Package weka.core

Class Capabilities

  • All Implemented Interfaces:
    Serializable, Cloneable, RevisionHandler
    public class Capabilities
extends Object
implements Cloneable, Serializable, RevisionHandler
    A class that describes the capabilites (e.g., handling certain types of attributes, missing values, types of classes, etc.) of a specific classifier. By default, the classifier is capable of nothing. This ensures that new features have to be enabled explicitly.

    A common code fragment for making use of the capabilities in a classifier would be this: 

     public void buildClassifier(Instances instances) throws Exception {
   // can the classifier handle the data?
   getCapabilities().testWithFail(instances);
   ...
   // possible deletion of instances with missing class labels, etc.
    For only testing a single attribute, use this: 
       ...
   Attribute att = instances.attribute(0);
   getCapabilities().testWithFail(att);
   ...
    Or for testing the class attribute (uses the capabilities that are especially for the class): 
       ...
   Attribute att = instances.classAttribute();
   getCapabilities().testWithFail(att, true);
   ...
    Version:
    $Revision$
    Author:
    FracPete (fracpete at waikato dot ac dot nz)
    See Also:
    Serialized Form

    • Field Detail

      • PROPERTIES_FILE

        public static final String PROPERTIES_FILE
        the properties file for managing the tests
        See Also:
        Constant Field Values

    • Constructor Detail

      • Capabilities

        public Capabilities​(CapabilitiesHandler owner)
        initializes the capabilities for the given owner
        Parameters:
        owner - the object that produced this Capabilities instance

    • Method Detail

      • doNotCheckCapabilities

        public boolean doNotCheckCapabilities()
        Does owner implement CapabilitiesIgnorer and does it not want capability checking to be performed?

      • clone

        public Object clone()
        Creates and returns a copy of this object.
        Returns:
        a clone of this object

      • assign

        public void assign​(Capabilities c)
        retrieves the data from the given Capabilities object
        Parameters:
        c - the capabilities object to initialize with

      • and

        public void and​(Capabilities c)
        performs an AND conjunction with the capabilities of the given Capabilities object and updates itself
        Parameters:
        c - the capabilities to AND with

      • or

        public void or​(Capabilities c)
        performs an OR conjunction with the capabilities of the given Capabilities object and updates itself
        Parameters:
        c - the capabilities to OR with

      • supports

        public boolean supports​(Capabilities c)
        Returns true if the currently set capabilities support at least all of the capabilities of the given Capabilities object (checks only the enum!)
        Parameters:
        c - the capabilities to support at least
        Returns:
        true if all the requested capabilities are supported

      • supportsMaybe

        public boolean supportsMaybe​(Capabilities c)
        Returns true if the currently set capabilities support (or have a dependency) at least all of the capabilities of the given Capabilities object (checks only the enum!)
        Parameters:
        c - the capabilities (or dependencies) to support at least
        Returns:
        true if all the requested capabilities are supported (or at least have a dependency)

      • setOwner

        public void setOwner​(CapabilitiesHandler value)
        sets the owner of this capabilities object
        Parameters:
        value - the new owner

      • getOwner

        public CapabilitiesHandler getOwner()
        returns the owner of this capabilities object
        Returns:
        the current owner of this capabilites object

      • setMinimumNumberInstances

        public void setMinimumNumberInstances​(int value)
        sets the minimum number of instances that have to be in the dataset
        Parameters:
        value - the minimum number of instances

      • getMinimumNumberInstances

        public int getMinimumNumberInstances()
        returns the minimum number of instances that have to be in the dataset
        Returns:
        the minimum number of instances

      • capabilities

        public Iterator<Capabilities.Capability> capabilities()
        Returns an Iterator over the stored capabilities
        Returns:
        iterator over the current capabilities

      • dependencies

        public Iterator<Capabilities.Capability> dependencies()
        Returns an Iterator over the stored dependencies
        Returns:
        iterator over the current dependencies

      • enable

        public void enable​(Capabilities.Capability c)
        enables the given capability. Enabling NOMINAL_ATTRIBUTES also enables BINARY_ATTRIBUTES, UNARY_ATTRIBUTES and EMPTY_NOMINAL_ATTRIBUTES. Enabling BINARY_ATTRIBUTES also enables UNARY_ATTRIBUTES and EMPTY_NOMINAL_ATTRIBUTES. Enabling UNARY_ATTRIBUTES also enables EMPTY_NOMINAL_ATTRIBUTES. But NOMINAL_CLASS only enables BINARY_CLASS, since normal schemes in Weka don't work with datasets that have only 1 class label (or none).
        Parameters:
        c - the capability to enable

