docker changes

bradparks · Dec 10, 2023 · 1f29033 · 1f29033
1 parent 59c7978
commit 1f29033
Show file tree

Hide file tree

Showing 21 changed files with 32 additions and 80 deletions.
diff --git a/Dockerfile b/Dockerfile
@@ -1,5 +1,5 @@
 # Use the base image
-FROM frooodle/stirling-pdf-base:testDontUseMe
+FROM frooodle/stirling-pdf-base:version6
 
 ARG VERSION_TAG
 

diff --git a/DockerfileBase b/DockerfileBase
@@ -1,48 +1,50 @@
 # Main stage
 FROM ubuntu:latest AS base
 
-RUN apt-get update && apt-get install -y --no-install-recommends software-properties-common gnupg2
 
-RUN add-apt-repository ppa:alex-p/tesseract-ocr5 && apt install -y --no-install-recommends tesseract-ocr
 
+# JDK for app
 RUN apt-get update && \
     apt-get install -y --no-install-recommends \
-    	openjdk-17-jre \
-        libreoffice-core-nogui \
+    	openjdk-17-jre
+
+
+# Doc conversion
+RUN apt-get update && \
+    apt-get install -y --no-install-recommends \
+    	libreoffice-core-nogui \
         libreoffice-common \
         libreoffice-writer-nogui \
         libreoffice-calc-nogui \
         libreoffice-impress-nogui \
-        python3-uno \
+		python3-uno \
+		unoconv
+
+
+# OCR MY PDF (unpaper for descew and other advanced featues)
+RUN apt-get update && apt-get install -y --no-install-recommends software-properties-common gnupg2 && \
+add-apt-repository ppa:alex-p/tesseract-ocr5 && apt install -y --no-install-recommends tesseract-ocr && \
+apt-get update && \
+    apt-get install -y --no-install-recommends \
         ghostscript \
         python3-pip \
         ocrmypdf \
-        unoconv && \
+        unpaper && \
     pip install --upgrade pip && \
-    pip install --no-cache-dir --user --upgrade ocrmypdf && \
-    pip install --no-cache-dir --upgrade pillow==10.0.1 reportlab==3.6.13 wheel==0.38.1 setuptools==65.5.1 pyjwt==2.4.0 cryptography==39.0.1 \
     pip install --no-cache-dir --upgrade ocrmypdf && \
-    pip install --no-cache-dir \
-        opencv-python-headless && \
-    rm -rf /var/lib/apt/lists/* && \
+    pip install --no-cache-dir --upgrade pillow==10.0.1 reportlab==3.6.13 wheel==0.38.1 setuptools==65.5.1 pyjwt==2.4.0 cryptography==39.0.1
+
+
+#CV    
+RUN pip install --no-cache-dir opencv-python-headless
+
+
+# cleanup and etc
+RUN rm -rf /var/lib/apt/lists/* && \
     mkdir /usr/share/tesseract-ocr-original && \
     cp -r /usr/share/tesseract-ocr/* /usr/share/tesseract-ocr-original && \
     rm -rf /usr/share/tesseract-ocr
 
 
-# Python packages stage
-FROM base AS python-packages
-# Install build tools and Python libraries
-RUN apt-get update && \
-    apt-get install -y --no-install-recommends \
-        build-essential \
-        libffi-dev \
-        libssl-dev \
-        zlib1g-dev \
-        libjpeg-dev 
-
-# Final stage: Copy necessary files from the previous stage
-FROM base
-COPY --from=python-packages /usr/local /usr/local
 
 
diff --git a/scripts/init.sh b/scripts/init.sh
@@ -5,6 +5,10 @@ echo "Copying original files without overwriting existing files"
 mkdir -p /usr/share/tesseract-ocr
 cp -rn /usr/share/tesseract-ocr-original/* /usr/share/tesseract-ocr
 
