// A naive suffix tree implementation sclass SuffixTree { Node root; S fullText; int nodeCount; Comparator childComparator = (a, b) -> a.firstCharOrMinus1(this)-b.firstCharOrMinus1(this); new DummyNode dummyNode; sinterface IFirstChar { int firstCharOrMinus1(SuffixTree tree); } sclass DummyNode implements IFirstChar { int firstChar; public int firstCharOrMinus1(SuffixTree tree) { ret firstChar; } } sclass Node implements IFirstChar { int from, to; // our range in fullText O children; // null, a Node or Node[] sorted by first character int position = -1; *() {} *(int *from, int *to) {} *(Substring s) { from = s.startIndex(); to = s.endIndex(); } Substring text(SuffixTree tree) { ret Substring(tree.fullText, from, to); } int lText() { ret to-from; } void addPosition(int i) { if (position >= 0) fail("Already have position"); position = i; } void addChild(SuffixTree tree, Node n) { if (children == null) children = n; else if (children cast Node) children = sortArrayInPlace(new Node[] {children, n}, tree.childComparator); else { Node[] c = cast children; Node[] x = new[l(c)+1]; arraycopy(c, 0, x, 0, l(c)); x[l(x)-1] = n; children = sortArrayInPlace(x, tree.childComparator); } } Cl children() { if (children == null) ret emptyList(); if (children cast Node) ret ll(children); ret wrapArrayAsList((Node[]) children); } CompactTreeSet makeEmptyChildrenSet(SuffixTree tree) { ret new CompactTreeSet<>(tree.childComparator); } public int firstCharOrMinus1(SuffixTree tree) { ret charToIntOrMinus1(text(tree)); } Node getChild(SuffixTree tree, int c) { if (children == null) null; if (children cast Node) ret children.firstCharOrMinus1(tree) == c ? children : null; tree.dummyNode.firstChar = c; Node[] x = cast children; int i = Arrays.binarySearch(x, tree.dummyNode, tree.childComparator); ret i >= 0 ? x[i] : null; } } *() {} *(S *fullText) { root = new Node(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(SuffixTree.this), s); s = s.substring(_n); if (_n >= node.lText()) { // 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.from+_n, node.to); ++nodeCount; nOld.position = node.position; node.position = -1; nOld.children = node.children; node.children = null; node.to = node.from+_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) { ret asList(indicesOf_iterator(pattern)); } public ItIt indicesOf_iterator(S pattern) { ret mapI_notNull(allNodesUnder(scanDown(root, pattern)), n -> n.position >= 0 ? n.position : null); } Node scanDown(Node node, S pattern) { int lPattern = l(pattern), iPattern = 0; while true { int n = lCommonPrefix_CharSequence(node.text(this), Substring(pattern, iPattern)); iPattern += n; if (iPattern >= lPattern) break; // pattern exhausted - done if (n < node.lText()) null; // mismatch, exit Node child = node.getChild(SuffixTree.this, charAtAsIntOrMinus1(pattern, iPattern)); if (child != null) continue with node = child; null; } ret node; } L getPositions(Node node) { new IntBuffer out; collectPositions(out, node); ret out.asVirtualList(); } void collectPositions(IntBuffer out, Node node) { if (node == null) ret; if (node.position >= 0) out.add(node.position); 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(this))) + (node.position < 0 ? "" : " [" + node.position + "]")); fOr (IFirstChar _n : node.children()) { Node n = cast _n; printNode(indent + " ", "[" + (n.lText() == 0 ? "end" : quote(n.text(this).charAt(0))) + "] ", n); } } ItIt allNodes() { ret allNodesUnder(root); } // includes the node itself ItIt allNodesUnder(Node node) { new L> stack; if (node != null) stack.add(iteratorLL(node)); 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; }); } }