From 86d9beb26e38f40b0b1314aa7dd36772142c9178 Mon Sep 17 00:00:00 2001
From: Ulderico Cirello <ulderico.cirello@gmail.com>
Date: Wed, 17 Jan 2024 11:39:07 -0800
Subject: [PATCH] Extract byteseeker into its own package

---
 filter_test.go                    | 56 +++----------------------------
 internal/byteseeker/byteseeker.go | 51 ++++++++++++++++++++++++++++
 utils.go                          | 12 ++++---
 3 files changed, 62 insertions(+), 57 deletions(-)
 create mode 100644 internal/byteseeker/byteseeker.go

diff --git a/filter_test.go b/filter_test.go
index 222f62b..688c109 100644
--- a/filter_test.go
+++ b/filter_test.go
@@ -2,59 +2,11 @@ package splitblockbloom
 
 import (
 	"bytes"
-	"errors"
 	"fmt"
-	"io"
 	"testing"
-)
-
-// ByteSliceReadSeeker implements the io.ReadSeeker interface for a byte slice.
-type ByteSliceReadSeeker struct {
-	slice  []byte
-	offset int64
-}
-
-// NewByteSliceReadSeeker creates a new ByteSliceReadSeeker.
-func NewByteSliceReadSeeker(slice []byte) *ByteSliceReadSeeker {
-	return &ByteSliceReadSeeker{slice: slice, offset: 0}
-}
-
-// Read implements the Read method of the io.ReadSeeker interface.
-func (r *ByteSliceReadSeeker) Read(p []byte) (int, error) {
-	if r.offset >= int64(len(r.slice)) {
-		return 0, io.EOF // end of slice
-	}
-
-	n := copy(p, r.slice[r.offset:])
-	r.offset += int64(n)
-	return n, nil
-}
 
-// Seek implements the Seek method of the io.ReadSeeker interface.
-func (r *ByteSliceReadSeeker) Seek(offset int64, whence int) (int64, error) {
-	var newOffset int64
-	switch whence {
-	case io.SeekStart:
-		newOffset = offset
-	case io.SeekCurrent:
-		newOffset = r.offset + offset
-	case io.SeekEnd:
-		newOffset = int64(len(r.slice)) + offset
-	default:
-		return 0, errors.New("invalid whence")
-	}
-
-	if newOffset < 0 {
-		return 0, errors.New("negative position")
-	}
-
-	if newOffset > int64(len(r.slice)) {
-		return 0, errors.New("position out of bounds")
-	}
-
-	r.offset = newOffset
-	return newOffset, nil
-}
+	"github.com/axiomhq/splitblockbloom/internal/byteseeker"
+)
 
 func TestFilter(t *testing.T) {
 	bb := NewFilter(1e6, 0.004)
@@ -90,7 +42,7 @@ func TestFilter(t *testing.T) {
 	t.Log("wrote:", n, "bytes", "len:", len(b))
 
 	for i := 0; i < 1e6; i++ {
-		ok, err := ContainsFromStream(NewByteSliceReadSeeker(b), len(bb), []byte(fmt.Sprintf("val%d", i)))
+		ok, err := ContainsFromStream(&byteseeker.Buffer{B: b}, len(bb), []byte(fmt.Sprintf("val%d", i)))
 		if err != nil {
 			t.Fatal(err)
 		}
@@ -101,7 +53,7 @@ func TestFilter(t *testing.T) {
 
 	errs = 0
 	for i := int(1e6); i < 10e6; i++ {
-		ok, err := ContainsFromStream(NewByteSliceReadSeeker(b), len(bb), []byte(fmt.Sprintf("val%d", i)))
+		ok, err := ContainsFromStream(&byteseeker.Buffer{B: b}, len(bb), []byte(fmt.Sprintf("val%d", i)))
 		if err != nil {
 			t.Fatal(err)
 		}
diff --git a/internal/byteseeker/byteseeker.go b/internal/byteseeker/byteseeker.go
new file mode 100644
index 0000000..a90f775
--- /dev/null
+++ b/internal/byteseeker/byteseeker.go
@@ -0,0 +1,51 @@
+package byteseeker
+
+import (
+	"errors"
+	"io"
+)
+
+var _ io.ReadSeeker = (*Buffer)(nil)
+
+// Buffer implements the io.ReadSeeker interface for a byte slice.
+type Buffer struct {
+	B      []byte
+	offset int64
+}
+
+// Read implements the Read method of the io.ReadSeeker interface.
+func (r *Buffer) Read(p []byte) (int, error) {
+	if r.offset >= int64(len(r.B)) {
+		return 0, io.EOF // end of slice
+	}
+
+	n := copy(p, r.B[r.offset:])
+	r.offset += int64(n)
+	return n, nil
+}
+
+// Seek implements the Seek method of the io.ReadSeeker interface.
+func (r *Buffer) Seek(offset int64, whence int) (int64, error) {
+	var newOffset int64
+	switch whence {
+	case io.SeekStart:
+		newOffset = offset
+	case io.SeekCurrent:
+		newOffset = r.offset + offset
+	case io.SeekEnd:
+		newOffset = int64(len(r.B)) + offset
+	default:
+		return 0, errors.New("invalid whence")
+	}
+
+	if newOffset < 0 {
+		return 0, errors.New("negative position")
+	}
+
+	if newOffset > int64(len(r.B)) {
+		return 0, errors.New("position out of bounds")
+	}
+
+	r.offset = newOffset
+	return newOffset, nil
+}
diff --git a/utils.go b/utils.go
index 3ae766c..e9df834 100644
--- a/utils.go
+++ b/utils.go
@@ -16,9 +16,11 @@ const (
 	blockSeed        = 0x5c6bfb31
 )
 
-func hash(val []byte, seed uint64) uint64 { return fnv1a.AddBytes64(seed, val) }
+func hash(val []byte, seed uint64) uint64 {
+	return fnv1a.AddBytes64(seed, val)
+}
 
-func calcFpp(ndv, bytes float64) float64 {
+func calcFalsePositiveRatio(ndv, bytes float64) float64 {
 	if ndv == 0 || bytes <= 0 || ndv/(bytes*8) > 3 {
 		return 1.0
 	}
@@ -41,9 +43,9 @@ func calcFpp(ndv, bytes float64) float64 {
 	return math.Min(result, 1.0)
 }
 
-func blockBytesNeeded(ndv, fpp float64) uint64 {
+func blockBytesNeeded(ndv, desiredFalsePositiveRatio float64) uint64 {
 	result := 1.0
-	for calcFpp(ndv, result) > fpp {
+	for calcFalsePositiveRatio(ndv, result) > desiredFalsePositiveRatio {
 		result *= 2
 	}
 	if result <= blockSizeInBytes {
@@ -53,7 +55,7 @@ func blockBytesNeeded(ndv, fpp float64) uint64 {
 	lo, hi := 0.0, result
 	for lo < hi-1 {
 		mid := lo + (hi-lo)/2
-		if calcFpp(ndv, mid) < fpp {
+		if calcFalsePositiveRatio(ndv, mid) < desiredFalsePositiveRatio {
 			hi = mid
 		} else {
 			lo = mid