API reference - Class EdgeFilter

Notation used in Ruby API documentation

Module: db

Description: A generic edge filter adaptor

Edge filters are an efficient way to filter edge from a Edges collection. To apply a filter, derive your own filter class and pass an instance to the Edges#filter or Edges#filtered method.

Conceptually, these methods take each edge from the collection and present it to the filter's 'selected' method. Based on the result of this evaluation, the edge is kept or discarded.

The magic happens when deep mode edge collections are involved. In that case, the filter will use as few calls as possible and exploit the hierarchical compression if possible. It needs to know however, how the filter behaves. You need to configure the filter by calling is_isotropic, is_scale_invariant or is_isotropic_and_scale_invariant before using the filter.

You can skip this step, but the filter algorithm will assume the worst case then. This usually leads to cell variant formation which is not always desired and blows up the hierarchy.

Here is some example that filters edges parallel to a given one:

class ParallelFilter < RBA::EdgeFilter

  # Constructor
  def initialize(ref_edge)
    self.is_scale_invariant   # orientation matters, but scale does not
    @ref_edge = ref_edge
  end
  
  # Select only parallel ones
  def selected(edge)
    return edge.is_parallel?(@ref_edge)
  end

end

edges = ... # some Edges object
ref_edge = ... # some Edge
parallel_only = edges.filtered(ParallelFilter::new(ref_edge))

This class has been introduced in version 0.29.

Public constructors

new EdgeFilter ptrnewCreates a new object of this class

Public methods

[const]EdgeFilter ptr_const_castReturns a non-const reference to self.
void_createEnsures the C++ object is created
void_destroyExplicitly destroys the object
[const]bool_destroyed?Returns a value indicating whether the object was already destroyed
[const]bool_is_const_object?Returns a value indicating whether the reference is a const reference
void_manageMarks the object as managed by the script side.
void_unmanageMarks the object as no longer owned by the script side.
voidis_isotropicIndicates that the filter has isotropic properties
voidis_isotropic_and_scale_invariantIndicates that the filter is isotropic and scale invariant
voidis_scale_invariantIndicates that the filter is scale invariant
voidrequires_raw_input=(bool flag)Sets a value indicating whether the filter needs raw (unmerged) input
[const]boolrequires_raw_input?Gets a value indicating whether the filter needs raw (unmerged) input
[virtual,const]boolselected(const Edge edge)Selects an edge
voidwants_variants=(bool flag)Sets a value indicating whether the filter prefers cell variants
[const]boolwants_variants?Gets a value indicating whether the filter prefers cell variants

Deprecated methods (protected, public, static, non-static and constructors)

voidcreateUse of this method is deprecated. Use _create instead
voiddestroyUse of this method is deprecated. Use _destroy instead
[const]booldestroyed?Use of this method is deprecated. Use _destroyed? instead
[const]boolis_const_object?Use of this method is deprecated. Use _is_const_object? instead

Detailed description

_const_cast

Signature: [const] EdgeFilter ptr _const_cast

Description: Returns a non-const reference to self.

Basically, this method allows turning a const object reference to a non-const one. This method is provided as last resort to remove the constness from an object. Usually there is a good reason for a const object reference, so using this method may have undesired side effects.

This method has been introduced in version 0.29.6.

_create

Signature: void _create

Description: Ensures the C++ object is created

Use this method to ensure the C++ object is created, for example to ensure that resources are allocated. Usually C++ objects are created on demand and not necessarily when the script object is created.

_destroy

Signature: void _destroy

Description: Explicitly destroys the object

Explicitly destroys the object on C++ side if it was owned by the script interpreter. Subsequent access to this object will throw an exception. If the object is not owned by the script, this method will do nothing.

_destroyed?

Signature: [const] bool _destroyed?

Description: Returns a value indicating whether the object was already destroyed

This method returns true, if the object was destroyed, either explicitly or by the C++ side. The latter may happen, if the object is owned by a C++ object which got destroyed itself.

_is_const_object?

Signature: [const] bool _is_const_object?

Description: Returns a value indicating whether the reference is a const reference

This method returns true, if self is a const reference. In that case, only const methods may be called on self.

_manage

Signature: void _manage

Description: Marks the object as managed by the script side.

After calling this method on an object, the script side will be responsible for the management of the object. This method may be called if an object is returned from a C++ function and the object is known not to be owned by any C++ instance. If necessary, the script side may delete the object if the script's reference is no longer required.

Usually it's not required to call this method. It has been introduced in version 0.24.

_unmanage

Signature: void _unmanage

Description: Marks the object as no longer owned by the script side.

Calling this method will make this object no longer owned by the script's memory management. Instead, the object must be managed in some other way. Usually this method may be called if it is known that some C++ object holds and manages this object. Technically speaking, this method will turn the script's reference into a weak reference. After the script engine decides to delete the reference, the object itself will still exist. If the object is not managed otherwise, memory leaks will occur.

