//===- llvm/ADT/SuffixTree.h - Tree for substrings --------------*- C++ -*-===//
 
//
 
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
 
// See https://llvm.org/LICENSE.txt for license information.
 
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 
//
 
//===----------------------------------------------------------------------===//
 
//
 
// This file defines the Suffix Tree class and Suffix Tree Node struct.
 
//
 
//===----------------------------------------------------------------------===//
 
#ifndef LLVM_SUPPORT_SUFFIXTREE_H
 
#define LLVM_SUPPORT_SUFFIXTREE_H
 
 
 
#include "llvm/ADT/ArrayRef.h"
 
#include "llvm/ADT/DenseMap.h"
 
#include "llvm/Support/Allocator.h"
 
#include <vector>
 
 
 
namespace llvm {
 
 
 
/// Represents an undefined index in the suffix tree.
 
const unsigned EmptyIdx = -1;
 
 
 
/// A node in a suffix tree which represents a substring or suffix.
 
///
 
/// Each node has either no children or at least two children, with the root
 
/// being a exception in the empty tree.
 
///
 
/// Children are represented as a map between unsigned integers and nodes. If
 
/// a node N has a child M on unsigned integer k, then the mapping represented
 
/// by N is a proper prefix of the mapping represented by M. Note that this,
 
/// although similar to a trie is somewhat different: each node stores a full
 
/// substring of the full mapping rather than a single character state.
 
///
 
/// Each internal node contains a pointer to the internal node representing
 
/// the same string, but with the first character chopped off. This is stored
 
/// in \p Link. Each leaf node stores the start index of its respective
 
/// suffix in \p SuffixIdx.
 
struct SuffixTreeNode {
 
 
 
  /// The children of this node.
 
  ///
 
  /// A child existing on an unsigned integer implies that from the mapping
 
  /// represented by the current node, there is a way to reach another
 
  /// mapping by tacking that character on the end of the current string.
 
  llvm::DenseMap<unsigned, SuffixTreeNode *> Children;
 
 
 
  /// The start index of this node's substring in the main string.
 
  unsigned StartIdx = EmptyIdx;
 
 
 
  /// The end index of this node's substring in the main string.
 
  ///
 
  /// Every leaf node must have its \p EndIdx incremented at the end of every
 
  /// step in the construction algorithm. To avoid having to update O(N)
 
  /// nodes individually at the end of every step, the end index is stored
 
  /// as a pointer.
 
  unsigned *EndIdx = nullptr;
 
 
 
  /// For leaves, the start index of the suffix represented by this node.
 
  ///
 
  /// For all other nodes, this is ignored.
 
  unsigned SuffixIdx = EmptyIdx;
 
 
 
  /// For internal nodes, a pointer to the internal node representing
 
  /// the same sequence with the first character chopped off.
 
  ///
 
  /// This acts as a shortcut in Ukkonen's algorithm. One of the things that
 
  /// Ukkonen's algorithm does to achieve linear-time construction is
 
  /// keep track of which node the next insert should be at. This makes each
 
  /// insert O(1), and there are a total of O(N) inserts. The suffix link
 
  /// helps with inserting children of internal nodes.
 
  ///
 
  /// Say we add a child to an internal node with associated mapping S. The
 
  /// next insertion must be at the node representing S - its first character.
 
  /// This is given by the way that we iteratively build the tree in Ukkonen's
 
  /// algorithm. The main idea is to look at the suffixes of each prefix in the
 
  /// string, starting with the longest suffix of the prefix, and ending with
 
  /// the shortest. Therefore, if we keep pointers between such nodes, we can
 
  /// move to the next insertion point in O(1) time. If we don't, then we'd
 
  /// have to query from the root, which takes O(N) time. This would make the
 
  /// construction algorithm O(N^2) rather than O(N).
 
  SuffixTreeNode *Link = nullptr;
 
 
 
  /// The length of the string formed by concatenating the edge labels from the
 
  /// root to this node.
 
  unsigned ConcatLen = 0;
 
 
 
  /// Returns true if this node is a leaf.
 
  bool isLeaf() const { return SuffixIdx != EmptyIdx; }
 
 
 
  /// Returns true if this node is the root of its owning \p SuffixTree.
 
  bool isRoot() const { return StartIdx == EmptyIdx; }
 
 
 
  /// Return the number of elements in the substring associated with this node.
 
  size_t size() const {
 
 
 
    // Is it the root? If so, it's the empty string so return 0.
 
    if (isRoot())
 
      return 0;
 
 
 
    assert(*EndIdx != EmptyIdx && "EndIdx is undefined!");
 
 
 
    // Size = the number of elements in the string.
 
    // For example, [0 1 2 3] has length 4, not 3. 3-0 = 3, so we have 3-0+1.
 
    return *EndIdx - StartIdx + 1;
 
  }
 
 
 
  SuffixTreeNode(unsigned StartIdx, unsigned *EndIdx, SuffixTreeNode *Link)
 
      : StartIdx(StartIdx), EndIdx(EndIdx), Link(Link) {}
 
 
 
  SuffixTreeNode() = default;
 
};
 
 
 
/// A data structure for fast substring queries.
 