+if [ -d /usr/share/tesseract-ocr/4.00/tessdata ]; then 
+	cp -r /usr/share/tesseract-ocr/4.00/tessdata/* /usr/share/tesseract-ocr/5/tessdata/ || true; 
+fi
+
 # Check if TESSERACT_LANGS environment variable is set and is not empty
 if [[ -n "$TESSERACT_LANGS" ]]; then
   # Convert comma-separated values to a space-separated list

diff --git a/src/main/resources/messages_ar_AR.properties b/src/main/resources/messages_ar_AR.properties
@@ -831,8 +831,5 @@ PDFToXML.submit=تحويل
 #PDFToCSV
 PDFToCSV.title=PDF ??? CSV
 PDFToCSV.header=PDF ??? CSV
-##########################
-###  TODO: Translate   ###
-##########################
 PDFToCSV.prompt=Choose page to extract table
 PDFToCSV.submit=??????
diff --git a/src/main/resources/messages_bg_BG.properties b/src/main/resources/messages_bg_BG.properties
@@ -831,8 +831,5 @@ PDFToXML.submit=Преобразуване
 #PDFToCSV
 PDFToCSV.title=PDF ??? CSV
 PDFToCSV.header=PDF ??? CSV
-##########################
-###  TODO: Translate   ###
-##########################
 PDFToCSV.prompt=Choose page to extract table
 PDFToCSV.submit=????????
diff --git a/src/main/resources/messages_ca_CA.properties b/src/main/resources/messages_ca_CA.properties
@@ -831,8 +831,5 @@ PDFToXML.submit=Converteix
 #PDFToCSV
 PDFToCSV.title=PDF a CSV
 PDFToCSV.header=PDF a CSV
-##########################
-###  TODO: Translate   ###
-##########################
 PDFToCSV.prompt=Choose page to extract table
 PDFToCSV.submit=Extracte
diff --git a/src/main/resources/messages_de_DE.properties b/src/main/resources/messages_de_DE.properties
@@ -831,8 +831,5 @@ PDFToXML.submit=Konvertieren
 #PDFToCSV
 PDFToCSV.title=PDF zu CSV
 PDFToCSV.header=PDF zu CSV
-##########################
-###  TODO: Translate   ###
-##########################
 PDFToCSV.prompt=Choose page to extract table
 PDFToCSV.submit=Extrakt
diff --git a/src/main/resources/messages_el_GR.properties b/src/main/resources/messages_el_GR.properties
@@ -831,8 +831,5 @@ PDFToXML.submit=\u039C\u03B5\u03C4\u03B1\u03C4\u03C1\u03BF\u03C0\u03AE
 #PDFToCSV
 PDFToCSV.title=PDF ?? CSV
 PDFToCSV.header=PDF ?? CSV
-##########################
-###  TODO: Translate   ###
-##########################
 PDFToCSV.prompt=Choose page to extract table
 PDFToCSV.submit=?????????
diff --git a/src/main/resources/messages_en_US.properties b/src/main/resources/messages_en_US.properties
@@ -831,8 +831,5 @@ PDFToXML.submit=Convert
 #PDFToCSV
 PDFToCSV.title=PDF to CSV
 PDFToCSV.header=PDF to CSV
-##########################
-###  TODO: Translate   ###
-##########################
 PDFToCSV.prompt=Choose page to extract table
 PDFToCSV.submit=Extract
diff --git a/src/main/resources/messages_es_ES.properties b/src/main/resources/messages_es_ES.properties
@@ -831,8 +831,5 @@ PDFToXML.submit=Convertir
 #PDFToCSV
 PDFToCSV.title=PDF a CSV
 PDFToCSV.header=PDF a CSV
-##########################
-###  TODO: Translate   ###
-##########################
 PDFToCSV.prompt=Choose page to extract table
 PDFToCSV.submit=Extracto
diff --git a/src/main/resources/messages_eu_ES.properties b/src/main/resources/messages_eu_ES.properties
@@ -831,8 +831,5 @@ PDFToXML.submit=Bihurtu
 #PDFToCSV
 PDFToCSV.title=PDF a CSV
 PDFToCSV.header=PDF a CSV
-##########################
-###  TODO: Translate   ###
-##########################
 PDFToCSV.prompt=Choose page to extract table
 PDFToCSV.submit=Extracto
diff --git a/src/main/resources/messages_fr_FR.properties b/src/main/resources/messages_fr_FR.properties
@@ -831,8 +831,5 @@ PDFToXML.submit=Convertir
 #PDFToCSV