Usually it's not required to call this method. It has been introduced in version 0.24.

create

Signature: void create

Description: Ensures the C++ object is created

Use of this method is deprecated. Use _create instead

Use this method to ensure the C++ object is created, for example to ensure that resources are allocated. Usually C++ objects are created on demand and not necessarily when the script object is created.

destroy

Signature: void destroy

Description: Explicitly destroys the object

Use of this method is deprecated. Use _destroy instead

Explicitly destroys the object on C++ side if it was owned by the script interpreter. Subsequent access to this object will throw an exception. If the object is not owned by the script, this method will do nothing.

destroyed?

Signature: [const] bool destroyed?

Description: Returns a value indicating whether the object was already destroyed

Use of this method is deprecated. Use _destroyed? instead

This method returns true, if the object was destroyed, either explicitly or by the C++ side. The latter may happen, if the object is owned by a C++ object which got destroyed itself.

is_const_object?

Signature: [const] bool is_const_object?

Description: Returns a value indicating whether the reference is a const reference

Use of this method is deprecated. Use _is_const_object? instead

This method returns true, if self is a const reference. In that case, only const methods may be called on self.

is_isotropic

Signature: void is_isotropic

Description: Indicates that the filter has isotropic properties

Call this method before using the filter to indicate that the selection is independent of the orientation of the shape. This helps the filter algorithm optimizing the filter run, specifically in hierarchical mode.

Examples for isotropic (polygon) filters are area or perimeter filters. The area or perimeter of a polygon depends on the scale, but not on the orientation of the polygon.

is_isotropic_and_scale_invariant

Signature: void is_isotropic_and_scale_invariant

Description: Indicates that the filter is isotropic and scale invariant

Call this method before using the filter to indicate that the selection is independent of the scale and orientation of the shape. This helps the filter algorithm optimizing the filter run, specifically in hierarchical mode.

An example for such a (polygon) filter is the square selector. Whether a polygon is a square or not does not depend on the polygon's orientation nor scale.

is_scale_invariant

Signature: void is_scale_invariant

Description: Indicates that the filter is scale invariant

Call this method before using the filter to indicate that the selection is independent of the scale of the shape. This helps the filter algorithm optimizing the filter run, specifically in hierarchical mode.

An example for a scale invariant (polygon) filter is the bounding box aspect ratio (height/width) filter. The definition of heigh and width depends on the orientation, but the ratio is independent on scale.

new

Signature: [static] new EdgeFilter ptr new

Description: Creates a new object of this class

Python specific notes:
This method is the default initializer of the object.

requires_raw_input=

Signature: void requires_raw_input= (bool flag)

Description: Sets a value indicating whether the filter needs raw (unmerged) input

This flag must be set before using this filter. It tells the filter implementation whether the filter wants to have raw input (unmerged). The default value is 'false', meaning that the filter will receive merged polygons ('merged semantics').

Setting this value to false potentially saves some CPU time needed for merging the polygons. Also, raw input means that strange shapes such as dot-like edges, self-overlapping polygons, empty or degenerated polygons are preserved.

Python specific notes:
The object exposes a writable attribute 'requires_raw_input'. This is the setter.

requires_raw_input?

Signature: [const] bool requires_raw_input?

Description: Gets a value indicating whether the filter needs raw (unmerged) input

See requires_raw_input= for details.

Python specific notes:
The object exposes a readable attribute 'requires_raw_input'. This is the getter.

selected

Signature: [virtual,const] bool selected (const Edge edge)

Description: Selects an edge

This method is the actual payload. It needs to be reimplemented in a derived class. It needs to analyze the edge and return 'true' if it should be kept and 'false' if it should be discarded.

wants_variants=

Signature: void wants_variants= (bool flag)

Description: Sets a value indicating whether the filter prefers cell variants

This flag must be set before using this filter for hierarchical applications (deep mode). It tells the filter implementation whether cell variants should be created (true, the default) or shape propagation will be applied (false).

This decision needs to be made, if the filter indicates that it will deliver different results for scaled or rotated versions of the shape (see is_isotropic and the other hints). If a cell is present with different qualities - as seen from the top cell - the respective instances need to be differentiated. Cell variant formation is one way, shape propagation the other way. Typically, cell variant formation is less expensive, but the hierarchy will be modified.

Python specific notes:
The object exposes a writable attribute 'wants_variants'. This is the setter.

wants_variants?

Signature: [const] bool wants_variants?

Description: Gets a value indicating whether the filter prefers cell variants

See wants_variants= for details.

Python specific notes:
The object exposes a readable attribute 'wants_variants'. This is the getter.