...
Run Format

Source file src/index/suffixarray/suffixarray.go

Documentation: index/suffixarray

  // Copyright 2010 The Go Authors. All rights reserved.
  // Use of this source code is governed by a BSD-style
  // license that can be found in the LICENSE file.
  
  // Package suffixarray implements substring search in logarithmic time using
  // an in-memory suffix array.
  //
  // Example use:
  //
  //	// create index for some data
  //	index := suffixarray.New(data)
  //
  //	// lookup byte slice s
  //	offsets1 := index.Lookup(s, -1) // the list of all indices where s occurs in data
  //	offsets2 := index.Lookup(s, 3)  // the list of at most 3 indices where s occurs in data
  //
  package suffixarray
  
  import (
  	"bytes"
  	"encoding/binary"
  	"io"
  	"regexp"
  	"sort"
  )
  
  // Index implements a suffix array for fast substring search.
  type Index struct {
  	data []byte
  	sa   []int // suffix array for data; len(sa) == len(data)
  }
  
  // New creates a new Index for data.
  // Index creation time is O(N*log(N)) for N = len(data).
  func New(data []byte) *Index {
  	return &Index{data, qsufsort(data)}
  }
  
  // writeInt writes an int x to w using buf to buffer the write.
  func writeInt(w io.Writer, buf []byte, x int) error {
  	binary.PutVarint(buf, int64(x))
  	_, err := w.Write(buf[0:binary.MaxVarintLen64])
  	return err
  }
  
  // readInt reads an int x from r using buf to buffer the read and returns x.
  func readInt(r io.Reader, buf []byte) (int, error) {
  	_, err := io.ReadFull(r, buf[0:binary.MaxVarintLen64]) // ok to continue with error
  	x, _ := binary.Varint(buf)
  	return int(x), err
  }
  
  // writeSlice writes data[:n] to w and returns n.
  // It uses buf to buffer the write.
  func writeSlice(w io.Writer, buf []byte, data []int) (n int, err error) {
  	// encode as many elements as fit into buf
  	p := binary.MaxVarintLen64
  	for ; n < len(data) && p+binary.MaxVarintLen64 <= len(buf); n++ {
  		p += binary.PutUvarint(buf[p:], uint64(data[n]))
  	}
  
  	// update buffer size
  	binary.PutVarint(buf, int64(p))
  
  	// write buffer
  	_, err = w.Write(buf[0:p])
  	return
  }
  
  // readSlice reads data[:n] from r and returns n.
  // It uses buf to buffer the read.
  func readSlice(r io.Reader, buf []byte, data []int) (n int, err error) {
  	// read buffer size
  	var size int
  	size, err = readInt(r, buf)
  	if err != nil {
  		return
  	}
  
  	// read buffer w/o the size
  	if _, err = io.ReadFull(r, buf[binary.MaxVarintLen64:size]); err != nil {
  		return
  	}
  
  	// decode as many elements as present in buf
  	for p := binary.MaxVarintLen64; p < size; n++ {
  		x, w := binary.Uvarint(buf[p:])
  		data[n] = int(x)
  		p += w
  	}
  
  	return
  }
  
  const bufSize = 16 << 10 // reasonable for BenchmarkSaveRestore
  
  // Read reads the index from r into x; x must not be nil.
  func (x *Index) Read(r io.Reader) error {
  	// buffer for all reads
  	buf := make([]byte, bufSize)
  
  	// read length
  	n, err := readInt(r, buf)
  	if err != nil {
  		return err
  	}
  
  	// allocate space
  	if 2*n < cap(x.data) || cap(x.data) < n {
  		// new data is significantly smaller or larger then
  		// existing buffers - allocate new ones
  		x.data = make([]byte, n)
  		x.sa = make([]int, n)
  	} else {
  		// re-use existing buffers
  		x.data = x.data[0:n]
  		x.sa = x.sa[0:n]
  	}
  
  	// read data
  	if _, err := io.ReadFull(r, x.data); err != nil {
  		return err
  	}
  
  	// read index
  	for sa := x.sa; len(sa) > 0; {
  		n, err := readSlice(r, buf, sa)
  		if err != nil {
  			return err
  		}
  		sa = sa[n:]
  	}
  	return nil
  }
  
  // Write writes the index x to w.
  func (x *Index) Write(w io.Writer) error {
  	// buffer for all writes
  	buf := make([]byte, bufSize)
  
  	// write length
  	if err := writeInt(w, buf, len(x.data)); err != nil {
  		return err
  	}
  
  	// write data
  	if _, err := w.Write(x.data); err != nil {
  		return err
  	}
  
  	// write index
  	for sa := x.sa; len(sa) > 0; {
  		n, err := writeSlice(w, buf, sa)
  		if err != nil {
  			return err
  		}
  		sa = sa[n:]
  	}
  	return nil
  }
  
