. This mechanism should give us enough flexibility to extend the API if necessary, e.g. in some cases we might want to retrieve the images with or without annotations

Most methods take such an options map which is built on the client-side using an option builder. The currently supported options are:

• annotator(Integer): If key exists but value null, annotations are retrieved for all objects in the hierarchy where they exist; if a valid experimenterID, annotations are only retrieved for that user. May not be used be all methods. Default: all annotations
• leaves(Boolean): if FALSE omits images from the returned hierarchy. May not be used by all methods. Default: true
• experimenter(Integer): inables filtering on a per-experimenter basis. This option has a method-specific (and possibly context-specific) meaning. Please see the individual methods.
• group(Integer): enables filtering on a per-group basis. The experimenter value is ignored if present and instead a similar filtering is done using all experimenters in the given group.

  Write API

  As outlined in TODO, the semantics of the Omero write API are based on three rules:

  1. IObject-valued fields for which isLoaded() returns false are assumed filtered
  2. Collection-valued fields that are null are assumed filtered
  3. Collection-valued fields for which getDetails().isFiltered(String collectionName) returns true are assumed filtered. TODO: should we accept isFiltered for all fields?
  In each of these cases, the server will reload that given field before attempting to save the graph.

  For all write calls, the options map (see below) must contain the userId and the userGroupId for the newly created objects. TODO umask.

Since:

OME1.0

Version:

1.0 (Internal version: $Revision$ $Date$)

Author:

Jean-Marie Burel j.burel@dundee.ac.uk,
Josh Moore josh.moore@gmx.de

To do:

possibly move optionBuilder to the common code. ome.common.utils, possibly move map description to marker interface with a see also link, add throws statements where necessary (IllegalArgument, ...)

Method Summary

Methods

Modifier and Type	Method and Description
<T extends IObject> T	createDataObject(T object, Parameters options) Creates the specified data object.
IObject[]	createDataObjects(IObject[] dataObjects, Parameters options) convenience method to save network calls.
<T extends IObject> Set<IObject>	findContainerHierarchies(Class<T> rootNodeType, Set<Long> imagesIds, Parameters options) Retrieves hierarchy trees in various hierarchies that contain the specified Images.
Map	getCollectionCount(String type, String property, Set<Long> ids, Parameters options) Counts the number of members in a collection for a given object.
<T extends IObject> Set<Image>	getImages(Class<T> rootNodeType, Set<Long> rootNodeIds, Parameters options) Retrieve a user's (or all users') images within any given container.
Set<Image>	getImagesByOptions(Parameters options) Retrieve images by options.
Map<Long,Map<Boolean,List<Long>>>	getImagesBySplitFilesets(Map<Class<? extends IObject>,List<Long>> included, Parameters options) Given a list of IDs of certain entity types, calculates which filesets are split such that a non-empty proper subset of their images are referenced, directly or indirectly, as being included.
Set<Image>	getUserImages(Parameters options) Retrieve a user's images.
ILink[]	link(ILink[] dataObjectLinks, Parameters options) Convenience method for creating links.
<T extends IObject> Set<IObject>	loadContainerHierarchy(Class<T> rootNodeType, Set<Long> rootNodeIds, Parameters options) Retrieves hierarchy trees rooted by a given node (unless orphan is specified -- See below) This method also retrieves the Experimenters linked to the objects in the tree.
Collection	retrieveCollection(IObject dataObject, String collectionName, Parameters options) Retrieves a collection with all members initialized ("loaded").
void	unlink(ILink[] dataOjectLinks, Parameters options) Removes links between OmeroDataObjects e.g Project-Dataset, Dataset-Image Note that the objects themselves aren't deleted, only the Link objects.
<T extends IObject> T	updateDataObject(T dataObject, Parameters options) Updates a data object.
IObject[]	updateDataObjects(IObject[] dataObjects, Parameters options) convenience method to save network calls.

Method Detail

loadContainerHierarchy

<T extends IObject> Set<IObject> loadContainerHierarchy(Class<T> rootNodeType,
                                                      Set<Long> rootNodeIds,
                                                      Parameters options)

Retrieves hierarchy trees rooted by a given node (unless orphan is specified -- See below)

This method also retrieves the Experimenters linked to the objects in the tree. Similarly, all Images will be linked to their Pixel objects if included.

