kyle-wip

container/strset/strset.go

@@ -0,0 +1,119 @@
+package strmap
+
+import (
+	"fmt"
+	"sort"
+	"strings"
+)
+
+// StrSet represents GC friendly string set implementation.
+// The pseudo code:
+//
+//	off := 0
+//	m := make(map[string]int)
+//	for _, s := range ss {
+//			m[s] = off
+//			off += len(s)
+//	}
+//
+// it only supports Get method after StrSet is created.
+// it's useful for scenario like `map[string]YourStruct` where YourStruct contains no pointer
+// so you can change it to `map[int]YourStruct` and use StrSet for keys for better performance.
+// you can also it for replacing `map[string]struct{}{}`
+type StrSet struct {
+	data  []byte
+	items []strItem
+}
+
+type strItem struct {
+	off  int
+	sz   int
+	hash uint32
+}
+
+// New creates a string set from []string
+func New(ss []string) *StrSet {
+	sz := 0
+	for _, s := range ss {
+		sz += len(s)
+	}
+	b := make([]byte, 0, sz)
+	items := make([]strItem, len(ss))
+	dup := make(map[string]struct{}, len(ss))
+	for i, s := range ss {
+		_, ok := dup[s]
+		if ok {
+			continue
+		}
+		dup[s] = struct{}{}
+		e := &items[i]
+		e.off = len(b)
+		e.sz = len(s)
+		e.hash = fnvhashstr(s)
+		b = append(b, s...)
+	}
+	ret := &StrSet{data: b, items: items}
+	ret.sort()
+	return ret
+}
+
+func (m *StrSet) sort() {
+	sort.Slice(m.items, func(i, j int) bool {
+		return m.items[i].hash < m.items[j].hash
+	})
+}
+
+// Get returns an ID for s, -1 if not found
+func (m *StrSet) Get(s string) int {
+	h := fnvhashstr(s)
+	// binary search
+	i, j := 0, len(m.items)
+	for i < j {
+		p := int(uint(i+j) >> 1) // (i+j)/2
+		// i ≤ p < j
+		if h > m.items[p].hash {
+			i = p + 1
+		} else {
+			j = p
+		}
+	}
+	//  i == j, check if found
+	for i < len(m.items) {
+		e := &m.items[i]
+		if e.hash != h {
+			return -1
+		}
+		if string(m.data[e.off:e.off+e.sz]) == s { // double check
+			return e.off
+		}
+		i++ // hash conflict
+	}
+	return -1
+}
+
+func (m *StrSet) String() string {
+	b := strings.Builder{}
+	b.WriteByte('[')
+	for i, e := range m.items {
+		if i != 0 {
+			b.WriteString(", ")
+		}
+		fmt.Fprintf(&b, "{off:%d, hash:%x, str:%q}", e.off, e.hash, string(m.data[e.off:e.off+e.sz]))
+	}
+	b.WriteByte(']')
+	return b.String()
+}
+
+const (
+	fnvHashOffset32 = uint32(2166136261)
+	fnvHashPrime32  = uint32(16777619)
+)
+
+func fnvhashstr(s string) uint32 {
+	h := fnvHashOffset32
+	for i := 0; i < len(s); i++ {
+		h *= fnvHashPrime32
+		h ^= uint32(s[i])
+	}
+	return h
+}

container/strset/strset_test.go

@@ -0,0 +1,50 @@
+package strmap
+
+import (
+	"math/rand"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/require"
+)
+
+var letters = []rune("abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ")
+
+func randString(n int) string {
+	b := make([]rune, n)
+	for i := range b {
+		b[i] = letters[rand.Intn(len(letters))]
+	}
+	return string(b)
+}
+
+func randStrings(m, n int) []string {
+	seed := time.Now().UnixNano()
+	seed = 1720779864101503000
+	println(seed)
+	r := rand.New(rand.NewSource(seed))
+	b := make([]byte, m)
+	ret := make([]string, 0, n)
+	for i := 0; i < n; i++ {
+		r.Read(b)
+		ret = append(ret, string(b[:1+r.Intn(m)]))
+	}
+	return ret
+}
+
+func TestStrMap(t *testing.T) {
+	ss := randStrings(5, 100)
+	off := 0
+	m := make(map[string]int)
+	for _, s := range ss {
+		_, ok := m[s]
+		if !ok {
+			m[s] = off
+			off += len(s)
+		}
+	}
+	strset := New(ss)
+	for i, s := range ss {
+		require.Equal(t, m[s], strset.Get(s), i)
+	}
+}