WIP: Compute dhash ourselves

4 files changed, 183 insertions, 33 deletions

diff --git a/go.mod b/go.mod
index 6d8c4f9..6aa6b75 100644
--- a/go.mod
+++ b/go.mod
@@ -2,4 +2,7 @@ module janouch.name/gallery
 
 go 1.21.4
 
-require github.com/mattn/go-sqlite3 v1.14.18
+require (
+	github.com/mattn/go-sqlite3 v1.14.19
+	golang.org/x/image v0.14.0
+)
diff --git a/go.sum b/go.sum
index 810a101..59d9ac4 100644
--- a/go.sum
+++ b/go.sum
@@ -1,2 +1,6 @@
 github.com/mattn/go-sqlite3 v1.14.18 h1:JL0eqdCOq6DJVNPSvArO/bIV9/P7fbGrV00LZHc+5aI=
 github.com/mattn/go-sqlite3 v1.14.18/go.mod h1:2eHXhiwb8IkHr+BDWZGa96P6+rkvnG63S2DGjv9HUNg=
+github.com/mattn/go-sqlite3 v1.14.19 h1:fhGleo2h1p8tVChob4I9HpmVFIAkKGpiukdrgQbWfGI=
+github.com/mattn/go-sqlite3 v1.14.19/go.mod h1:2eHXhiwb8IkHr+BDWZGa96P6+rkvnG63S2DGjv9HUNg=
+golang.org/x/image v0.14.0 h1:tNgSxAFe3jC4uYqvZdTr84SZoM1KfwdC9SKIFrLjFn4=
+golang.org/x/image v0.14.0/go.mod h1:HUYqC05R2ZcZ3ejNQsIHQDQiwWM4JBqmm6MKANTp4LE=
diff --git a/main.go b/main.go
index ae0a5f7..c9fef9a 100644
--- a/main.go
+++ b/main.go
@@ -11,9 +11,12 @@ import (
 	"errors"
 	"fmt"
 	"html/template"
+	"image"
+	"image/color"
 	"io"
 	"io/fs"
 	"log"
+	"math"
 	"math/bits"
 	"net"
 	"net/http"
@@ -31,6 +34,8 @@ import (
 	"time"
 
 	"github.com/mattn/go-sqlite3"
+	"golang.org/x/image/draw"
+	"golang.org/x/image/webp"
 )
 
 var (
@@ -54,7 +59,7 @@ func hammingDistance(a, b int64) int {
 func init() {
 	sql.Register("sqlite3_custom", &sqlite3.SQLiteDriver{
 		ConnectHook: func(conn *sqlite3.SQLiteConn) error {
-			return conn.RegisterFunc("hamming", hammingDistance, true)
+			return conn.RegisterFunc("hamming", hammingDistance, true /*pure*/)
 		},
 	})
 }
@@ -439,7 +444,7 @@ type webSimilarImage struct {
 	Paths       []string `json:"paths"`
 }
 
-func getSimilar(sha1 string, pixels int64, distance int) (
+func getSimilar(sha1 string, dhash int64, pixels int64, distance int) (
 	result []webSimilarImage, err error) {
 	// For distance ∈ {0, 1}, this query is quite inefficient.
 	// In exchange, it's generic.
@@ -448,9 +453,8 @@ func getSimilar(sha1 string, pixels int64, distance int) (
 	// so not bothering with IFNULL on them.
 	rows, err := db.Query(`
 		SELECT sha1, width * height, IFNULL(thumbw, 0), IFNULL(thumbh, 0)
-		FROM image
-		WHERE hamming(dhash, (SELECT dhash FROM image WHERE sha1 = ?)) = ?
-		AND sha1 <> ?`, sha1, distance, sha1)
+		FROM image WHERE sha1 <> ? AND dhash IS NOT NULL
+		AND hamming(dhash, ?) = ?`, sha1, dhash, distance)
 	if err != nil {
 		return nil, err
 	}
@@ -475,6 +479,19 @@ func getSimilar(sha1 string, pixels int64, distance int) (
 	return result, rows.Err()
 }
 
+func getSimilarGroups(sha1 string, dhash int64, pixels int64,
+	output map[string][]webSimilarImage) error {
+	var err error
+	for distance := 0; distance <= 1; distance++ {
+		output[fmt.Sprintf("Perceptual distance %d", distance)], err =
+			getSimilar(sha1, dhash, pixels, distance)
+		if err != nil {
+			return err
+		}
+	}
+	return nil
+}
+
 func handleAPISimilar(w http.ResponseWriter, r *http.Request) {
 	var params struct {
 		SHA1 string
@@ -497,10 +514,13 @@ func handleAPISimilar(w http.ResponseWriter, r *http.Request) {
 		result.Info.Paths = paths
 	}
 
