Modernise code

.gitattributes

@@ -0,0 +1,2 @@
+# Generated protobuf code: collapse in diffs/reviews and exclude from language stats.
+internal/zetasketch/*.pb.go linguist-generated=true

.github/workflows/test.yml

@@ -8,23 +8,4 @@ on:
       - main
 jobs:
   go:
-    runs-on: ubuntu-latest
-    strategy:
-      matrix:
-        go-version: [1.26.x]
-    steps:
-      - name: Checkout
-        uses: actions/checkout@v4
-      - name: Cache dependencies
-        uses: actions/cache@v4
-        with:
-          path: ~/go/pkg/mod
-          key: ${{ runner.os }}-go-${{ hashFiles('**/go.sum') }}
-          restore-keys: |
-            ${{ runner.os }}-go-
-      - name: Setup Go
-        uses: actions/setup-go@v5
-        with:
-          go-version: ${{ matrix.go-version }}
-      - name: Run tests
-        run: make test
+    uses: bsm/misc/.github/workflows/test-go.yml@main

go.mod

@@ -2,8 +2,4 @@ module github.com/bsm/zetasketch
 
 go 1.26.0
 
-require (
-	github.com/bsm/ginkgo v1.16.5
-	github.com/bsm/gomega v1.27.10
-	google.golang.org/protobuf v1.36.11
-)
+require google.golang.org/protobuf v1.36.11

go.sum

@@ -1,7 +1,3 @@
-github.com/bsm/ginkgo v1.16.5 h1:uTeeWv0Yx1PnDeCk76PFyGrOMVw3D+r9bTNKNcIjDdQ=
-github.com/bsm/ginkgo v1.16.5/go.mod h1:RabIZLzOCPghgHJKUqHZpqrQETA5AnF4aCSIYy5C1bk=
-github.com/bsm/gomega v1.27.10 h1:yeMWxP2pV2fG3FgAODIY8EiRE3dy0aeFYt4l7wh6yKA=
-github.com/bsm/gomega v1.27.10/go.mod h1:JyEr/xRbxbtgWNi8tIEVPUYZ5Dzef52k01W3YH0H+O0=
 github.com/google/go-cmp v0.7.0 h1:wk8382ETsv4JYUZwIsn6YpYiWiBsYLSJiTsyBybVuN8=
 github.com/google/go-cmp v0.7.0/go.mod h1:pXiqmnSA92OHEEa9HXL2W4E7lf9JzCmGVUdgjX3N/iU=
 google.golang.org/protobuf v1.36.11 h1:fV6ZwhNocDyBLK0dj+fg8ektcVegBBuEolpbTQyBNVE=

hll.go

@@ -4,7 +4,6 @@ import (
 	"fmt"
 
 	"github.com/bsm/zetasketch/hllplus"
-	"github.com/bsm/zetasketch/internal/zetasketch"
 	pb "github.com/bsm/zetasketch/internal/zetasketch"
 	"google.golang.org/protobuf/proto"
 )
@@ -85,7 +84,7 @@ func (h *HLL) proto() *pb.AggregatorStateProto {
 		EncodingVersion: &encodingVersion,
 		NumValues:       &numValues,
 	}
-	proto.SetExtension(msg, zetasketch.E_HyperloglogplusUniqueState, h.h.Proto())
+	proto.SetExtension(msg, pb.E_HyperloglogplusUniqueState, h.h.Proto())
 	return msg
 }
 
@@ -100,7 +99,7 @@ func (h *HLL) fromProto(msg *pb.AggregatorStateProto) error {
 		return fmt.Errorf("incompatible binary message: no num values")
 	}
 
-	ext := proto.GetExtension(msg, zetasketch.E_HyperloglogplusUniqueState)
+	ext := proto.GetExtension(msg, pb.E_HyperloglogplusUniqueState)
 	hState, ok := ext.(*pb.HyperLogLogPlusUniqueStateProto)
 	if !ok {
 		return fmt.Errorf("incompatible binary message: invalid HyperLogLog++ state")

hll_test.go

@@ -1,57 +1,81 @@
 package zetasketch_test
 
 import (
-	"github.com/bsm/zetasketch"
+	"testing"
 
