Update benchmarks and add WithCopyStrings option (#36)

Fixes #35

Author: klauspost

README.md

@@ -29,20 +29,35 @@ func benchmarkFromFile(b *testing.B, filename string) {
 	}
 	msg := loadCompressed(b, filename)
 
-	b.SetBytes(int64(len(msg)))
-	b.ReportAllocs()
-	b.ResetTimer()
-
-	pj := &ParsedJson{}
-
-	for i := 0; i < b.N; i++ {
-		// Reset tape
-		var err error
-		pj, err = Parse(msg, pj)
-		if err != nil {
-			b.Fatal(err)
+	b.Run("copy", func(b *testing.B) {
+		pj := &ParsedJson{}
+		b.SetBytes(int64(len(msg)))
+		b.ReportAllocs()
+		b.ResetTimer()
+		for i := 0; i < b.N; i++ {
+			// Reset tape
+			var err error
+			pj, err = Parse(msg, pj, WithCopyStrings(true))
+			if err != nil {
+				b.Fatal(err)
+			}
 		}
-	}
+	})
+	b.Run("nocopy", func(b *testing.B) {
+		pj := &ParsedJson{}
+		b.SetBytes(int64(len(msg)))
+		b.ReportAllocs()
+		b.ResetTimer()
+		for i := 0; i < b.N; i++ {
+			// Reset tape
+			var err error
+			pj, err = Parse(msg, pj, WithCopyStrings(false))
+			if err != nil {
+				b.Fatal(err)
+			}
+		}
+	})
+
 }
 
 func BenchmarkApache_builds(b *testing.B)  { benchmarkFromFile(b, "apache_builds") }

benchmarks_test.go

@@ -241,7 +241,7 @@ func verifyDemoNdjson(pj internalParsedJson, t *testing.T, object int) {
 		//}
 		//fmt.Printf("{%s, 0x%x},\n", c, tp&0xffffffffffffff)
 		expected := tc.expected[ii].val | (uint64(tc.expected[ii].c) << 56)
-		if !alwaysCopyStrings && tp != expected {
+		if !pj.copyStrings && tp != expected {
 			t.Errorf("verifyDemoNdjson(%d): got: %016x want: %016x", ii, tp, expected)
 		}
 	}

ndjson_test.go

@@ -0,0 +1,18 @@
+package simdjson
+
+// ParserOption is a parser option.
+type ParserOption func(pj *internalParsedJson) error
+
+// WithCopyStrings will copy strings so they no longer reference the input.
+// For enhanced performance, simdjson-go can point back into the original JSON buffer for strings,
+// however this can lead to issues in streaming use cases scenarios, or scenarios in which
+// the underlying JSON buffer is reused. So the default behaviour is to create copies of all
+// strings (not just those transformed anyway for unicode escape characters) into the separate
+// Strings buffer (at the expense of using more memory and less performance).
+// Default: true - strings are copied.
+func WithCopyStrings(b bool) ParserOption {
+	return func(pj *internalParsedJson) error {
+		pj.copyStrings = b
+		return nil
+	}
+}

options.go

@@ -33,14 +33,14 @@ func _parse_string(src, dst, pcurrent_string_buf_loc unsafe.Pointer) (res uint64
 
 // Disable new -d=checkptr behaviour for Go 1.14
 //go:nocheckptr
-func parseStringSimdValidateOnly(buf []byte, maxStringSize, dst_length *uint64, need_copy *bool) bool {
+func parseStringSimdValidateOnly(buf []byte, maxStringSize, dstLength *uint64, needCopy *bool) bool {
 
 	src := uintptr(unsafe.Pointer(&buf[1])) // Use buf[1] in order to skip opening quote
 	src_length := uint64(0)
 
-	success := _parse_string_validate_only(unsafe.Pointer(src), unsafe.Pointer(&maxStringSize), unsafe.Pointer(&src_length), unsafe.Pointer(dst_length))
+	success := _parse_string_validate_only(unsafe.Pointer(src), unsafe.Pointer(&maxStringSize), unsafe.Pointer(&src_length), unsafe.Pointer(dstLength))
 
-	*need_copy = alwaysCopyStrings || src_length != *dst_length
+	*needCopy = *needCopy || src_length != *dstLength
 	return success != 0
 }
 

parse_string_amd64.go

@@ -26,15 +26,6 @@ import (
 	"strconv"
 )
 
-//
-// For enhanced performance, simdjson-go can point back into the original JSON buffer for strings,
-// however this can lead to issues in streaming use cases scenarios, or scenarios in which
-// the underlying JSON buffer is reused. So the default behaviour is to create copies of all
-// strings (not just those transformed anyway for unicode escape characters) into the separate
-// Strings buffer (at the expense of using more memory and less performance).
-//
-const alwaysCopyStrings = true
-
 const JSONVALUEMASK = 0xffffffffffffff
 const JSONTAGMASK = 0xff << 56
 const STRINGBUFBIT = 0x80000000000000
@@ -96,6 +87,7 @@ type internalParsedJson struct {
 	buffers               [indexSlots][indexSize]uint32
 	buffersOffset         uint64
 	ndjson                uint64
+	copyStrings           bool
 }
 
 // Iter returns a new Iter.

parsed_json.go

@@ -37,9 +37,7 @@ func SupportedCPU() bool {
 	return cpuid.CPU.Supports(cpuid.AVX2, cpuid.CLMUL)
 }
 
