// A naive suffix tree implementation sclass SuffixTree { Node root; S fullText; int nodeCount; Comparator childComparator = (a, b) -> a.firstCharOrMinus1()-b.firstCharOrMinus1(); new DummyNode dummyNode; sinterface IFirstChar { int firstCharOrMinus1(); } sclass DummyNode implements IFirstChar { int firstChar; public int firstCharOrMinus1() { ret firstChar; } } sclass Node implements IFirstChar { Substring text; CompactTreeSet children; // actually, IntBuffer positions; *() {} *(Substring *text) {} void addPosition(int i) { if (positions == null) positions = new IntBuffer; positions.add(i); } void addChild(SuffixTree tree, Node n) { if (children == null) children = makeEmptyChildrenSet(tree); children.add(n); } CompactTreeSet makeEmptyChildrenSet(SuffixTree tree) { ret new CompactTreeSet<>(tree.childComparator); } public int firstCharOrMinus1() { ret charToIntOrMinus1(text); } Node getChild(SuffixTree tree, int c) { if (children == null) null; tree.dummyNode.firstChar = c; ret (Node) children.find(tree.dummyNode); } } *() {} *(S *fullText) { root = new Node(Substring(fullText, 0, 0)); ++nodeCount; for i over fullText: { addSuffix(Substring(fullText, i), i); if (((i+1) % 1000000) == 0) print((i+1) + " suffixes added (" + nNodes(nodeCount) + ")"); } } void addSuffix(Substring s, int position) { Node node = root; while (!empty(s)) { int _n = lCommonPrefix_CharSequence(node.text, s); s = s.substring(_n); if (_n >= l(node.text)) { // node text exhausted if (empty(s)) { // pattern also exhausted - done node.addPosition(position); ret; } else { Node n = node.getChild(SuffixTree.this, charToIntOrMinus1(s)); if (n == null) { n = new Node(s); ++nodeCount; n.addPosition(position); node.addChild(SuffixTree.this, n); ret; } else node = n; } } else { // node text not exhausted // split node. first, move all the node's vitals to a new node nOld Node nOld = new Node(node.text.substring(_n)); ++nodeCount; nOld.positions = node.positions; node.positions = null; nOld.children = node.children; node.children = null; node.text = node.text.substring(0, _n); node.addChild(SuffixTree.this, nOld); // now add a new node Node nNew = new Node(s); ++nodeCount; nNew.addPosition(position); node.addChild(SuffixTree.this, nNew); ret; } } } public L indicesOf(S pattern) { new IntBuffer out; getIndicesOf(out, root, pattern, 0); ret out.asVirtualList(); } void getIndicesOf(IntBuffer out, Node node, S pattern, int iPattern) { int lPattern = l(pattern); while true { int n = lCommonPrefix_CharSequence(node.text, Substring(pattern, iPattern)); iPattern += n; if (iPattern >= lPattern) break; // pattern exhausted - done if (n < l(node.text)) ret; // mismatch, exit Node child = node.getChild(SuffixTree.this, charAtAsIntOrMinus1(pattern, iPattern)); if (child != null) continue with node = child; ret; } collectPositions(out, node); } L getPositions(Node node) { new IntBuffer out; collectPositions(out, node); ret out.asVirtualList(); } void collectPositions(IntBuffer out, Node node) { if (node == null) ret; out.addAll(node.positions); fOr (IFirstChar n : node.children) collectPositions(out, (Node) n); } void printMe() { printNode("", "", root); } void printNode(S indent, S pre, Node node) { print(indent + pre + quote(shorten(20, node.text)) + (empty(node.positions) ? "" : " " + node.positions)); fOr (IFirstChar _n : node.children) { Node n = cast _n; printNode(indent + " ", "[" + (empty(n.text) ? "end" : quote(n.text.charAt(0))) + "] ", n); } } ItIt allNodes() { new L> stack; stack.add(iteratorLL(root)); ret iteratorFromFunction_if0(() -> { while (nempty(stack)) { if (!last(stack).hasNext()) popLast(stack); else { Node n = last(stack).next(); stack.add((Iterator) iterator(n.children)); ret n; } } null; }); } }