...
Run Format

Source file src/index/suffixarray/suffixarray_test.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
  
  import (
  	"bytes"
  	"math/rand"
  	"regexp"
  	"sort"
  	"strings"
  	"testing"
  )
  
  type testCase struct {
  	name     string   // name of test case
  	source   string   // source to index
  	patterns []string // patterns to lookup
  }
  
  var testCases = []testCase{
  	{
  		"empty string",
  		"",
  		[]string{
  			"",
  			"foo",
  			"(foo)",
  			".*",
  			"a*",
  		},
  	},
  
  	{
  		"all a's",
  		"aaaaaaaaaa", // 10 a's
  		[]string{
  			"",
  			"a",
  			"aa",
  			"aaa",
  			"aaaa",
  			"aaaaa",
  			"aaaaaa",
  			"aaaaaaa",
  			"aaaaaaaa",
  			"aaaaaaaaa",
  			"aaaaaaaaaa",
  			"aaaaaaaaaaa", // 11 a's
  			".",
  			".*",
  			"a+",
  			"aa+",
  			"aaaa[b]?",
  			"aaa*",
  		},
  	},
  
  	{
  		"abc",
  		"abc",
  		[]string{
  			"a",
  			"b",
  			"c",
  			"ab",
  			"bc",
  			"abc",
  			"a.c",
  			"a(b|c)",
  			"abc?",
  		},
  	},
  
  	{
  		"barbara*3",
  		"barbarabarbarabarbara",
  		[]string{
  			"a",
  			"bar",
  			"rab",
  			"arab",
  			"barbar",
  			"bara?bar",
  		},
  	},
  
  	{
  		"typing drill",
  		"Now is the time for all good men to come to the aid of their country.",
  		[]string{
  			"Now",
  			"the time",
  			"to come the aid",
  			"is the time for all good men to come to the aid of their",
  			"to (come|the)?",
  		},
  	},
  
  	{
  		"godoc simulation",
  		"package main\n\nimport(\n    \"rand\"\n    ",
  		[]string{},
  	},
  }
  
  // find all occurrences of s in source; report at most n occurrences
  func find(src, s string, n int) []int {
  	var res []int
  	if s != "" && n != 0 {
  		// find at most n occurrences of s in src
  		for i := -1; n < 0 || len(res) < n; {
  			j := strings.Index(src[i+1:], s)
  			if j < 0 {
  				break
  			}
  			i += j + 1
  			res = append(res, i)
  		}
  	}
  	return res
  }
  
  func testLookup(t *testing.T, tc *testCase, x *Index, s string, n int) {
  	res := x.Lookup([]byte(s), n)
  	exp := find(tc.source, s, n)
  
  	// check that the lengths match
  	if len(res) != len(exp) {
  		t.Errorf("test %q, lookup %q (n = %d): expected %d results; got %d", tc.name, s, n, len(exp), len(res))
  	}
  
  	// if n >= 0 the number of results is limited --- unless n >= all results,
  	// we may obtain different positions from the Index and from find (because
  	// Index may not find the results in the same order as find) => in general
  	// we cannot simply check that the res and exp lists are equal
  
  	// check that each result is in fact a correct match and there are no duplicates
  	sort.Ints(res)
  	for i, r := range res {
  		if r < 0 || len(tc.source) <= r {
  			t.Errorf("test %q, lookup %q, result %d (n = %d): index %d out of range [0, %d[", tc.name, s, i, n, r, len(tc.source))
  		} else if !strings.HasPrefix(tc.source[r:], s) {
  			t.Errorf("test %q, lookup %q, result %d (n = %d): index %d not a match", tc.name, s, i, n, r)
  		}
  		if i > 0 && res[i-1] == r {
  			t.Errorf("test %q, lookup %q, result %d (n = %d): found duplicate index %d", tc.name, s, i, n, r)
  		}
  	}
  
  	if n < 0 {
  		// all results computed - sorted res and exp must be equal
  		for i, r := range res {
  			e := exp[i]
  			if r != e {
  				t.Errorf("test %q, lookup %q, result %d: expected index %d; got %d", tc.name, s, i, e, r)
  			}
  		}
  	}
  }
  
  func testFindAllIndex(t *testing.T, tc *testCase, x *Index, rx *regexp.Regexp, n int) {
  	res := x.FindAllIndex(rx, n)
  	exp := rx.FindAllStringIndex(tc.source, n)
  
  	// check that the lengths match
  	if len(res) != len(exp) {
  		t.Errorf("test %q, FindAllIndex %q (n = %d): expected %d results; got %d", tc.name, rx, n, len(exp), len(res))
  	}
  
  	// if n >= 0 the number of results is limited --- unless n >= all results,
  	// we may obtain different positions from the Index and from regexp (because
  	// Index may not find the results in the same order as regexp) => in general
  	// we cannot simply check that the res and exp lists are equal
  