///
 
/// Suffix trees represent the suffixes of their input strings in their leaves.
 
/// A suffix tree is a type of compressed trie structure where each node
 
/// represents an entire substring rather than a single character. Each leaf
 
/// of the tree is a suffix.
 
///
 
/// A suffix tree can be seen as a type of state machine where each state is a
 
/// substring of the full string. The tree is structured so that, for a string
 
/// of length N, there are exactly N leaves in the tree. This structure allows
 
/// us to quickly find repeated substrings of the input string.
 
///
 
/// In this implementation, a "string" is a vector of unsigned integers.
 
/// These integers may result from hashing some data type. A suffix tree can
 
/// contain 1 or many strings, which can then be queried as one large string.
 
///
 
/// The suffix tree is implemented using Ukkonen's algorithm for linear-time
 
/// suffix tree construction. Ukkonen's algorithm is explained in more detail
 
/// in the paper by Esko Ukkonen "On-line construction of suffix trees. The
 
/// paper is available at
 
///
 
/// https://www.cs.helsinki.fi/u/ukkonen/SuffixT1withFigs.pdf
 
class SuffixTree {
 
public:
 
  /// Each element is an integer representing an instruction in the module.
 
  llvm::ArrayRef<unsigned> Str;
 
 
 
  /// A repeated substring in the tree.
 
  struct RepeatedSubstring {
 
    /// The length of the string.
 
    unsigned Length;
 
 
 
    /// The start indices of each occurrence.
 
    std::vector<unsigned> StartIndices;
 
  };
 
 
 
private:
 
  /// Maintains each node in the tree.
 
  llvm::SpecificBumpPtrAllocator<SuffixTreeNode> NodeAllocator;
 
 
 
  /// The root of the suffix tree.
 
  ///
 
  /// The root represents the empty string. It is maintained by the
 
  /// \p NodeAllocator like every other node in the tree.
 
  SuffixTreeNode *Root = nullptr;
 
 
 
  /// Maintains the end indices of the internal nodes in the tree.
 
  ///
 
  /// Each internal node is guaranteed to never have its end index change
 
  /// during the construction algorithm; however, leaves must be updated at
 
  /// every step. Therefore, we need to store leaf end indices by reference
 
  /// to avoid updating O(N) leaves at every step of construction. Thus,
 
  /// every internal node must be allocated its own end index.
 
  llvm::BumpPtrAllocator InternalEndIdxAllocator;
 
 
 
  /// The end index of each leaf in the tree.
 
  unsigned LeafEndIdx = -1;
 
 
 
  /// Helper struct which keeps track of the next insertion point in
 
  /// Ukkonen's algorithm.
 
  struct ActiveState {
 
    /// The next node to insert at.
 
    SuffixTreeNode *Node = nullptr;
 
 
 
    /// The index of the first character in the substring currently being added.
 
    unsigned Idx = EmptyIdx;
 
 
 
    /// The length of the substring we have to add at the current step.
 
    unsigned Len = 0;
 
  };
 
 
 
  /// The point the next insertion will take place at in the
 
  /// construction algorithm.
 
  ActiveState Active;
 
 
 
  /// Allocate a leaf node and add it to the tree.
 
  ///
 
  /// \param Parent The parent of this node.
 
  /// \param StartIdx The start index of this node's associated string.
 
  /// \param Edge The label on the edge leaving \p Parent to this node.
 
  ///
 
  /// \returns A pointer to the allocated leaf node.
 
  SuffixTreeNode *insertLeaf(SuffixTreeNode &Parent, unsigned StartIdx,
 
                             unsigned Edge);
 
 
 
  /// Allocate an internal node and add it to the tree.
 
  ///
 
  /// \param Parent The parent of this node. Only null when allocating the root.
 
  /// \param StartIdx The start index of this node's associated string.
 
  /// \param EndIdx The end index of this node's associated string.
 
  /// \param Edge The label on the edge leaving \p Parent to this node.
 
  ///
 
  /// \returns A pointer to the allocated internal node.
 
  SuffixTreeNode *insertInternalNode(SuffixTreeNode *Parent, unsigned StartIdx,
 
                                     unsigned EndIdx, unsigned Edge);
 
 
 
  /// Set the suffix indices of the leaves to the start indices of their
 
  /// respective suffixes.
 
  void setSuffixIndices();
 
 
 
  /// Construct the suffix tree for the prefix of the input ending at
 
  /// \p EndIdx.
 
  ///
 
  /// Used to construct the full suffix tree iteratively. At the end of each
 
  /// step, the constructed suffix tree is either a valid suffix tree, or a
 
  /// suffix tree with implicit suffixes. At the end of the final step, the
 
  /// suffix tree is a valid tree.
 