      • enableDependency

        public void enableDependency​(Capabilities.Capability c)
        enables the dependency flag for the given capability Enabling NOMINAL_ATTRIBUTES also enables BINARY_ATTRIBUTES, UNARY_ATTRIBUTES and EMPTY_NOMINAL_ATTRIBUTES. Enabling BINARY_ATTRIBUTES also enables UNARY_ATTRIBUTES and EMPTY_NOMINAL_ATTRIBUTES. Enabling UNARY_ATTRIBUTES also enables EMPTY_NOMINAL_ATTRIBUTES. But NOMINAL_CLASS only enables BINARY_CLASS, since normal schemes in Weka don't work with datasets that have only 1 class label (or none).
        Parameters:
        c - the capability to enable the dependency flag for

      • enableAll

        public void enableAll()
        enables all attribute and class types (including dependencies)

      • disable

        public void disable​(Capabilities.Capability c)
        disables the given capability Disabling NOMINAL_ATTRIBUTES also disables BINARY_ATTRIBUTES, UNARY_ATTRIBUTES and EMPTY_NOMINAL_ATTRIBUTES. Disabling BINARY_ATTRIBUTES also disables UNARY_ATTRIBUTES and EMPTY_NOMINAL_ATTRIBUTES. Disabling UNARY_ATTRIBUTES also disables EMPTY_NOMINAL_ATTRIBUTES. The same hierarchy applies to the class capabilities.
        Parameters:
        c - the capability to disable

      • disableDependency

        public void disableDependency​(Capabilities.Capability c)
        disables the dependency of the given capability Disabling NOMINAL_ATTRIBUTES also disables BINARY_ATTRIBUTES, UNARY_ATTRIBUTES and EMPTY_NOMINAL_ATTRIBUTES. Disabling BINARY_ATTRIBUTES also disables UNARY_ATTRIBUTES and EMPTY_NOMINAL_ATTRIBUTES. Disabling UNARY_ATTRIBUTES also disables EMPTY_NOMINAL_ATTRIBUTES. The same hierarchy applies to the class capabilities.
        Parameters:
        c - the capability to disable the dependency flag for

      • disableAll

        public void disableAll()
        disables all attribute and class types (including dependencies)

      • getOtherCapabilities

        public Capabilities getOtherCapabilities()
        returns all other capabilities, besides class and attribute related ones
        Returns:
        all other capabilities, besides class and attribute related ones

      • handles

        public boolean handles​(Capabilities.Capability c)
        returns true if the classifier handler has the specified capability
        Parameters:
        c - the capability to test
        Returns:
        true if the classifier handler has the capability

      • hasDependency

        public boolean hasDependency​(Capabilities.Capability c)
        returns true if the classifier handler has a dependency for the specified capability
        Parameters:
        c - the capability to test
        Returns:
        true if the classifier handler has a dependency for the capability

      • hasDependencies

        public boolean hasDependencies()
        Checks whether there are any dependencies at all
        Returns:
        true if there is at least one dependency for a capability

      • getFailReason

        public Exception getFailReason()
        returns the reason why the tests failed, is null if tests succeeded
        Returns:
        the reason why the tests failed

      • test

        public boolean test​(Attribute att)
        Test the given attribute, whether it can be processed by the handler, given its capabilities. The method assumes that the specified attribute is not the class attribute.
        Parameters:
        att - the attribute to test
        Returns:
        true if all the tests succeeded
        See Also:
        test(Attribute, boolean)

      • test

        public boolean test​(Attribute att,
                    boolean isClass)
        Test the given attribute, whether it can be processed by the handler, given its capabilities.
        Parameters:
        att - the attribute to test
        isClass - whether this attribute is the class attribute
        Returns:
        true if all the tests succeeded
        See Also:
        m_AttributeTest

      • test

        public boolean test​(Instances data)
        Tests the given data, whether it can be processed by the handler, given its capabilities. Classifiers implementing the MultiInstanceCapabilitiesHandler interface are checked automatically for their multi-instance Capabilities (if no bags, then only the bag-structure, otherwise only the first bag).
        Parameters:
        data - the data to test
        Returns:
        true if all the tests succeeded
        See Also:
        test(Instances, int, int)

