argus-json-notes.txt

Argus renders JSON (from taxomachine and treemachine).

It expects two objects:
    1. a tree
    2. object with a 'domsource' attribute (e.g. {"domsource":"ottol"})

The tree is nested set of node description. Each node has the attributes:
    "children" -> list of child nodes.

    "nleaves" -> 0 for a leaf, # of descendant "leaves" for the node. Note that
        the # of leaves refers to the # of leaves in the tree that has been
        pruned to "maxnodedepth" (see below), not the # of nodes in the source
        graph that are descendants of the node.

    "nTipDescendants" -> 0 for a leaf, # of tips within the node's clade,
        *excluding* the node itself. (In other words, this reflects the full
        source graph, irrespective of "maxonodedepth".)

    "name" -> string with the node's name.

    "nodeid" -> string that is a node ID in the webservice that generated the JSON
        this ID is not guaranteed to be stable (at least, I don't think it is)
        across multiple instances of the service. Perhaps not even stable if the 
        neo4j db is recreated. Argus uses this to enable navigation by attaching a 
        callback that requests more of the tree starting at node that corresponds to 
        this nodeid.

In addition the root is expected to have:
    "maxnodedepth" -> number of node levels that include their children.

When this JSON is retrieved on the client, it is deserialized by a "reviver"
function that converts some elements (those with the 'nodeid' attribute) from
simple objects to instances of the ArgusNode pseudo-class.