not_streaming

    is: ro, isa: Bool, default: false

not_streaming determines the behavior for the opening tag parsed. If what is being parsed is not a stream, the document will be parsed in full then placed into the finished_nodes attribute. Otherwise, the opening tag will be placed immediately into the finished_nodes bucket.

current_node

    is: rw, isa: POE::Filter::XML::Node

current_node holds the node being immediately parsed.

finished_nodes

    is: ro, isa: ArrayRef, traits: Array

finished_nodes holds the nodes that have been completely parsed. Access to this attribute is provided through the following methods:

    handles =>
    {
        all_finished_nodes => 'elements',
        has_finished_nodes => 'count',
        add_finished_node => 'push',
        get_finished_node => 'shift',
    }

depth_stack

    is: ro, isa: ArrayRef, traits: Array

depth_stack holds the operating stack for the parsed nodes. As nodes are processed, ancendants of the current node are stored in the stack. When done they are popped off the stack. Access to this attribute is provided through the following methods:

    handles =>
    {
        push_depth_stack => 'push',
        pop_depth_stack => 'pop',
        depth => 'count',
    }

reset

override start_element

    (HashRef $data)

start_element is overriden from the XML::SAX::Base class to provide our custom behavior for dealing with streaming vs. non-streaming data. It builds Nodes then attaches them to either the root node (non-streaming) or as stand-alone top level fragments (streaming) sets them to the current node. Children nodes are appended to their parents before getting set as the current node. Then the base class method is called via super()

override end_element

    (HashRef $data)

end_element is overriden from the XML::SAX::Base class to provide our custom behavior for dealing with streaming vs. non-streaming data. Mostly this method is in charge of stack management when the depth of the stack reaches certain points. In streaming documents, this means that top level fragments (not root) are popped off the stack and added to the finished_nodes collection. Otherwise a Node is created with stream_end set and added to the finished nodes.

Then the base class method is called via super()

override characters

    (HashRef $data)

characters merely applies the character data as text to the current node being processed. It then calls the base class method via super().