List of usage examples for org.apache.pdfbox.text TextPosition getUnicode
public String getUnicode()
From source file:PDFTextExtract.java
License:Apache License
/** * Override the default functionality of PDFTextStripper. */// w w w .j a v a 2 s .com @Override protected void writeString(String _string, List<TextPosition> textPositions) throws IOException { for (TextPosition text : textPositions) { { _tmp.add(new Text(text.getXDirAdj(), text.getYDirAdj(), text.getFontSize(), text.getXScale(), text.getHeightDir(), text.getWidthOfSpace(), text.getWidthDirAdj(), text.getUnicode())); } } }
From source file:com.formkiq.core.service.generator.pdfbox.TextToPDFieldMapper.java
License:Apache License
/** * Remove Non Printable Characters and extra spaces. * @param textPositions {@link List} of {@link TextPosition} * @return {@link List} of {@link TextPosition} *//*from www . j av a2 s . com*/ private List<TextPosition> removeNonPrintableAndExtraSpaces(final List<TextPosition> textPositions) { List<TextPosition> list = textPositions.stream() .filter(s -> cleanTextContent(s.getUnicode()).equals(s.getUnicode())).collect(Collectors.toList()); int c = 0; Iterator<TextPosition> itr = list.iterator(); while (itr.hasNext()) { TextPosition tp = itr.next(); if (isEmpty(tp.getUnicode().trim())) { c++; if (c > 2) { itr.remove(); } } else { c = 0; } } return list; }
From source file:com.repeatability.pdf.PDFTextStripper.java
License:Apache License
/** * This will print the text of the processed page to "output". It will estimate, based on the coordinates of the * text, where newlines and word spacings should be placed. The text will be sorted only if that feature was * enabled.//from w w w.j av a2s . c o m * * @throws IOException If there is an error writing the text. */ protected void writePage() throws IOException { float maxYForLine = MAX_Y_FOR_LINE_RESET_VALUE; float minYTopForLine = MIN_Y_TOP_FOR_LINE_RESET_VALUE; float endOfLastTextX = END_OF_LAST_TEXT_X_RESET_VALUE; float lastWordSpacing = LAST_WORD_SPACING_RESET_VALUE; float maxHeightForLine = MAX_HEIGHT_FOR_LINE_RESET_VALUE; PositionWrapper lastPosition = null; PositionWrapper lastLineStartPosition = null; boolean startOfPage = true; // flag to indicate start of page boolean startOfArticle; if (charactersByArticle.size() > 0) { writePageStart(); } for (List<TextPosition> textList : charactersByArticle) { if (getSortByPosition()) { TextPositionComparator comparator = new TextPositionComparator(); // because the TextPositionComparator is not transitive, but // JDK7+ enforces transitivity on comparators, we need to use // a custom quicksort implementation (which is slower, unfortunately). if (useCustomQuickSort) { QuickSort.sort(textList, comparator); } else { Collections.sort(textList, comparator); } } Iterator<TextPosition> textIter = textList.iterator(); startArticle(); startOfArticle = true; // Now cycle through to print the text. // We queue up a line at a time before we print so that we can convert // the line from presentation form to logical form (if needed). List<LineItem> line = new ArrayList<LineItem>(); textIter = textList.iterator(); // start from the beginning again // PDF files don't always store spaces. We will need to guess where we should add // spaces based on the distances between TextPositions. Historically, this was done // based on the size of the space character provided by the font. In general, this // worked but there were cases where it did not work. Calculating the average character // width and using that as a metric works better in some cases but fails in some cases // where the spacing worked. So we use both. NOTE: Adobe reader also fails on some of // these examples. // Keeps track of the previous average character width float previousAveCharWidth = -1; while (textIter.hasNext()) { TextPosition position = textIter.next(); PositionWrapper current = new PositionWrapper(position); String characterValue = position.getUnicode(); // Resets the average character width when we see a change in font // or a change in the font size if (lastPosition != null && (position.getFont() != lastPosition.getTextPosition().getFont() || position.getFontSize() != lastPosition.getTextPosition().getFontSize())) { previousAveCharWidth = -1; } float positionX; float positionY; float positionWidth; float positionHeight; // If we are