 PDFToCSV.title=PDF en CSV
 PDFToCSV.header=PDF en CSV
-##########################
-###  TODO: Translate   ###
-##########################
 PDFToCSV.prompt=Choose page to extract table
 PDFToCSV.submit=Extrait
diff --git a/src/main/resources/messages_it_IT.properties b/src/main/resources/messages_it_IT.properties
@@ -831,8 +831,5 @@ PDFToXML.submit=Converti
 #PDFToCSV
 PDFToCSV.title=Da PDF a CSV
 PDFToCSV.header=Da PDF a CSV
-##########################
-###  TODO: Translate   ###
-##########################
 PDFToCSV.prompt=Choose page to extract table
 PDFToCSV.submit=Estratto
diff --git a/src/main/resources/messages_ja_JP.properties b/src/main/resources/messages_ja_JP.properties
@@ -831,8 +831,5 @@ PDFToXML.submit=変換
 #PDFToCSV
 PDFToCSV.title=PDF??CSV?
 PDFToCSV.header=PDF??CSV?
-##########################
-###  TODO: Translate   ###
-##########################
 PDFToCSV.prompt=Choose page to extract table
 PDFToCSV.submit=????
diff --git a/src/main/resources/messages_ko_KR.properties b/src/main/resources/messages_ko_KR.properties
@@ -831,8 +831,5 @@ PDFToXML.submit=변환
 #PDFToCSV
 PDFToCSV.title=PDF? CSV?
 PDFToCSV.header=PDF? CSV?
-##########################
-###  TODO: Translate   ###
-##########################
 PDFToCSV.prompt=Choose page to extract table
 PDFToCSV.submit=??
diff --git a/src/main/resources/messages_nl_NL.properties b/src/main/resources/messages_nl_NL.properties
@@ -831,8 +831,5 @@ PDFToXML.submit=Converteren
 #PDFToCSV
 PDFToCSV.title=PDF naar CSV
 PDFToCSV.header=PDF naar CSV
-##########################
-###  TODO: Translate   ###
-##########################
 PDFToCSV.prompt=Choose page to extract table
 PDFToCSV.submit=Extract
diff --git a/src/main/resources/messages_pl_PL.properties b/src/main/resources/messages_pl_PL.properties
@@ -831,8 +831,5 @@ PDFToXML.submit=Konwertuj
 #PDFToCSV
 PDFToCSV.title=PDF na CSV
 PDFToCSV.header=PDF na CSV
-##########################
-###  TODO: Translate   ###
-##########################
 PDFToCSV.prompt=Choose page to extract table
 PDFToCSV.submit=Wyci?g
diff --git a/src/main/resources/messages_ru_RU.properties b/src/main/resources/messages_ru_RU.properties
@@ -831,8 +831,5 @@ PDFToXML.submit=Конвертировать
 #PDFToCSV
 PDFToCSV.title=PDF ? CSV
 PDFToCSV.header=PDF ? CSV
-##########################
-###  TODO: Translate   ###
-##########################
 PDFToCSV.prompt=Choose page to extract table
 PDFToCSV.submit=???????
diff --git a/src/main/resources/messages_sv_SE.properties b/src/main/resources/messages_sv_SE.properties
@@ -831,8 +831,5 @@ PDFToXML.submit=Konvertera
 #PDFToCSV
 PDFToCSV.title=PDF till CSV
 PDFToCSV.header=PDF till CSV
-##########################
-###  TODO: Translate   ###
-##########################
 PDFToCSV.prompt=Choose page to extract table
 PDFToCSV.submit=Navvit
diff --git a/src/main/resources/messages_tr_TR.properties b/src/main/resources/messages_tr_TR.properties
@@ -831,8 +831,5 @@ PDFToXML.submit=Dönüştür
 #PDFToCSV
 PDFToCSV.title=PDF to CSV
 PDFToCSV.header=PDF to CSV
-##########################
-###  TODO: Translate   ###
-##########################
 PDFToCSV.prompt=Choose page to extract table
 PDFToCSV.submit=Extract
diff --git a/src/main/resources/messages_zh_CN.properties b/src/main/resources/messages_zh_CN.properties
@@ -831,8 +831,5 @@ PDFToXML.submit=转换
 #PDFToCSV
 PDFToCSV.title=PDF ? CSV
 PDFToCSV.header=PDF ? CSV
-##########################
-###  TODO: Translate   ###
-##########################
 PDFToCSV.prompt=Choose page to extract table
 PDFToCSV.submit=??