123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338 |
- // SiYuan - Build Your Eternal Digital Garden
- // Copyright (c) 2020-present, b3log.org
- //
- // This program is free software: you can redistribute it and/or modify
- // it under the terms of the GNU Affero General Public License as published by
- // the Free Software Foundation, either version 3 of the License, or
- // (at your option) any later version.
- //
- // This program is distributed in the hope that it will be useful,
- // but WITHOUT ANY WARRANTY; without even the implied warranty of
- // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
- // GNU Affero General Public License for more details.
- //
- // You should have received a copy of the GNU Affero General Public License
- // along with this program. If not, see <https://www.gnu.org/licenses/>.
- package util
- import (
- "bytes"
- "context"
- "io"
- "os"
- "os/exec"
- "path/filepath"
- "regexp"
- "runtime"
- "runtime/debug"
- "strings"
- "sync"
- "time"
- "github.com/88250/gulu"
- "github.com/dustin/go-humanize"
- "github.com/panjf2000/ants/v2"
- "github.com/siyuan-note/logging"
- )
- var (
- tesseractEnabled bool
- tesseractLangs []string
- assetsTexts = map[string]string{}
- assetsTextsLock = sync.Mutex{}
- assetsTextsChanged = false
- )
- func GetAssetText(asset string) string {
- assetsTextsLock.Lock()
- ret, ok := assetsTexts[asset]
- assetsTextsLock.Unlock()
- if ok {
- return ret
- }
- assetsPath := GetDataAssetsAbsPath()
- assetAbsPath := strings.TrimPrefix(asset, "assets")
- assetAbsPath = filepath.Join(assetsPath, assetAbsPath)
- ret = Tesseract(assetAbsPath)
- assetsTextsLock.Lock()
- assetsTexts[asset] = ret
- assetsTextsLock.Unlock()
- return ret
- }
- func Tesseract(imgAbsPath string) string {
- if ContainerStd != Container || !tesseractEnabled {
- return ""
- }
- info, err := os.Stat(imgAbsPath)
- if nil != err {
- return ""
- }
- defer logging.Recover()
- ctx, cancel := context.WithTimeout(context.Background(), 7*time.Second)
- defer cancel()
- now := time.Now()
- cmd := exec.CommandContext(ctx, "tesseract", "-c", "debug_file=/dev/null", imgAbsPath, "stdout", "-l", strings.Join(tesseractLangs, "+"))
- gulu.CmdAttr(cmd)
- output, err := cmd.CombinedOutput()
- if ctx.Err() == context.DeadlineExceeded {
- logging.LogWarnf("tesseract [path=%s, size=%d] timeout", imgAbsPath, info.Size())
- return ""
- }
- if nil != err {
- logging.LogWarnf("tesseract [path=%s, size=%d] failed: %s", imgAbsPath, info.Size(), err)
- return ""
- }
- ret := string(output)
- ret = strings.ReplaceAll(ret, "\r", "")
- ret = strings.ReplaceAll(ret, "\n", "")
- ret = strings.ReplaceAll(ret, "\t", " ")
- reg := regexp.MustCompile("\\s{2,}")
- ret = reg.ReplaceAllString(ret, " ")
- logging.LogInfof("tesseract [path=%s, size=%d, text=%s, elapsed=%dms]", imgAbsPath, info.Size(), ret, time.Since(now).Milliseconds())
- return ret
- }
- func AutoOCRAssets() {
- if !tesseractEnabled {
- return
- }
- for {
- autoOCRAssets()
- time.Sleep(7 * time.Second)
- }
- }
- func autoOCRAssets() {
- defer logging.Recover()
- assetsPath := GetDataAssetsAbsPath()
- assets := getUnOCRAssetsAbsPaths()
- poolSize := runtime.NumCPU()
- if 4 < poolSize {
- poolSize = 4
- }
- waitGroup := &sync.WaitGroup{}
- p, _ := ants.NewPoolWithFunc(poolSize, func(arg interface{}) {
- defer waitGroup.Done()
- assetAbsPath := arg.(string)
- text := Tesseract(assetAbsPath)
- p := strings.TrimPrefix(assetAbsPath, assetsPath)
- p = "assets" + filepath.ToSlash(p)
- assetsTextsLock.Lock()
- assetsTexts[p] = text
- assetsTextsLock.Unlock()
- assetsTextsChanged = true
- })
- for _, assetAbsPath := range assets {
- waitGroup.Add(1)
- p.Invoke(assetAbsPath)
- }
- waitGroup.Wait()
- p.Release()
- cleanNotFoundAssetsTexts()
- }
- func cleanNotFoundAssetsTexts() {
- tmp := assetsTexts
- assetsPath := GetDataAssetsAbsPath()
- var toRemoves []string
- for asset, _ := range tmp {
- assetAbsPath := strings.TrimPrefix(asset, "assets")
- assetAbsPath = filepath.Join(assetsPath, assetAbsPath)
- if !gulu.File.IsExist(assetAbsPath) {
- toRemoves = append(toRemoves, asset)
- }
- }
- assetsTextsLock.Lock()
- for _, asset := range toRemoves {
