Question: Per-word "Identity" Value at Serialization/Deserialization

[dgerding]

Hi,
Thanks for the great writeup and highly functional DAWG.

I wonder if there is a way to infer (?) a unique numeric identity at serialization. For example, is there some value to grab (a numeric representation of path to node as id sequence?) that I could treat as a unique numeric identifier for that string? My goal is to know some unique value that persists both to and from serialization so that I can map a word value to a related data item.

When I ran the File_TapTopTapsTops test, which specifies four entries for the dawg:

var dawg = Dawg.CreateBuilder(new[] {"tap", "top", "taps", "tops"}) .Build();

I expected I would hit the debug breakpoint I put on the modified version of WriteLists in DawgBuilder 4 times, but only hit it twice. I expect that's because I have the wrong mental model of the encoding.

But what I was hoping for was to find some form of "per word numeric identity" in the graph, lurking at the moment of serialization.

` void WriteLists(Queue pending, LongestListsPerHash longest, IDawgStateWriter writer)
{
while (pending. Count > 0)
{
var current = pending. Dequeue();

            writer.MoveToNode(current.NodeIndex);
            writer.IsLastSibling = current.IsLastSibling;
            writer.SymbolId = current.SymbolId;
            writer.IsEndOfWord = current.IsTerminal;

            if (current.IsTerminal)
            {
                var test = pending.ToString();
            }`

Any suggestions or help appreciated. I'm going back to reread. Thanks again.! :)