Note that objects are never duplicated. For example, if an Experimenter owns all the objects in the retrieved tree, then those objects will be linked to the same instance of Experimenter. Or if an Image is contained in more than one Dataset in the retrieved tree, then all enclosing Dataset objects will point to the same Image object. And so on.

Parameters:

rootNodeType - The type of the root node. Can be Project, Dataset, Screen or Plate. Cannot be null.

rootNodeIds - The ids of the root nodes. Can be null if an Experimenter is specified in options, otherwise an Exception is thrown to prevent all images in the entire database from being downloaded.

options - Parameters as above. annotator, leaves, orphan, acquisition data used. acquisition data is only relevant for images and taken into account if the images are loaded. If rootNodeIds==null, experimenter|group must be set and filtering will be applied at the Class-level; e.g. to retrieve a user's Projects, or user's Datasets. If rootNodeIds!=null, the result will be filtered by the experimenter|group at the Image and intermediate levels if available. Due to the amount of data potentially linked a Screen/Plate, the leaves option is not taken into account. when the root node is a Screen. orphan implies that objects which are not contained in an object of rootNodeType should also be returned.

Returns:

a set of hierarchy trees. The requested node as root and all of its descendants. The type of the returned value will be rootNodeType, unless orphan is specified in which case objects of type rootNodeType and below may be returned.

To do:

should it be applied at all levels?

findContainerHierarchies

<T extends IObject> Set<IObject> findContainerHierarchies(Class<T> rootNodeType,
                                                        Set<Long> imagesIds,
                                                        Parameters options)

Retrieves hierarchy trees in various hierarchies that contain the specified Images.

This method will look for all the containers containing the specified Images and then for all containers containing those containers and on up the container hierarchy.

This method returns a Set with all root nodes that were found. Every root node is linked to the found objects and so on until the leaf nodes, which are Image objects. Note that the type of any root node in the returned set can be the given rootNodeType, any of its containees or an image.

For example, say that you pass in the ids of six Images: i1, i2, i3, i4, i5, i6. If the P/D/I hierarchy in the DB looks like this:

                   __p1__
                  /      \    
                _d1_    _d2_      d3
               /    \  /    \     |
              i1     i2     i3    i4    i5  i6
 

Then the returned set will contain p1, d3, i5, i6. All objects will be properly linked up.

Finally, this method will only retrieve the nodes that are connected in a tree to the specified leaf image nodes. Back to the previous example, if d1 contained image img500, then the returned object would not contain img500. In a similar way, if p1 contained ds300 and this dataset weren't linked to any of the i1, i2, i3, i4, i5, i6 images, then ds300 would not be part of the returned tree rooted by p1.

Parameters:

rootNodeType - top-most type which will be searched for Can be Project. Not null.

imagesIds - Contains the ids of the Images that sit at the bottom of the trees. Not null.

options - Parameters as above. annotator used. experimenter|group may be applied at the top-level only or at each level in the hierarchy, but will not apply to the leaf (Image) level.

Returns:

A Set with all root nodes that were found.

To do:

decide on use of experimenter option

getImages

<T extends IObject> Set<Image> getImages(Class<T> rootNodeType,
                                       Set<Long> rootNodeIds,
                                       Parameters options)

Retrieve a user's (or all users') images within any given container. For example, all images in project, applying temporal filtering or pagination.

Parameters:

rootNodeType - A Class which will have its hierarchy searched for Images. Not null. TODO types?

rootNodeIds - A set of ids of type rootNodeType Not null.

options - Parameters as above. No notion of leaves. experimenter|group apply at the Image level. OPTIONS: - startTime and/or endTime should be Timestamp.valueOf("YYYY-MM-DD hh:mm:ss.ms");

limit and offset are applied at the Image-level. That is, calling with Dataset.class, limit == 10 and offset == 0 will first perform one query to get an effective set of rootNodeIds, then getImages will be called with an effective rootNodeType of Image.class and the new ids.

. acquisition data is only relevant for images.

Returns:

A set of images.

See Also:

ome.util.builders.PojoOptions#paginate(int, int), ome.util.builders.PojoOptions#startTime(java.sql.Timestamp), ome.util.builders.PojoOptions#endTime(java.sql.Timestamp)

getImagesByOptions

Set<Image> getImagesByOptions(Parameters options)

Retrieve images by options.

