Add tesseract support for docker container deployments. Add dockerfile with tesseract and languages installed.

k · k · commit 458ccda335e5 · 2024-12-12T02:21:35.000-08:00
diff --git a/Dockerfile-tesseract b/Dockerfile-tesseract
@@ -0,0 +1,15 @@
+ARG VERSION=latest
+
+FROM b3log/siyuan:${VERSION}
+
+RUN apk update && apk add --no-cache --update tesseract-ocr
+
+RUN cd /usr/share/tessdata \
+	&& wget https://github.com/tesseract-ocr/tessdata/raw/main/eng.traineddata \
+	&& wget https://github.com/tesseract-ocr/tessdata/raw/main/chi_sim.traineddata \
+	&& wget https://github.com/tesseract-ocr/tessdata/raw/main/chi_tra.traineddata
+	# Add more langs if needed
+
+ENTRYPOINT ["/opt/siyuan/entrypoint.sh"]
+
+CMD ["/opt/siyuan/kernel"]
diff --git a/kernel/util/ocr.go b/kernel/util/ocr.go
@@ -187,7 +187,7 @@ func IsTesseractExtractable(p string) bool {
 var tesseractOCRLock = sync.Mutex{}
 
 func Tesseract(imgAbsPath string) (ret []map[string]interface{}) {
-	if ContainerStd != Container || !TesseractEnabled {
+	if !TesseractEnabled {
 		return
 	}
 
@@ -343,8 +343,23 @@ func filterTesseractLangs(langs []string) (ret []string) {
 	return ret
 }
 
+func isContainerTesseractSupported(container string) bool {
+	supportedContainers := []string{
+		ContainerStd,
+		ContainerDocker,
+	}
+
+	for _, v := range supportedContainers {
+		if v == container {
+			return true
+		}
+	}
+
+	return false
+}
+
 func getTesseractVer() (ret string) {
-	if ContainerStd != Container {
+	if !isContainerTesseractSupported(Container) {
 		return
 	}
 