-	var width, height int64
+	var (
+		width, height int64
+		dhash         sql.NullInt64
+	)
 	err := db.QueryRow(`
-		SELECT width, height, IFNULL(thumbw, 0), IFNULL(thumbh, 0)
-		FROM image WHERE sha1 = ?`, params.SHA1).Scan(&width, &height,
+		SELECT width, height, dhash, IFNULL(thumbw, 0), IFNULL(thumbh, 0)
+		FROM image WHERE sha1 = ?`, params.SHA1).Scan(&width, &height, &dhash,
 		&result.Info.ThumbW, &result.Info.ThumbH)
 	if err != nil {
 		http.Error(w, err.Error(), http.StatusInternalServerError)
@@ -508,10 +528,9 @@ func handleAPISimilar(w http.ResponseWriter, r *http.Request) {
 	}
 
 	result.Groups = make(map[string][]webSimilarImage)
-	for distance := 0; distance <= 1; distance++ {
-		result.Groups[fmt.Sprintf("Perceptual distance %d", distance)], err =
-			getSimilar(params.SHA1, width*height, distance)
-		if err != nil {
+	if dhash.Valid {
+		if err := getSimilarGroups(
+			params.SHA1, dhash.Int64, width*height, result.Groups); err != nil {
 			http.Error(w, err.Error(), http.StatusInternalServerError)
 			return
 		}
@@ -535,6 +554,7 @@ type webDuplicateImage struct {
 const duplicatesCTE = `WITH
 	duplicated(dhash, count) AS (
 		SELECT dhash, COUNT(*) AS count FROM image
+		WHERE dhash IS NOT NULL
 		GROUP BY dhash HAVING count > 1
 	),
 	multipathed(sha1, count) AS (
@@ -1847,6 +1867,7 @@ func cmdThumbnail(args []string) error {
 			defer wg.Done()
 			if err := makeThumbnailFor(sha1); err != nil {
 				if ee, ok := err.(*exec.ExitError); ok {
+					// FIXME: Not in the goroutine, or lock it.
 					pb.Stop()
 					log.Printf("%s: %s\n", sha1, ee.Stderr)
 					pb.Update()
@@ -1866,27 +1887,133 @@ func cmdThumbnail(args []string) error {
 
 // --- Perceptual hash ---------------------------------------------------------
 
-func makeDhash(hasher, pathThumb string) (uint64, error) {
-	out, err := exec.Command(hasher, pathThumb).Output()
+type linearImage struct {
+	img image.Image
+}
+
+func newLinearImage(img image.Image) *linearImage {
+	return &linearImage{img: img}
+}
+
+func (l *linearImage) ColorModel() color.Model { return l.img.ColorModel() }
+func (l *linearImage) Bounds() image.Rectangle { return l.img.Bounds() }
+
+func unSRGB(c uint32) uint8 {
+	n := float64(c) / 0xffff
+	if n <= 0.04045 {
+		return uint8(n * (255.0 / 12.92))
+	}
+	return uint8(math.Pow((n+0.055)/(1.055), 2.4) * 255.0)
+}
+
+func (l *linearImage) At(x, y int) color.Color {
+	r, g, b, a := l.img.At(x, y).RGBA()
+	return color.RGBA{
+		R: unSRGB(r), G: unSRGB(g), B: unSRGB(b), A: uint8(a >> 8)}
+}
+
+// - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+
+// isWebPAnimation returns whether the given ReadSeeker starts a WebP animation.
+// See https://developers.google.com/speed/webp/docs/riff_container
+func isWebPAnimation(rs io.ReadSeeker) (bool, error) {
+	b := make([]byte, 17)
+	if _, err := rs.Read(b); err != nil {
+		return false, err
+	}
+	if _, err := rs.Seek(0, io.SeekStart); err != nil {
+		return false, err
+	}
+
+	return bytes.Equal(b[:4], []byte("RIFF")) &&
+		bytes.Equal(b[8:16], []byte("WEBPVP8X")) &&
+		b[16]&0b00000010 != 0, nil
+}
+
+var errIsAnimation = errors.New("cannot perceptually hash animations")
+
+func dhashWebP(rs io.ReadSeeker) (uint64, error) {
+	if a, err := isWebPAnimation(rs); err != nil {
+		return 0, err
+	} else if a {
+		return 0, errIsAnimation
+	}
+
+	// Doing this entire thing in Go is SLOW, but convenient.
+	source, err := webp.Decode(rs)
 	if err != nil {
 		return 0, err
 	}
 