Parameters:

options - Parameters as above. No notion of leaves. experimenter|group apply at the Image level and must be present. OPTIONS: - startTime and/or endTime should be Timestamp.valueOf("YYYY-MM-DD hh:mm:ss.ms"). acquisition data is only relevant for images.

Returns:

A set of images.

getUserImages

Set<Image> getUserImages(Parameters options)

Retrieve a user's images.

Parameters:

options - Parameters as above. No notion of leaves. experimenter|group apply at the Image level and must be present.

Returns:

A set of images.

getCollectionCount

Map getCollectionCount(String type,
                     String property,
                     Set<Long> ids,
                     Parameters options)

Counts the number of members in a collection for a given object. For example, if you wanted to retrieve the number of Images contained in a Dataset you would pass TODO.

Parameters:

type - The fully-qualified classname of the object to be tested

property - Name of the property on that class, omitting getters and setters.

ids - Set of Longs, the ids of the objects to test

options - Parameters. Unused.

Returns:

A map from id integer to count integer

getImagesBySplitFilesets

Map<Long,Map<Boolean,List<Long>>> getImagesBySplitFilesets(Map<Class<? extends IObject>,List<Long>> included,
                                                         Parameters options)

Given a list of IDs of certain entity types, calculates which filesets are split such that a non-empty proper subset of their images are referenced, directly or indirectly, as being included. The return value lists both the fileset IDs and the image IDs in ascending order, the image ID lists separated by if they were included. Warning: following discussion in trac ticket 11019 the return type may be changed.

Parameters:

included - the entities included

options - parameters, presently ignored

Returns:

the partially included filesets

retrieveCollection

Collection retrieveCollection(IObject dataObject,
                            String collectionName,
                            Parameters options)

Retrieves a collection with all members initialized ("loaded"). This is useful when a collection has been nulled in a previous query.

Parameters:

dataObject - Can be "unloaded".

collectionName - public final static String from the IObject.class

options - Parameters. Unused.

Returns:

An initialized collection.

createDataObject

<T extends IObject> T createDataObject(T object,
                                     Parameters options)

Creates the specified data object.

A placeholder parent object is created if the data object is to be put in a collection.

For example, if the object is a Dataset, we first create a Project as parent then we set the Dataset parent as follows: //pseudo-code TODO Project p = new Project(id,false); dataset.addProject(p); then for each parent relationship a DataObject ILink#link is created.

Parameters:

object - IObject. Supported: Project, Dataset, Annotation, Group, Experimenter. Not null.

Returns:

the created object

createDataObjects

IObject[] createDataObjects(IObject[] dataObjects,
                          Parameters options)

convenience method to save network calls. Loops over the array of IObjects calling createDataObject.

Parameters:

dataObjects - Array of Omero IObjects

options - Parameters as above.

See Also:

createDataObject

unlink

void unlink(ILink[] dataOjectLinks,
          Parameters options)

Removes links between OmeroDataObjects e.g Project-Dataset, Dataset-Image Note that the objects themselves aren't deleted, only the Link objects.

Parameters:

dataObjectLinks - Not null.

options - Parameters as above.

link

ILink[] link(ILink[] dataObjectLinks,
           Parameters options)

Convenience method for creating links. Functionality also availeble from createDataObject(IObject, Parameters)

Parameters:

dataObjectLinks - Array of links to be created.

options - Parameters as above.

Returns:

the created links

updateDataObject

<T extends IObject> T updateDataObject(T dataObject,
                                     Parameters options)

Updates a data object.

To link or unlink objects to the specified object, we should call the methods link or unlink. TODO Or do we use for example dataset.setProjects(set of projects) to add. Tink has to be set as follows dataset->project and project->dataset. Alternatively, you can make sure that the collection is exactly how it should be in the database. If you can't guarantee this, it's best to send all your collections back as null

Parameters:

dataObject - Pojos-based IObject. Supported: supported: Project, Dataset, Annotation, Group, Experimenter.

options - Parameters as above.

Returns:

created data object

updateDataObjects

IObject[] updateDataObjects(IObject[] dataObjects,
                          Parameters options)

convenience method to save network calls. Loops over the array of IObjects calling updateDataObject.

Parameters:

dataObjects -

options - Parameters as above.

Returns:

created data objects.

See Also:

updateDataObject(IObject, Parameters)