Allow setting of title in PDF metadata

rmtheis · Aug 23, 2015 · 9e99699 · 9e99699
1 parent 42bde29
commit 9e99699
Show file tree

Hide file tree

Showing 6 changed files with 40 additions and 11 deletions.
diff --git a/tess-two-test/src/com/googlecode/tesseract/android/test/TessPdfRendererTest.java b/tess-two-test/src/com/googlecode/tesseract/android/test/TessPdfRendererTest.java
@@ -69,7 +69,7 @@ public void testAddPageToDocument() throws IOException {
                 + pdfBasename);
 
         // Start the PDF writing process.
-        boolean beginSuccess = baseApi.beginDocument(pdfRenderer);
+        boolean beginSuccess = baseApi.beginDocument(pdfRenderer, "title");
         // assertTrue(beginSuccess);
 
         // Add a page to the PDF.

diff --git a/tess-two/jni/com_googlecode_tesseract_android/src/api/baseapi.cpp b/tess-two/jni/com_googlecode_tesseract_android/src/api/baseapi.cpp
@@ -987,8 +987,7 @@ bool TessBaseAPI::ProcessPagesFileList(FILE *flist,
   }
 
   // Begin producing output
-  const char* kUnknownTitle = "";
-  if (renderer && !renderer->BeginDocument(kUnknownTitle)) {
+  if (renderer && !renderer->BeginDocument(unknown_title_)) {
     return false;
   }
 
@@ -1162,8 +1161,7 @@ bool TessBaseAPI::ProcessPagesInternal(const char* filename,
   }
 
   // Begin the output
-  const char* kUnknownTitle = "";
-  if (renderer && !renderer->BeginDocument(kUnknownTitle)) {
+  if (renderer && !renderer->BeginDocument(unknown_title_)) {
     pixDestroy(&pix);
     return false;
   }

diff --git a/tess-two/jni/com_googlecode_tesseract_android/src/api/baseapi.h b/tess-two/jni/com_googlecode_tesseract_android/src/api/baseapi.h
@@ -879,6 +879,12 @@ class TESS_API TessBaseAPI {
                                  int timeout_millisec,
                                  TessResultRenderer* renderer,
                                  int tessedit_page_number);
+  // There's currently no way to pass a document title from the
+  // Tesseract command line, and we have multiple places that choose
+  // to set the title to an empty string. Using a single named
+  // variable will hopefully reduce confusion if the situation changes
+  // in the future.
+  const char *unknown_title_ = "";
 };  // class TessBaseAPI.
 
 /** Escape a char string - remove &<>"' with HTML codes. */

diff --git a/tess-two/jni/com_googlecode_tesseract_android/src/api/renderer.h b/tess-two/jni/com_googlecode_tesseract_android/src/api/renderer.h
@@ -77,7 +77,7 @@ class TESS_API TessResultRenderer {
     bool EndDocument();
 
     const char* file_extension() const { return file_extension_; }
-    const char* title() const { return title_; }
+    const char* title() const { return title_.c_str(); }
 
     /**
      * Returns the index of the last image given to AddImage
@@ -126,7 +126,7 @@ class TESS_API TessResultRenderer {
 
   private:
     const char* file_extension_;  // standard extension for generated output
-    const char* title_;           // title of document being renderered
+    STRING title_;                // title of document being renderered
     int imagenum_;                // index of last image added
 
     FILE* fout_;                  // output file pointer

diff --git a/tess-two/jni/com_googlecode_tesseract_android/tessbaseapi.cpp b/tess-two/jni/com_googlecode_tesseract_android/tessbaseapi.cpp
@@ -612,10 +612,21 @@ void Java_com_googlecode_tesseract_android_TessPdfRenderer_nativeRecycle(JNIEnv
 
 jboolean Java_com_googlecode_tesseract_android_TessBaseAPI_nativeBeginDocument(JNIEnv *env,
                                                                                jobject thiz,
-                                                                               jlong jRenderer) {
+                                                                               jlong jRenderer,
+                                                                               jstring title) {
 
+  const char *c_title = env->GetStringUTFChars(title, NULL);
   tesseract::TessPDFRenderer* pdfRenderer = (tesseract::TessPDFRenderer*) jRenderer;
-  return pdfRenderer->BeginDocument("");
+
+  jboolean res = JNI_TRUE;
+
+  if (pdfRenderer->BeginDocument(c_title)) {
+    res = JNI_FALSE;
+  }
+
+  env->ReleaseStringUTFChars(title, c_title);
+
+  return res;
 }
 
 jboolean Java_com_googlecode_tesseract_android_TessBaseAPI_nativeEndDocument(JNIEnv *env,

diff --git a/tess-two/src/com/googlecode/tesseract/android/TessBaseAPI.java b/tess-two/src/com/googlecode/tesseract/android/TessBaseAPI.java
@@ -856,11 +856,25 @@ protected void onProgressValues(final int percent, final int left,
 
     /**
      * Starts a new document. This clears the contents of the output data.
+     * 
+     * Caller is responsible for escaping the provided title.
      *
+     * @param title a title to be used in the document metadata
+     * @return {@code true} on success. {@code false} on failure
+     */
+    public boolean beginDocument(TessPdfRenderer tessPdfRenderer, String title) {
+        return nativeBeginDocument(tessPdfRenderer.getNativePdfRenderer(), 
+                title);
+    }
+
+    /**
+     * Starts a new document with no title.
+     * 
      * @return {@code true} on success. {@code false} on failure
+     * @see #beginDocument(TessPdfRenderer, String)
      */
     public boolean beginDocument(TessPdfRenderer tessPdfRenderer) {
-        return nativeBeginDocument(tessPdfRenderer.getNativePdfRenderer());
+        return nativeBeginDocument(tessPdfRenderer.getNativePdfRenderer(), "");
     }
 
     /**
@@ -963,7 +977,7 @@ private native void nativeSetImageBytes(
 
     private native int nativeStop();
 
-    private native boolean nativeBeginDocument(long rendererPointer);
+    private native boolean nativeBeginDocument(long rendererPointer, String title);
 
     private native boolean nativeEndDocument(long rendererPointer);