Merge pull request gali8#186 from ws233/master

Added PDF output support. Fixes gali8#172. Fixes gali8#173.

.gitignore

@@ -71,3 +71,4 @@ crashlytics-build.properties
 leptonica-*
 tesseract-*
 
+TesseractOCR/lib

.gitmodules

@@ -0,0 +1,6 @@
+[submodule "TesseractOCR/tesseract-ocr"]
+	path = TesseractOCR/tesseract-ocr
+	url = https://github.com/tesseract-ocr/tesseract.git
+[submodule "TesseractOCR/libtiff-ios"]
+	path = TesseractOCR/libtiff-ios
+	url = https://github.com/ashtons/libtiff-ios.git

Podfile.lock

@@ -7,4 +7,4 @@ DEPENDENCIES:
 SPEC CHECKSUMS:
   Kiwi: f038a6c61f7a9e4d7766bff5717aa3b3fdb75f55
 
-COCOAPODS: 0.36.1
+COCOAPODS: 0.36.4

README.md

@@ -8,7 +8,10 @@ Easy and fast.**
 These are the current versions of the upstream bundled libraries within the framework that this repository provides:
 
 * Tesseract 3.03-rc1 ([homepage](https://code.google.com/p/tesseract-ocr/))
-* Leptonica 1.71 ([homepage](http://leptonica.org/))
+* Leptonica 1.72 ([homepage](http://leptonica.org/))
+* Libtiff 4.0.4 ([homepage](http://www.remotesensing.org/libtiff/))
+* Libpng 1.6.17 ([homepage](http://www.libpng.org/pub/png/libpng.html))
+* Libjpeg 9a ([homepage](http://libjpeg.sourceforge.net/))
 
 Getting Started
 =================

Template Framework Project/Template Framework Project.xcodeproj/project.pbxproj

@@ -312,6 +312,7 @@
 				OTHER_LDFLAGS = (
 					"-lstdc++",
 					"-ObjC",
+					"-lz",
 				);
 				PRODUCT_NAME = "$(TARGET_NAME)";
 				WRAPPER_EXTENSION = app;
@@ -337,6 +338,7 @@
 				OTHER_LDFLAGS = (
 					"-lstdc++",
 					"-ObjC",
+					"-lz",
 				);
 				PRODUCT_NAME = "$(TARGET_NAME)";
 				WRAPPER_EXTENSION = app;

TesseractOCR/G8Tesseract.h

@@ -146,6 +146,14 @@ extern NSInteger const kG8MaxCredibleResolution;
  */
 - (NSString *)recognizedHOCRForPageNumber:(int)pageNumber;
 
+/**
+ *  Produces a PDF output with the pages sent to the function
+ *  @param  images  An array of the input images being recognized and 
+ *                  included into the output PDF file.
+ *  @return NSData representing output PDF file
+ */
+- (NSData *)recognizedPDFForImages:(NSArray*)images;
+
 /**
  *  Run Tesseract's page analysis on the target image.
  *

TesseractOCR/G8Tesseract.mm

@@ -22,6 +22,7 @@
 #import "allheaders.h"
 #import "genericvector.h"
 #import "strngs.h"
+#import "renderer.h"
 
 NSInteger const kG8DefaultResolution = 72;
 NSInteger const kG8MinCredibleResolution = 70;
@@ -644,6 +645,49 @@ - (NSString *)recognizedHOCRForPageNumber:(int)pageNumber {
     return nil;
 }
 
+- (NSData *)recognizedPDFForImages:(NSArray*)images {
+
+  NSString *path = [self.absoluteDataPath stringByAppendingPathComponent:@"tessdata"];
+  tesseract::TessPDFRenderer *renderer = new tesseract::TessPDFRenderer(path.fileSystemRepresentation);
+
+  // Begin producing output
+  const char* kUnknownTitle = "Unknown Title";
+  if (renderer && !renderer->BeginDocument(kUnknownTitle)) {
+    return nil;
+  }
+
+  bool result = YES;
+  for (int page = 0; page < images.count && result; page++) {
+    UIImage *image = images[page];
+    if ([image isKindOfClass:[UIImage class]]) {
+      Pix *pixs = [self pixForImage:image];
+      Pix *pix = pixConvertTo1(pixs, UINT8_MAX / 2);
+      pixDestroy(&pixs);
+
+      const char *pagename = [NSString stringWithFormat:@"page #%i", page].UTF8String;
+      result = _tesseract->ProcessPage(pix, page, pagename, NULL, 0, renderer);
+      pixDestroy(&pix);
+    }
+  }
+
+  //  error
+  if (!result) {
+    return nil;
+  }
+
+  // Finish producing output
+  if (renderer && !renderer->EndDocument()) {
+    return nil;
+  }
+
+  const char *pdfData = NULL;
+  int pdfDataLength = 0;
+  renderer->GetOutput(&pdfData, &pdfDataLength);
+
+  NSData *data = [NSData dataWithBytes:pdfData length:pdfDataLength];
+  return data;
+}
+
 - (UIImage *)imageWithBlocks:(NSArray *)blocks drawText:(BOOL)drawText thresholded:(BOOL)thresholded
 {
     UIImage *image = thresholded ? self.thresholdedImage : self.image;

TesseractOCR/README_howto_compile_libaries.md

@@ -1,5 +1,3 @@
-# Compiling the .a library files your self
-
 
 ### Step 1 - Prerequisites
 First you need to install these tools:
@@ -12,16 +10,12 @@ First you need to install these tools:
 
 
 ### Step 2 - Download and unzip
-Download the source code for `leptonica-1.71` from the Internet here: http://leptonica.org/source/leptonica-1.71.tar.gz.
-Unzip it in this folder.
-
-
-Download the source code for `tesseract-3.03` from the Internet here: https://tesseract-ocr.googlecode.com/archive/3.03-rc1.tar.gz.
+Download the source code for `leptonica-1.72` from the Internet here: http://leptonica.org/source/leptonica-1.72.tar.gz.
 Unzip it in this folder.
 