  ///
 
  /// \param EndIdx The end index of the current prefix in the main string.
 
  /// \param SuffixesToAdd The number of suffixes that must be added
 
  /// to complete the suffix tree at the current phase.
 
  ///
 
  /// \returns The number of suffixes that have not been added at the end of
 
  /// this step.
 
  unsigned extend(unsigned EndIdx, unsigned SuffixesToAdd);
 
 
 
public:
 
  /// Construct a suffix tree from a sequence of unsigned integers.
 
  ///
 
  /// \param Str The string to construct the suffix tree for.
 
  SuffixTree(const std::vector<unsigned> &Str);
 
 
 
  /// Iterator for finding all repeated substrings in the suffix tree.
 
  struct RepeatedSubstringIterator {
 
  private:
 
    /// The current node we're visiting.
 
    SuffixTreeNode *N = nullptr;
 
 
 
    /// The repeated substring associated with this node.
 
    RepeatedSubstring RS;
 
 
 
    /// The nodes left to visit.
 
    std::vector<SuffixTreeNode *> ToVisit;
 
 
 
    /// The minimum length of a repeated substring to find.
 
    /// Since we're outlining, we want at least two instructions in the range.
 
    /// FIXME: This may not be true for targets like X86 which support many
 
    /// instruction lengths.
 
    const unsigned MinLength = 2;
 
 
 
    /// Move the iterator to the next repeated substring.
 
    void advance() {
 
      // Clear the current state. If we're at the end of the range, then this
 
      // is the state we want to be in.
 
      RS = RepeatedSubstring();
 
      N = nullptr;
 
 
 
      // Each leaf node represents a repeat of a string.
 
      std::vector<SuffixTreeNode *> LeafChildren;
 
 
 
      // Continue visiting nodes until we find one which repeats more than once.
 
      while (!ToVisit.empty()) {
 
        SuffixTreeNode *Curr = ToVisit.back();
 
        ToVisit.pop_back();
 
        LeafChildren.clear();
 
 
 
        // Keep track of the length of the string associated with the node. If
 
        // it's too short, we'll quit.
 
        unsigned Length = Curr->ConcatLen;
 
 
 
        // Iterate over each child, saving internal nodes for visiting, and
 
        // leaf nodes in LeafChildren. Internal nodes represent individual
 
        // strings, which may repeat.
 
        for (auto &ChildPair : Curr->Children) {
 
          // Save all of this node's children for processing.
 
          if (!ChildPair.second->isLeaf())
 
            ToVisit.push_back(ChildPair.second);
 
 
 
          // It's not an internal node, so it must be a leaf. If we have a
 
          // long enough string, then save the leaf children.
 
          else if (Length >= MinLength)
 
            LeafChildren.push_back(ChildPair.second);
 
        }
 
 
 
        // The root never represents a repeated substring. If we're looking at
 
        // that, then skip it.
 
        if (Curr->isRoot())
 
          continue;
 
 
 
        // Do we have any repeated substrings?
 
        if (LeafChildren.size() >= 2) {
 
          // Yes. Update the state to reflect this, and then bail out.
 
          N = Curr;
 
          RS.Length = Length;
 
          for (SuffixTreeNode *Leaf : LeafChildren)
 
            RS.StartIndices.push_back(Leaf->SuffixIdx);
 
          break;
 
        }
 
      }
 
 
 
      // At this point, either NewRS is an empty RepeatedSubstring, or it was
 
      // set in the above loop. Similarly, N is either nullptr, or the node
 
      // associated with NewRS.
 
    }
 
 
 
  public:
 
    /// Return the current repeated substring.
 
    RepeatedSubstring &operator*() { return RS; }
 
 
 
    RepeatedSubstringIterator &operator++() {
 
      advance();
 
      return *this;
 
    }
 
 
 
    RepeatedSubstringIterator operator++(int I) {
 
      RepeatedSubstringIterator It(*this);
 
      advance();
 
      return It;
 
    }
 
 
 
    bool operator==(const RepeatedSubstringIterator &Other) const {
 
      return N == Other.N;
 
    }
 
    bool operator!=(const RepeatedSubstringIterator &Other) const {
 
      return !(*this == Other);
 
    }
 
 
 
    RepeatedSubstringIterator(SuffixTreeNode *N) : N(N) {
 
      // Do we have a non-null node?
 
      if (N) {
 
        // Yes. At the first step, we need to visit all of N's children.
 
        // Note: This means that we visit N last.
 
        ToVisit.push_back(N);
 
        advance();
 
      }
 
    }
 
  };
 
 
 
  typedef RepeatedSubstringIterator iterator;
 
  iterator begin() { return iterator(Root); }
 
  iterator end() { return iterator(nullptr); }
 
};
 
 
 
} // namespace llvm
 
 
 
#endif // LLVM_SUPPORT_SUFFIXTREE_H