not add space character for CJK.

manisandro · Jun 22, 2018 · fe04fd3 · fe04fd3
1 parent 2a40945
commit fe04fd3
Show file tree

Hide file tree

Showing 3 changed files with 32 additions and 3 deletions.
diff --git a/qt/src/Utils.cc b/qt/src/Utils.cc
@@ -221,3 +221,19 @@ QString Utils::getSpellingLanguage(const QString& lang) {
 	}
 	return syslang;
 }
+
+// Unicode blocks http://www.fileformat.info/info/unicode/block/index.htm
+bool Utils::spacedWord(const QString& text, bool prevWord) {
+	short unicode = (prevWord ? text.back() : text.front()).unicode();
+	// CJK Word
+	std::vector<std::pair<int, int>> cjkWordRange{{0x2480, 0x303f}, {0x31c0, 0x9fff}
+		, {0xf900, 0xfaff}, {0xfe30, 0xfe4f}, {0x20000, 0x2fa1f}};
+	for(int i = 0; i < cjkWordRange.size(); i++) {
+		if(unicode < cjkWordRange[i].first) {
+			return true;
+		} else if(unicode >= cjkWordRange[i].first && unicode <= cjkWordRange[i].second) {
+			return false;
+		}
+	}
+	return true;
+}
diff --git a/qt/src/Utils.hh b/qt/src/Utils.hh
@@ -47,6 +47,8 @@ QByteArray download(QUrl url, QString& messages, int timeout = 60000);
 
 QString getSpellingLanguage(const QString& lang = QString());
 
+bool spacedWord(const QString& text, bool prevWord);
+
 template<typename T>
 class AsyncQueue {
 public:

diff --git a/qt/src/hocr/HOCRPdfExporter.cc b/qt/src/hocr/HOCRPdfExporter.cc
@@ -688,6 +688,8 @@ void HOCRPdfExporter::printChildren(PDFPainter& painter, const HOCRItem* item, c
 	QString itemClass = item->itemClass();
 	QRect itemRect = item->bbox();
 	int childCount = item->children().size();
+	bool prevSpacedWord, currentSpacedWord;
+	prevSpacedWord = currentSpacedWord = false;
 	if(itemClass == "ocr_par" && pdfSettings.uniformizeLineSpacing) {
 		double yInc = double(itemRect.height()) / childCount;
 		double y = itemRect.top() + yInc;
@@ -706,15 +708,24 @@ void HOCRPdfExporter::printChildren(PDFPainter& painter, const HOCRItem* item, c
 				if(pdfSettings.fontSize == -1) {
 					painter.setFontSize(wordItem->fontSize() * pdfSettings.detectedFontScaling);
 				}
+
+				prevWordRight = wordRect.right();
+				QString text = wordItem->text();
+				currentSpacedWord = Utils::spacedWord(text, false);
 				// If distance from previous word is large, keep the space
 				if(wordRect.x() - prevWordRight > pdfSettings.preserveSpaceWidth * painter.getAverageCharWidth() / px2pu) {
 					x = wordRect.x();
+				} else {
+					//need space
+					if(currentSpacedWord && prevSpacedWord ) {
+						x += painter.getTextWidth(" ") / px2pu;
+					}
 				}
-				prevWordRight = wordRect.right();
-				QString text = wordItem->text();
+
 				double wordBaseline = (x - itemRect.x()) * baseline.first + baseline.second;
 				painter.drawText(x * px2pu, (y + wordBaseline) * px2pu, text);
-				x += painter.getTextWidth(text + " ") / px2pu;
+				x += painter.getTextWidth(text) / px2pu;
+				prevSpacedWord = Utils::spacedWord(text, true);
 			}
 		}
 	} else if(itemClass == "ocr_line" && !pdfSettings.uniformizeLineSpacing) {