Can this be made to work with more complex graph?

Nobody has claimed this yet.

I have family tree data in a levelgraph structured like:
```
{subject: INDI, predicate: 'child_of', object: FAMI}
{subject: FAMI, predicate: 'has_child', object: INDI}

{subject: INDI, predicate: 'marriage', object: FAMI}
{subject: FAMI, predicate: 'parent', object: INDI}
```
where INDI is a unique id for an individual and FAMI is a unique id for a family group the consists of 0..2 'parent' and 0..n 'has_child'.

So for example:
db.get({subject: 'I0169', predicate: 'child_of'}) returns the family group where 'I0169' is a child, say 'F0917'
db.get({subject: 'F0917', predicate: 'has_child'}) returns all the children of 'F0917'
db.get({subject: 'F0917', predicate: 'parent'}) returns the parents of this family group
db.get({subject: 'I0169', predicate: 'marriage'}) returns the family group id for each marriage of 'I0169'

I'm looking to get 3 things from the graph:

1. list of descendants of an INDI
2. list of ancestors of an INDI
3. the shortest path relationship from INDI_A to INDI_B

for ancestor and descendants list if would be nice to also get the level that each result is at so level is incremented/decremented as you move up/down through the tree.

Start by reviewing the repository's existing breadth-first and depth-first traversal entry points and how they query LevelGraph. Define how family and individual nodes should be traversed, then specify results for descendants, ancestors with levels, and shortest paths before implementing and testing the behavior.

Written by the indexing model from the issue text.