Document apoc.convert.toTree

[jexp]

Especially the config options

• lower case relationships
• exclusion / inclusion filters for properties by label/type

perhaps also how to convert individual nodes/rels into paths using the paths functions

and if/how virtual entities work with this procedure

[fabio-s-franco]

The documenation for apoc.toTree is completely absent in conversion functions of the documentation. I am trying to figure out how to include property of relationships in the result, but the documentation is not there. So I don't even know if it's possible.

[bburns]

FWIW here's the current code -

neo4j-apoc-procedures/core/src/main/java/apoc/convert/Json.java

Lines 73 to 114 in 16581f3

	@Procedure("apoc.convert.toTree")
	@Description("apoc.convert.toTree([paths],[lowerCaseRels=true], [config]) creates a stream of nested documents representing the at least one root of these paths")
	// todo optinally provide root node
	public Stream<MapResult> toTree(@Name("paths") List<Path> paths, @Name(value = "lowerCaseRels",defaultValue = "true") boolean lowerCaseRels, @Name(value = "config", defaultValue = "{}") Map<String, Object> config) {
	if (paths.isEmpty()) return Stream.of(new MapResult(Collections.emptyMap()));
	ConvertConfig conf = new ConvertConfig(config);
	Map<String, List<String>> nodes = conf.getNodes();
	Map<String, List<String>> rels = conf.getRels();
	Map<Long, Map<String, Object>> maps = new HashMap<>(paths.size() * 100);
	for (Path path : paths) {
	Iterator<Entity> it = path.iterator();
	while (it.hasNext()) {
	Node n = (Node) it.next();
	Map<String, Object> nMap = maps.computeIfAbsent(n.getId(), (id) -> toMap(n, nodes));
	if (it.hasNext()) {
	Relationship r = (Relationship) it.next();
	Node m = r.getOtherNode(n);
	String typeName = lowerCaseRels ? r.getType().name().toLowerCase() : r.getType().name();
	// todo take direction into account and create collection into outgoing direction ??
	// parent-[:HAS_CHILD]->(child) vs. (parent)<-[:PARENT_OF]-(child)
	if (!nMap.containsKey(typeName)) nMap.put(typeName, new ArrayList<>(16));
	List<Map<String, Object>> list = (List) nMap.get(typeName);
	Optional<Map<String, Object>> optMap = list.stream()
	.filter(elem -> elem.get("_id").equals(m.getId()))
	.findFirst();
	if (!optMap.isPresent()) {
	Map<String, Object> mMap = toMap(m, nodes);
	mMap = addRelProperties(mMap, typeName, r, rels);
	maps.put(m.getId(), mMap);
	list.add(maps.get(m.getId()));
	}
	}
	}
	}
	return paths.stream()
	.map(Path::startNode)
	.distinct()
	.map(n -> maps.remove(n.getId()))
	.map(m -> m == null ? Collections.<String,Object>emptyMap() : m)
	.map(MapResult::new);
	}

and tests - https://github.com/neo4j-contrib/neo4j-apoc-procedures/blob/16581f3253f82e75506428dfd2a498346c6bfb1d/core/src/test/java/apoc/convert/ConvertJsonTest.java

The tests include some examples of including/excluding properties - eg

CALL apoc.convert.toTree(ps, true, {nodes: {Category: ['name']}, rels: {subcategory:['-id']}}) yield value

[jexp]

@conker84 @mneedham can one of you add docs with these examples?