-// Parse a block of data and return the parsed JSON.
-// An optional block of previously parsed json can be supplied to reduce allocations.
-func Parse(b []byte, reuse *ParsedJson) (*ParsedJson, error) {
+func newInternalParsedJson(reuse *ParsedJson, opts []ParserOption) (*internalParsedJson, error) {
 	if !SupportedCPU() {
 		return nil, errors.New("Host CPU does not meet target specs")
 	}
@@ -53,7 +51,23 @@ func Parse(b []byte, reuse *ParsedJson) (*ParsedJson, error) {
 	if pj == nil {
 		pj = &internalParsedJson{}
 	}
-	err := pj.parseMessage(b)
+	pj.copyStrings = true
+	for _, opt := range opts {
+		if err := opt(pj); err != nil {
+			return nil, err
+		}
+	}
+	return pj, nil
+}
+
+// Parse a block of data and return the parsed JSON.
+// An optional block of previously parsed json can be supplied to reduce allocations.
+func Parse(b []byte, reuse *ParsedJson, opts ...ParserOption) (*ParsedJson, error) {
+	pj, err := newInternalParsedJson(reuse, opts)
+	if err != nil {
+		return nil, err
+	}
+	err = pj.parseMessage(b)
 	if err != nil {
 		return nil, err
 	}
@@ -64,17 +78,12 @@ func Parse(b []byte, reuse *ParsedJson) (*ParsedJson, error) {
 
 // ParseND will parse newline delimited JSON.
 // An optional block of previously parsed json can be supplied to reduce allocations.
-func ParseND(b []byte, reuse *ParsedJson) (*ParsedJson, error) {
-	if !SupportedCPU() {
-		return nil, errors.New("Host CPU does not meet target specs")
-	}
-	var pj internalParsedJson
-	if reuse != nil {
-		pj.ParsedJson = *reuse
+func ParseND(b []byte, reuse *ParsedJson, opts ...ParserOption) (*ParsedJson, error) {
+	pj, err := newInternalParsedJson(reuse, opts)
+	if err != nil {
+		return nil, err
 	}
-	b = bytes.TrimSpace(b)
-
-	err := pj.parseMessageNdjson(b)
+	err = pj.parseMessageNdjson(bytes.TrimSpace(b))
 	if err != nil {
 		return nil, err
 	}
@@ -166,6 +175,7 @@ func ParseNDStream(r io.Reader, res chan<- Stream, reuse <-chan *ParsedJson) {
 				queue <- result
 				go func() {
 					var pj internalParsedJson
+					pj.copyStrings = true
 					select {
 					case v := <-reuse:
 						if cap(v.Message) >= tmpSize+1024 {

simdjson_amd64.go

@@ -71,9 +71,8 @@ func peekSize(pj *internalParsedJson) uint64 {
 	return uint64(pj.indexesChan.indexes[pj.indexesChan.index])
 }
 
-func parseString(pj *ParsedJson, idx uint64, maxStringSize uint64) bool {
+func parseString(pj *ParsedJson, idx uint64, maxStringSize uint64, needCopy bool) bool {
 	size := uint64(0)
-	need_copy := false
 	buf := pj.Message[idx:]
 	// Make sure that we have at least one full YMM word available after maxStringSize into the buffer
 	if len(buf)-int(maxStringSize) < 64 {
@@ -87,10 +86,10 @@ func parseString(pj *ParsedJson, idx uint64, maxStringSize uint64) bool {
 			buf = paddedBuf[:]
 		}
 	}
-	if !parseStringSimdValidateOnly(buf, &maxStringSize, &size, &need_copy) {
+	if !parseStringSimdValidateOnly(buf, &maxStringSize, &size, &needCopy) {
 		return false
 	}
-	if !need_copy {
+	if !needCopy {
 		pj.write_tape(idx+1, '"')
 	} else {
 		// Make sure we account for at least 32 bytes additional space due to
@@ -237,7 +236,7 @@ object_begin:
 	}
 	switch buf[idx] {
 	case '"':
-		if !parseString(&pj.ParsedJson, idx, peekSize(pj)) {
+		if !parseString(&pj.ParsedJson, idx, peekSize(pj), pj.copyStrings) {
 			goto fail
 		}
 		goto object_key_state
@@ -259,7 +258,7 @@ object_key_state:
 	}
 	switch buf[idx] {
 	case '"':
-		if !parseString(&pj.ParsedJson, idx, peekSize(pj)) {
+		if !parseString(&pj.ParsedJson, idx, peekSize(pj), pj.copyStrings) {
 			goto fail
 		}
 
@@ -319,7 +318,7 @@ objectContinue:
 		if buf[idx] != '"' {
 			goto fail
 		}
-		if !parseString(&pj.ParsedJson, idx, peekSize(pj)) {
+		if !parseString(&pj.ParsedJson, idx, peekSize(pj), pj.copyStrings) {
 			goto fail
 		}
 		goto object_key_state
@@ -365,7 +364,7 @@ mainArraySwitch:
 	// on paths that can accept a close square brace (post-, and at start)
 	switch buf[idx] {
 	case '"':
-		if !parseString(&pj.ParsedJson, idx, peekSize(pj)) {
+		if !parseString(&pj.ParsedJson, idx, peekSize(pj), pj.copyStrings) {
 			goto fail
 		}
 	case 't':

stage2_build_tape_amd64.go

@@ -184,7 +184,7 @@ func TestStage2BuildTape(t *testing.T) {
 			//}
 			//fmt.Printf("{%s, 0x%x},\n", c, tp&0xffffffffffffff)
 			expected := tc.expected[ii].val | (uint64(tc.expected[ii].c) << 56)
-			if !alwaysCopyStrings && tp != expected {
+			if !pj.copyStrings && tp != expected {
 				t.Errorf("TestStage2BuildTape(%d): got: %d want: %d", ii, tp, expected)
 			}
 		}