+	var (
+		linear  = newLinearImage(source)
+		resized = image.NewNRGBA64(image.Rect(0, 0, 9, 8))
+	)
+	draw.CatmullRom.Scale(resized, resized.Bounds(),
+		linear, linear.Bounds(), draw.Src, nil)
+
 	var hash uint64
-	_, err = fmt.Fscanf(bytes.NewReader(out), "%x", &hash)
-	return hash, err
+	for y := 0; y < 8; y++ {
+		var grey [9]float32
+		for x := 0; x < 9; x++ {
+			rgba := resized.NRGBA64At(x, y)
+			grey[x] = 0.2126*float32(rgba.R) +
+				0.7152*float32(rgba.G) +
+				0.0722*float32(rgba.B)
+		}
+
+		var row uint64
+		if grey[0] < grey[1] {
+			row |= 1 << 7
+		}
+		if grey[1] < grey[2] {
+			row |= 1 << 6
+		}
+		if grey[2] < grey[3] {
+			row |= 1 << 5
+		}
+		if grey[3] < grey[4] {
+			row |= 1 << 4
+		}
+		if grey[4] < grey[5] {
+			row |= 1 << 3
+		}
+		if grey[5] < grey[6] {
+			row |= 1 << 2
+		}
+		if grey[6] < grey[7] {
+			row |= 1 << 1
+		}
+		if grey[7] < grey[8] {
+			row |= 1 << 0
+		}
+		hash = hash<<8 | row
+	}
+	return hash, nil
+}
+
+// - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+
+func makeDhash(sha1 string) (uint64, error) {
+	pathThumb := thumbPath(sha1)
+	f, err := os.Open(pathThumb)
+	if err != nil {
+		return 0, err
+	}
+	defer f.Close()
+	return dhashWebP(f)
 }
 
 // cmdDhash generates perceptual hash from thumbnails.
 func cmdDhash(args []string) error {
 	if len(args) < 1 {
-		return errors.New("usage: GD HASHER [SHA1...]")
+		return errors.New("usage: GD [SHA1...]")
 	}
 	if err := openDB(args[0]); err != nil {
 		return err
 	}
 
-	hasher, hexSHA1 := args[1], args[2:]
+	hexSHA1 := args[1:]
 	if len(hexSHA1) == 0 {
 		var err error
 		hexSHA1, err = dbCollectStrings(`
@@ -1899,23 +2026,35 @@ func cmdDhash(args []string) error {
 	pb := newProgressBar(len(hexSHA1))
 	defer pb.Stop()
 
-	// TODO: Also run the hasher in parallel, once it becomes a problem.
-	// And/or run it in batches, since start-up time of the hasher
-	// poses considerable overhead with large amounts of images.
+	ctx, cancel := context.WithCancelCause(context.Background())
+	wg := sync.WaitGroup{}
 	for _, sha1 := range hexSHA1 {
-		pathThumb := thumbPath(sha1)
-		hash, err := makeDhash(hasher, pathThumb)
-		if err != nil {
-			return err
-		}
-
-		_, err = db.Exec(`UPDATE image SET dhash = ? WHERE sha1 = ?`,
-			int64(hash), sha1)
-		if err != nil {
-			return err
+		if taskSemaphore.acquire(ctx) != nil {
+			break
 		}
 
-		pb.Step()
+		wg.Add(1)
+		go func(sha1 string) {
+			defer taskSemaphore.release()
+			defer wg.Done()
+			if hash, err := makeDhash(sha1); errors.Is(err, errIsAnimation) {
+				// Ignoring this common condition.
+			} else if err != nil {
+				// FIXME: Not in the goroutine, or lock it.
+				pb.Stop()
+				log.Printf("%s: %s\n", sha1, err)
+				pb.Update()
+			} else if _, err = db.Exec(
+				`UPDATE image SET dhash = ? WHERE sha1 = ?`,
+				int64(hash), sha1); err != nil {
+				cancel(err)
+			}
+			pb.Step()
+		}(sha1)
+	}
+	wg.Wait()
+	if ctx.Err() != nil {
+		return context.Cause(ctx)
 	}
 	return nil
 }
diff --git a/public/gallery.js b/public/gallery.js
index ca1e511..970a3bf 100644
--- a/public/gallery.js
+++ b/public/gallery.js
@@ -1,6 +1,10 @@
 'use strict'
 
 function call(method, params) {
+	// TODO: Make it apparent when results result in errors:
+	//  - With responseType == "json", m.request() always expects JSON,
+	//    and error.message is null if it fails, but we can handle it manually.
+	//  - Go can wrap all errors into trivial strings before writing.
 	return m.request({
 		method: "POST",
 		url: `/api/${method}`,