 
 ### Step 3 - Compile
-Run `build_dependencies.sh` from the terminal. This first compiles leptonica and then tesseract for every architecture iOS/simulator uses (arm7 arm7s arm64 i386 x86_64), and then combines the resulting libs into one library file. It does this for both leptonica and tesseract, so the final results of the script are "libtesseract_all.a", "liblept.a", and "include" directories for both leptonica and tesseract. Finally, the script copies these results into the "lib" and "include" directories inside this directory.
+Run `build_dependencies.sh` from the terminal. This first compiles dependent libraries (png, jpeg, tiff, leptonica) and then tesseract for every architecture iOS/simulator uses (arm7 arm7s arm64 i386 x86_64), and then combines the resulting libs into one library file. It does this for both dependent libraries and tesseract, so the final results of the script are "libpng.a", "libpng16.a", "libjpeg.a", "libtiff.a", "libtiffxx.a", "libtesseract_all.a", "liblept.a", and "include" directories for both leptonica and tesseract. Finally, the script copies these results into the "lib" and "include" directories inside this directory.
 
 
 ### Step 4 - Build

TesseractOCR/build_dependencies.sh

@@ -7,8 +7,9 @@
 
 GLOBAL_OUTDIR="`pwd`/build"
 LOCAL_OUTDIR="./outdir"
-LEPTON_LIB="`pwd`/leptonica-1.71"
-TESSERACT_LIB="`pwd`/tesseract-3.03"
+TESSERACT_LIB="`pwd`/tesseract-ocr"
+LEPTON_LIB="`pwd`/leptonica-1.72"
+IMAGE_LIBS="`pwd`/libtiff-ios/dependencies"
 
 IOS_BASE_SDK="7.0"
 IOS_DEPLOY_TGT="7.0"
@@ -37,9 +38,10 @@ arch_names=(arm-apple-darwin7 arm-apple-darwin7s arm-apple-darwin64 i386-apple-d
 
 setenv_all() {
 # Add internal libs
-export CFLAGS="$CFLAGS -I$GLOBAL_OUTDIR/include -L$GLOBAL_OUTDIR/lib -Qunused-arguments"
+export LIBS="-lz -lpng -ljpeg -ltiff"
+export CFLAGS="$CFLAGS -I$IMAGE_LIBS/include -L$IMAGE_LIBS/lib -I$GLOBAL_OUTDIR/include -L$GLOBAL_OUTDIR/lib -Qunused-arguments"
 
-export LDFLAGS="-L$SDKROOT/usr/lib/"
+export LDFLAGS="-L$SDKROOT/usr/lib/ -L$IMAGE_LIBS/lib/"
 
 export CPPFLAGS=$CFLAGS
 export CXXFLAGS=$CFLAGS
@@ -126,6 +128,25 @@ cd -
 
 rm -rf $GLOBAL_OUTDIR lib include
 
+#######################
+# Download dependencies
+#######################
+
+git submodule init
+git submodule update
+
+#######################
+# Build libtiff and all of it's dependencies
+#######################
+
+cd libtiff-ios
+./build-png.sh
+./build-jpg.sh
+./build-tiff.sh
+cd ..
+
+mkdir -p $GLOBAL_OUTDIR/lib && cp -rvf $IMAGE_LIBS/lib/lib*.a $GLOBAL_OUTDIR/lib
+
 #######################
 # LEPTONLIB
 #######################
@@ -138,7 +159,8 @@ mkdir -p "$LOCAL_OUTDIR/${archs[$n]}"
 make clean 2> /dev/null
 make distclean 2> /dev/null
 eval "setenv_${archs[$n]}"
-./configure --host="${arch_names[$n]}" --enable-shared=no --disable-programs --without-zlib --without-libpng --without-jpeg --without-giflib --without-libtiff
+env
+./configure --host="${arch_names[$n]}" --enable-shared=no --disable-programs --with-zlib --with-libpng --with-jpeg --without-giflib --with-libtiff
 make -j12
 cp -rvf src/.libs/lib*.a "$LOCAL_OUTDIR/${archs[$n]}"
 done
@@ -182,6 +204,6 @@ cd ..
 
 cp -rf $GLOBAL_OUTDIR/include .
 mkdir -p lib
-cp -rf $GLOBAL_OUTDIR/lib/libtesseract_all.a $GLOBAL_OUTDIR/lib/liblept.a lib/
+cp -rf $GLOBAL_OUTDIR/lib/lib*.a lib/
 
 echo "Finished!"