- delete(assetsTexts, asset)
- assetsTextsChanged = true
- }
- assetsTextsLock.Unlock()
- return
- }
- func getUnOCRAssetsAbsPaths() (ret []string) {
- assetsPath := GetDataAssetsAbsPath()
- var assetsPaths []string
- filepath.Walk(assetsPath, func(path string, info os.FileInfo, err error) error {
- name := info.Name()
- if info.IsDir() {
- if strings.HasPrefix(name, ".") {
- return filepath.SkipDir
- }
- return nil
- }
- lowerName := strings.ToLower(name)
- if !strings.HasSuffix(lowerName, ".png") && !strings.HasSuffix(lowerName, ".jpg") && !strings.HasSuffix(lowerName, ".jpeg") {
- return nil
- }
- assetsPaths = append(assetsPaths, path)
- return nil
- })
- assetsTextsTmp := assetsTexts
- for _, absPath := range assetsPaths {
- p := strings.TrimPrefix(absPath, assetsPath)
- p = "assets" + filepath.ToSlash(p)
- if _, ok := assetsTextsTmp[p]; ok {
- continue
- }
- ret = append(ret, absPath)
- }
- return
- }
- func AutoFlushAssetsTexts() {
- for {
- SaveAssetsTexts()
- time.Sleep(7 * time.Second)
- }
- }
- func LoadAssetsTexts() {
- assetsPath := GetDataAssetsAbsPath()
- assetsTextsPath := filepath.Join(assetsPath, "ocr-texts.json")
- if !gulu.File.IsExist(assetsTextsPath) {
- return
- }
- start := time.Now()
- var err error
- fh, err := os.OpenFile(assetsTextsPath, os.O_RDWR, 0644)
- if nil != err {
- logging.LogErrorf("open assets texts failed: %s", err)
- return
- }
- defer fh.Close()
- data, err := io.ReadAll(fh)
- if nil != err {
- logging.LogErrorf("read assets texts failed: %s", err)
- return
- }
- assetsTextsLock.Lock()
- if err = gulu.JSON.UnmarshalJSON(data, &assetsTexts); nil != err {
- logging.LogErrorf("unmarshal assets texts failed: %s", err)
- if err = os.RemoveAll(assetsTextsPath); nil != err {
- logging.LogErrorf("removed corrupted assets texts failed: %s", err)
- }
- return
- }
- assetsTextsLock.Unlock()
- debug.FreeOSMemory()
- if elapsed := time.Since(start).Seconds(); 2 < elapsed {
- logging.LogWarnf("read assets texts [%s] to [%s], elapsed [%.2fs]", humanize.Bytes(uint64(len(data))), assetsTextsPath, elapsed)
- }
- return
- }
- func SaveAssetsTexts() {
- if !assetsTextsChanged {
- return
- }
- start := time.Now()
- assetsTextsLock.Lock()
- data, err := gulu.JSON.MarshalIndentJSON(assetsTexts, "", " ")
- if nil != err {
- logging.LogErrorf("marshal assets texts failed: %s", err)
- return
- }
- assetsTextsLock.Unlock()
- assetsPath := GetDataAssetsAbsPath()
- assetsTextsPath := filepath.Join(assetsPath, "ocr-texts.json")
- if err = gulu.File.WriteFileSafer(assetsTextsPath, data, 0644); nil != err {
- logging.LogErrorf("write assets texts failed: %s", err)
- return
- }
- debug.FreeOSMemory()
- if elapsed := time.Since(start).Seconds(); 2 < elapsed {
- logging.LogWarnf("save assets texts [size=%s] to [%s], elapsed [%.2fs]", humanize.Bytes(uint64(len(data))), assetsTextsPath, elapsed)
- }
- assetsTextsChanged = false
- }
- func initTesseract() {
- ver := getTesseractVer()
- if "" == ver {
- return
- }
- tesseractLangs = getTesseractLangs()
- if 1 > len(tesseractLangs) {
- logging.LogWarnf("no tesseract langs found")
- tesseractEnabled = false
- return
- }
- logging.LogInfof("tesseract-ocr enabled [ver=%s, langs=%s]", ver, strings.Join(tesseractLangs, "+"))
- }
- func getTesseractVer() (ret string) {
- if ContainerStd != Container {
- return
- }
- cmd := exec.Command("tesseract", "--version")
- gulu.CmdAttr(cmd)
- data, err := cmd.CombinedOutput()
- if nil == err && strings.HasPrefix(string(data), "tesseract ") {
- parts := bytes.Split(data, []byte("\n"))
- if 0 < len(parts) {
- ret = strings.TrimPrefix(string(parts[0]), "tesseract ")
- ret = strings.TrimSpace(ret)
- tesseractEnabled = true
- }
- return
- }
- return
- }
- func getTesseractLangs() (ret []string) {
- if !tesseractEnabled {
- return nil
- }
- cmd := exec.Command("tesseract", "--list-langs")
- gulu.CmdAttr(cmd)
- data, err := cmd.CombinedOutput()
- if nil != err {
- return nil
- }
- parts := bytes.Split(data, []byte("\n"))
- if 0 < len(parts) {
- parts = parts[1:]
- }
- for _, part := range parts {
- part = bytes.TrimSpace(part)
- if 0 == len(part) {
- continue
- }
- ret = append(ret, string(part))
- }
- return
- }
|