      • test

        public boolean test​(Instances data,
                    int fromIndex,
                    int toIndex)
        Tests a certain range of attributes of the given data, whether it can be processed by the handler, given its capabilities. Classifiers implementing the MultiInstanceCapabilitiesHandler interface are checked automatically for their multi-instance Capabilities (if no bags, then only the bag-structure, otherwise only the first bag).
        Parameters:
        data - the data to test
        fromIndex - the range of attributes - start (incl.)
        toIndex - the range of attributes - end (incl.)
        Returns:
        true if all the tests succeeded
        See Also:
        MultiInstanceCapabilitiesHandler, m_InstancesTest, m_MissingValuesTest, m_MissingClassValuesTest, m_MinimumNumberInstancesTest

      • testWithFail

        public void testWithFail​(Attribute att)
                  throws Exception
        tests the given attribute by calling the test(Attribute,boolean) method and throws an exception if the test fails. The method assumes that the specified attribute is not the class attribute.
        Parameters:
        att - the attribute to test
        Throws:
        Exception - in case the attribute doesn't pass the tests
        See Also:
        test(Attribute,boolean)

      • testWithFail

        public void testWithFail​(Attribute att,
                         boolean isClass)
                  throws Exception
        tests the given attribute by calling the test(Attribute,boolean) method and throws an exception if the test fails.
        Parameters:
        att - the attribute to test
        isClass - whether this attribute is the class attribute
        Throws:
        Exception - in case the attribute doesn't pass the tests
        See Also:
        test(Attribute,boolean)

      • testWithFail

        public void testWithFail​(Instances data,
                         int fromIndex,
                         int toIndex)
                  throws Exception
        tests the given data by calling the test(Instances,int,int) method and throws an exception if the test fails.
        Parameters:
        data - the data to test
        fromIndex - the range of attributes - start (incl.)
        toIndex - the range of attributes - end (incl.)
        Throws:
        Exception - in case the data doesn't pass the tests
        See Also:
        test(Instances,int,int)

      • testWithFail

        public void testWithFail​(Instances data)
                  throws Exception
        tests the given data by calling the test(Instances) method and throws an exception if the test fails.
        Parameters:
        data - the data to test
        Throws:
        Exception - in case the data doesn't pass the tests
        See Also:
        test(Instances)

      • listCapabilities

        public String listCapabilities()
        returns a comma-separated list of all the capabilities, excluding interface-based ones.
        Returns:
        the string describing the capabilities

      • addCapabilities

        public String addCapabilities​(String title)
        generates a string from the capabilities, suitable to add to the help text.
        Parameters:
        title - the title for the capabilities
        Returns:
        a string describing the capabilities

      • toString

        public String toString()
        returns a string representation of the capabilities
        Overrides:
        toString in class Object
        Returns:
        a string representation of this object

      • toSource

        public String toSource​(String objectname)
        turns the capabilities object into source code. The returned source code is a block that creates a Capabilities object named 'objectname' and enables all the capabilities of this Capabilities object.
        Parameters:
        objectname - the name of the Capabilities object being instantiated
        Returns:
        the generated source code

      • toSource

        public String toSource​(String objectname,
                       int indent)
        turns the capabilities object into source code. The returned source code is a block that creates a Capabilities object named 'objectname' and enables all the capabilities of this Capabilities object.
        Parameters:
        objectname - the name of the Capabilities object being instantiated
        indent - the number of blanks to indent
        Returns:
        the generated source code

      • forInstances

        public static Capabilities forInstances​(Instances data)
                                 throws Exception
        returns a Capabilities object specific for this data. The multi-instance capability is not checked as well as the minimum number of instances is not set.
        Parameters:
        data - the data to base the capabilities on
        Returns:
        a data-specific capabilities object
        Throws:
        Exception - in case an error occurrs, e.g., an unknown attribute type

      • forInstances

        public static Capabilities forInstances​(Instances data,
                                        boolean multi)
                                 throws Exception
        returns a Capabilities object specific for this data. The minimum number of instances is not set, the check for multi-instance data is optional.
        Parameters:
        data - the data to base the capabilities on
        multi - if true then the structure is checked, too
        Returns:
        a data-specific capabilities object
        Throws:
        Exception - in case an error occurrs, e.g., an unknown attribute type

      • main

        public static void main​(String[] args)
                 throws Exception
        loads the given dataset and prints the Capabilities necessary to process it.

        Valid parameters:

        -file filename
        the file to load -c index the explicit index of the class attribute (default: none)

        Parameters:
        args - the commandline arguments
        Throws:
        Exception - if something goes wrong