-	. "github.com/bsm/ginkgo"
-	. "github.com/bsm/gomega"
+	"github.com/bsm/zetasketch"
 )
 
-var _ = Describe("HLL", func() {
-	var subject *zetasketch.HLL
-	var _ zetasketch.Aggregator = subject
-
-	BeforeEach(func() {
-		subject = zetasketch.NewHLL(nil)
-
-		for i := 0; i < 1_000; i++ {
-			subject.Add(zetasketch.Uint64Value(uint64(i)))
-		}
-		for i := 500; i < 1_000; i++ {
-			subject.Add(zetasketch.Uint64Value(uint64(i)))
-		}
-	})
-
-	It("should count values", func() {
-		Expect(subject.NumValues()).To(BeNumerically("==", 1_500))
-	})
-
-	It("should estimate uniques", func() {
-		Expect(subject.Result()).To(BeNumerically("==", 1_000))
-	})
-
-	It("should merge", func() {
-		other := zetasketch.NewHLL(nil)
-		for i := 800; i < 1_200; i++ {
-			other.Add(zetasketch.Uint64Value(uint64(i)))
-		}
-
-		Expect(subject.Merge(other)).To(Succeed())
-		Expect(subject.NumValues()).To(BeNumerically("==", 1_900))
-		Expect(subject.Result()).To(BeNumerically("==", 1_207))
-
-		// `other` is not modified:
-		Expect(other.NumValues()).To(BeNumerically("==", 400))
-		Expect(other.Result()).To(BeNumerically("==", 400))
-	})
-
-	It("should marshal/unmarshal binary", func() {
-		data, err := subject.MarshalBinary()
-		Expect(err).NotTo(HaveOccurred())
-
-		subject = new(zetasketch.HLL)
-		Expect(subject.UnmarshalBinary(data)).To(Succeed())
-		Expect(subject.NumValues()).To(BeNumerically("==", 1_500))
-		Expect(subject.Result()).To(BeNumerically("==", 1_000))
-	})
-})
+var _ zetasketch.Aggregator = (*zetasketch.HLL)(nil)
+
+func newTestHLL() *zetasketch.HLL {
+	subject := zetasketch.NewHLL(nil)
+	for i := range 1_000 {
+		subject.Add(zetasketch.Uint64Value(uint64(i)))
+	}
+	for i := 500; i < 1_000; i++ {
+		subject.Add(zetasketch.Uint64Value(uint64(i)))
+	}
+	return subject
+}
+
+func TestHLL_NumValues(t *testing.T) {
+	subject := newTestHLL()
+	if got, exp := subject.NumValues(), int64(1_500); got != exp {
+		t.Errorf("got %d, want %d", got, exp)
+	}
+}
+
+func TestHLL_Result(t *testing.T) {
+	subject := newTestHLL()
+	if got, exp := subject.Result(), int64(1_000); got != exp {
+		t.Errorf("got %d, want %d", got, exp)
+	}
+}
+
+func TestHLL_Merge(t *testing.T) {
+	subject := newTestHLL()
+
+	other := zetasketch.NewHLL(nil)
+	for i := 800; i < 1_200; i++ {
+		other.Add(zetasketch.Uint64Value(uint64(i)))
+	}
+
+	if err := subject.Merge(other); err != nil {
+		t.Fatal(err)
+	}
+	if got, exp := subject.NumValues(), int64(1_900); got != exp {
+		t.Errorf("NumValues: got %d, want %d", got, exp)
+	}
+	if got, exp := subject.Result(), int64(1_207); got != exp {
+		t.Errorf("Result: got %d, want %d", got, exp)
+	}
+
+	// `other` is not modified:
+	if got, exp := other.NumValues(), int64(400); got != exp {
+		t.Errorf("other.NumValues: got %d, want %d", got, exp)
+	}
+	if got, exp := other.Result(), int64(400); got != exp {
+		t.Errorf("other.Result: got %d, want %d", got, exp)
+	}
+}
+
+func TestHLL_MarshalBinary(t *testing.T) {
+	subject := newTestHLL()
+
+	data, err := subject.MarshalBinary()
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	subject = new(zetasketch.HLL)
+	if err := subject.UnmarshalBinary(data); err != nil {
+		t.Fatal(err)
+	}
+	if got, exp := subject.NumValues(), int64(1_500); got != exp {
+		t.Errorf("NumValues: got %d, want %d", got, exp)
+	}
+	if got, exp := subject.Result(), int64(1_000); got != exp {
+		t.Errorf("Result: got %d, want %d", got, exp)
+	}
+}