  	// check that each result is in fact a correct match and the result is sorted
  	for i, r := range res {
  		if r[0] < 0 || r[0] > r[1] || len(tc.source) < r[1] {
  			t.Errorf("test %q, FindAllIndex %q, result %d (n == %d): illegal match [%d, %d]", tc.name, rx, i, n, r[0], r[1])
  		} else if !rx.MatchString(tc.source[r[0]:r[1]]) {
  			t.Errorf("test %q, FindAllIndex %q, result %d (n = %d): [%d, %d] not a match", tc.name, rx, i, n, r[0], r[1])
  		}
  	}
  
  	if n < 0 {
  		// all results computed - sorted res and exp must be equal
  		for i, r := range res {
  			e := exp[i]
  			if r[0] != e[0] || r[1] != e[1] {
  				t.Errorf("test %q, FindAllIndex %q, result %d: expected match [%d, %d]; got [%d, %d]",
  					tc.name, rx, i, e[0], e[1], r[0], r[1])
  			}
  		}
  	}
  }
  
  func testLookups(t *testing.T, tc *testCase, x *Index, n int) {
  	for _, pat := range tc.patterns {
  		testLookup(t, tc, x, pat, n)
  		if rx, err := regexp.Compile(pat); err == nil {
  			testFindAllIndex(t, tc, x, rx, n)
  		}
  	}
  }
  
  // index is used to hide the sort.Interface
  type index Index
  
  func (x *index) Len() int           { return len(x.sa) }
  func (x *index) Less(i, j int) bool { return bytes.Compare(x.at(i), x.at(j)) < 0 }
  func (x *index) Swap(i, j int)      { x.sa[i], x.sa[j] = x.sa[j], x.sa[i] }
  func (a *index) at(i int) []byte    { return a.data[a.sa[i]:] }
  
  func testConstruction(t *testing.T, tc *testCase, x *Index) {
  	if !sort.IsSorted((*index)(x)) {
  		t.Errorf("failed testConstruction %s", tc.name)
  	}
  }
  
  func equal(x, y *Index) bool {
  	if !bytes.Equal(x.data, y.data) {
  		return false
  	}
  	for i, j := range x.sa {
  		if j != y.sa[i] {
  			return false
  		}
  	}
  	return true
  }
  
  // returns the serialized index size
  func testSaveRestore(t *testing.T, tc *testCase, x *Index) int {
  	var buf bytes.Buffer
  	if err := x.Write(&buf); err != nil {
  		t.Errorf("failed writing index %s (%s)", tc.name, err)
  	}
  	size := buf.Len()
  	var y Index
  	if err := y.Read(&buf); err != nil {
  		t.Errorf("failed reading index %s (%s)", tc.name, err)
  	}
  	if !equal(x, &y) {
  		t.Errorf("restored index doesn't match saved index %s", tc.name)
  	}
  	return size
  }
  
  func TestIndex(t *testing.T) {
  	for _, tc := range testCases {
  		x := New([]byte(tc.source))
  		testConstruction(t, &tc, x)
  		testSaveRestore(t, &tc, x)
  		testLookups(t, &tc, x, 0)
  		testLookups(t, &tc, x, 1)
  		testLookups(t, &tc, x, 10)
  		testLookups(t, &tc, x, 2e9)
  		testLookups(t, &tc, x, -1)
  	}
  }
  
  // Of all possible inputs, the random bytes have the least amount of substring
  // repetition, and the repeated bytes have the most. For most algorithms,
  // the running time of every input will be between these two.
  func benchmarkNew(b *testing.B, random bool) {
  	b.StopTimer()
  	data := make([]byte, 1e6)
  	if random {
  		for i := range data {
  			data[i] = byte(rand.Intn(256))
  		}
  	}
  	b.StartTimer()
  	for i := 0; i < b.N; i++ {
  		New(data)
  	}
  }
  
  func BenchmarkNewIndexRandom(b *testing.B) {
  	benchmarkNew(b, true)
  }
  func BenchmarkNewIndexRepeat(b *testing.B) {
  	benchmarkNew(b, false)
  }
  
  func BenchmarkSaveRestore(b *testing.B) {
  	b.StopTimer()
  	r := rand.New(rand.NewSource(0x5a77a1)) // guarantee always same sequence
  	data := make([]byte, 1<<20)             // 1MB of data to index
  	for i := range data {
  		data[i] = byte(r.Intn(256))
  	}
  	x := New(data)
  	size := testSaveRestore(nil, nil, x)       // verify correctness
  	buf := bytes.NewBuffer(make([]byte, size)) // avoid growing
  	b.SetBytes(int64(size))
  	b.StartTimer()
  	for i := 0; i < b.N; i++ {
  		x.Write(buf)
  		var y Index
  		y.Read(buf)
  	}
  }
  

View as plain text