  // Bytes returns the data over which the index was created.
  // It must not be modified.
  //
  func (x *Index) Bytes() []byte {
  	return x.data
  }
  
  func (x *Index) at(i int) []byte {
  	return x.data[x.sa[i]:]
  }
  
  // lookupAll returns a slice into the matching region of the index.
  // The runtime is O(log(N)*len(s)).
  func (x *Index) lookupAll(s []byte) []int {
  	// find matching suffix index range [i:j]
  	// find the first index where s would be the prefix
  	i := sort.Search(len(x.sa), func(i int) bool { return bytes.Compare(x.at(i), s) >= 0 })
  	// starting at i, find the first index at which s is not a prefix
  	j := i + sort.Search(len(x.sa)-i, func(j int) bool { return !bytes.HasPrefix(x.at(j+i), s) })
  	return x.sa[i:j]
  }
  
  // Lookup returns an unsorted list of at most n indices where the byte string s
  // occurs in the indexed data. If n < 0, all occurrences are returned.
  // The result is nil if s is empty, s is not found, or n == 0.
  // Lookup time is O(log(N)*len(s) + len(result)) where N is the
  // size of the indexed data.
  //
  func (x *Index) Lookup(s []byte, n int) (result []int) {
  	if len(s) > 0 && n != 0 {
  		matches := x.lookupAll(s)
  		if n < 0 || len(matches) < n {
  			n = len(matches)
  		}
  		// 0 <= n <= len(matches)
  		if n > 0 {
  			result = make([]int, n)
  			copy(result, matches)
  		}
  	}
  	return
  }
  
  // FindAllIndex returns a sorted list of non-overlapping matches of the
  // regular expression r, where a match is a pair of indices specifying
  // the matched slice of x.Bytes(). If n < 0, all matches are returned
  // in successive order. Otherwise, at most n matches are returned and
  // they may not be successive. The result is nil if there are no matches,
  // or if n == 0.
  //
  func (x *Index) FindAllIndex(r *regexp.Regexp, n int) (result [][]int) {
  	// a non-empty literal prefix is used to determine possible
  	// match start indices with Lookup
  	prefix, complete := r.LiteralPrefix()
  	lit := []byte(prefix)
  
  	// worst-case scenario: no literal prefix
  	if prefix == "" {
  		return r.FindAllIndex(x.data, n)
  	}
  
  	// if regexp is a literal just use Lookup and convert its
  	// result into match pairs
  	if complete {
  		// Lookup returns indices that may belong to overlapping matches.
  		// After eliminating them, we may end up with fewer than n matches.
  		// If we don't have enough at the end, redo the search with an
  		// increased value n1, but only if Lookup returned all the requested
  		// indices in the first place (if it returned fewer than that then
  		// there cannot be more).
  		for n1 := n; ; n1 += 2 * (n - len(result)) /* overflow ok */ {
  			indices := x.Lookup(lit, n1)
  			if len(indices) == 0 {
  				return
  			}
  			sort.Ints(indices)
  			pairs := make([]int, 2*len(indices))
  			result = make([][]int, len(indices))
  			count := 0
  			prev := 0
  			for _, i := range indices {
  				if count == n {
  					break
  				}
  				// ignore indices leading to overlapping matches
  				if prev <= i {
  					j := 2 * count
  					pairs[j+0] = i
  					pairs[j+1] = i + len(lit)
  					result[count] = pairs[j : j+2]
  					count++
  					prev = i + len(lit)
  				}
  			}
  			result = result[0:count]
  			if len(result) >= n || len(indices) != n1 {
  				// found all matches or there's no chance to find more
  				// (n and n1 can be negative)
  				break
  			}
  		}
  		if len(result) == 0 {
  			result = nil
  		}
  		return
  	}
  
  	// regexp has a non-empty literal prefix; Lookup(lit) computes
  	// the indices of possible complete matches; use these as starting
  	// points for anchored searches
  	// (regexp "^" matches beginning of input, not beginning of line)
  	r = regexp.MustCompile("^" + r.String()) // compiles because r compiled
  
  	// same comment about Lookup applies here as in the loop above
  	for n1 := n; ; n1 += 2 * (n - len(result)) /* overflow ok */ {
  		indices := x.Lookup(lit, n1)
  		if len(indices) == 0 {
  			return
  		}
  		sort.Ints(indices)
  		result = result[0:0]
  		prev := 0
  		for _, i := range indices {
  			if len(result) == n {
  				break
  			}
  			m := r.FindIndex(x.data[i:]) // anchored search - will not run off
  			// ignore indices leading to overlapping matches
  			if m != nil && prev <= i {
  				m[0] = i // correct m
  				m[1] += i
  				result = append(result, m)
  				prev = m[1]
  			}
  		}
  		if len(result) >= n || len(indices) != n1 {
  			// found all matches or there's no chance to find more
  			// (n and n1 can be negative)
  			break
  		}
  	}
  	if len(result) == 0 {
  		result = nil
  	}
  	return
  }
  

View as plain text