hllplus/data.go

@@ -17,7 +17,6 @@
 package hllplus
 
 import (
-	"math"
 	"math/bits"
 	"sort"
 )
@@ -748,9 +747,10 @@ func closestBiases(estimate float64, precision uint8) weightedBiases {
 
 	res := make(weightedBiases, 0, max-min)
 	for i := min; i < max; i++ {
+		d := means[i] - estimate
 		res = append(res, weightedBias{
 			Bias:     biases[i],
-			Distance: math.Pow(means[i]-estimate, 2),
+			Distance: d * d,
 		})
 	}
 

hllplus/data_test.go

@@ -1,24 +1,35 @@
 package hllplus_test
 
 import (
-	"github.com/bsm/zetasketch/hllplus"
+	"math"
+	"testing"
 
-	. "github.com/bsm/ginkgo"
-	. "github.com/bsm/gomega"
+	"github.com/bsm/zetasketch/hllplus"
 )
 
-var _ = Describe("Data", func() {
-	It("should estimate bias", func() {
-		Expect(hllplus.EstimateBias(0, 15)).To(BeNumerically("==", 0.0))
-		Expect(hllplus.EstimateBias(1, 15)).To(BeNumerically("==", 0.0))
-		Expect(hllplus.EstimateBias(10_000, 15)).To(BeNumerically("==", 0.0))
-		Expect(hllplus.EstimateBias(100_000, 15)).To(BeNumerically("~", 888.1, 0.1))
-		Expect(hllplus.EstimateBias(200_000, 15)).To(BeNumerically("==", 0.0))
+func TestEstimateBias(t *testing.T) {
+	cases := []struct {
+		e     float64
+		p     uint8
+		exp   float64
+		delta float64
+	}{
+		{0, 15, 0.0, 0},
+		{1, 15, 0.0, 0},
+		{10_000, 15, 0.0, 0},
+		{100_000, 15, 888.1, 0.1},
+		{200_000, 15, 0.0, 0},
 
-		Expect(hllplus.EstimateBias(50_000, 13)).To(BeNumerically("==", 0.0))
-		Expect(hllplus.EstimateBias(50_000, 14)).To(BeNumerically("~", 449.7, 0.1))
-		Expect(hllplus.EstimateBias(50_000, 15)).To(BeNumerically("~", 7820.2, 0.1))
-		Expect(hllplus.EstimateBias(50_000, 16)).To(BeNumerically("~", 44513.2, 0.1))
-		Expect(hllplus.EstimateBias(50_000, 17)).To(BeNumerically("==", 0.0))
-	})
-})
+		{50_000, 13, 0.0, 0},
+		{50_000, 14, 449.7, 0.1},
+		{50_000, 15, 7820.2, 0.1},
+		{50_000, 16, 44513.2, 0.1},
+		{50_000, 17, 0.0, 0},
+	}
+	for _, tc := range cases {
+		got := hllplus.EstimateBias(tc.e, tc.p)
+		if math.Abs(got-tc.exp) > tc.delta {
+			t.Errorf("EstimateBias(%v, %d) = %v, want %v (±%v)", tc.e, tc.p, got, tc.exp, tc.delta)
+		}
+	}
+}

hllplus/ext_test.go

@@ -6,10 +6,7 @@ func EstimateBias(e float64, p uint8) float64 {
 }
 
 func NewNormal(precision uint8) (*HLL, error) {
-	pp := precision + 5
-	if pp > MaxSparsePrecision {
-		pp = MaxSparsePrecision
-	}
+	pp := min(precision+5, MaxSparsePrecision)
 
 	s, err := New(precision, pp)
 	if err != nil {