sorting, then we need to use the text direction // adjusted coordinates, because they were used in the sorting. if (getSortByPosition()) { positionX = position.getXDirAdj(); positionY = position.getYDirAdj(); positionWidth = position.getWidthDirAdj(); positionHeight = position.getHeightDir(); } else { positionX = position.getX(); positionY = position.getY(); positionWidth = position.getWidth(); positionHeight = position.getHeight(); } // The current amount of characters in a word int wordCharCount = position.getIndividualWidths().length; // Estimate the expected width of the space based on the // space character with some margin. float wordSpacing = position.getWidthOfSpace(); float deltaSpace; if (wordSpacing == 0 || Float.isNaN(wordSpacing)) { deltaSpace = Float.MAX_VALUE; } else { if (lastWordSpacing < 0) { deltaSpace = wordSpacing * getSpacingTolerance(); } else { deltaSpace = (wordSpacing + lastWordSpacing) / 2f * getSpacingTolerance(); } } // Estimate the expected width of the space based on the average character width // with some margin. This calculation does not make a true average (average of // averages) but we found that it gave the best results after numerous experiments. // Based on experiments we also found that .3 worked well. float averageCharWidth; if (previousAveCharWidth < 0) { averageCharWidth = positionWidth / wordCharCount; } else { averageCharWidth = (previousAveCharWidth + positionWidth / wordCharCount) / 2f; } float deltaCharWidth = averageCharWidth * getAverageCharTolerance(); // Compares the values obtained by the average method and the wordSpacing method // and picks the smaller number. float expectedStartOfNextWordX = EXPECTED_START_OF_NEXT_WORD_X_RESET_VALUE; if (endOfLastTextX != END_OF_LAST_TEXT_X_RESET_VALUE) { if (deltaCharWidth > deltaSpace) { expectedStartOfNextWordX = endOfLastTextX + deltaSpace; } else { expectedStartOfNextWordX = endOfLastTextX + deltaCharWidth; } } if (lastPosition != null) { if (startOfArticle) { lastPosition.setArticleStart(); startOfArticle = false; } // RDD - Here we determine whether this text object is on the current // line. We use the lastBaselineFontSize to handle the superscript // case, and the size of the current font to handle the subscript case. // Text must overlap with the last rendered baseline text by at least // a small amount in order to be considered as being on the same line. // XXX BC: In theory, this check should really check if the next char is in // full range seen in this line. This is what I tried to do with minYTopForLine, // but this caused a lot of regression test failures. So, I'm leaving it be for // now if (!overlap(positionY, positionHeight, maxYForLine, maxHeightForLine)) { writeLine(normalize(line)); line.clear(); lastLineStartPosition = handleLineSeparation(current, lastPosition, lastLineStartPosition, maxHeightForLine); expectedStartOfNextWordX = EXPECTED_START_OF_NEXT_WORD_X_RESET_VALUE; maxYForLine = MAX_Y_FOR_LINE_RESET_VALUE; maxHeightForLine = MAX_HEIGHT_FOR_LINE_RESET_VALUE; minYTopForLine = MIN_Y_TOP_FOR_LINE_RESET_VALUE; } // test if our TextPosition starts after a new word would be expected to start if (expectedStartOfNextWordX != EXPECTED_START_OF_NEXT_WORD_X_RESET_VALUE && expectedStartOfNextWordX < positionX && // only bother adding a space if the last character was not a space lastPosition.getTextPosition().getUnicode() != null && !lastPosition.getTextPosition().getUnicode().endsWith(" ")) { line.add(LineItem.getWordSeparator()); } } if (positionY >= maxYForLine) { maxYForLine = positionY; } // RDD - endX is what PDF considers to be the x coordinate of the // end position of the text. We use it in computing our metrics below. endOfLastTextX = positionX + positionWidth; // add it to the list if (characterValue != null) { if (startOfPage && lastPosition == null) { writeParagraphStart();// not sure this is correct for RTL? } line.add(new LineItem(position)); } maxHeightForLine = Math.max(maxHeightForLine, positionHeight); minYTopForLine = Math.min(minYTopForLine, positionY - positionHeight); lastPosition = current; if (startOfPage) { lastPosition.setParagraphStart(); lastPosition.setLineStart(); lastLineStartPosition = lastPosition; startOfPage = false; } lastWordSpacing = wordSpacing; previousAveCharWidth = averageCharWidth; } // print the final line if (line.size() > 0) { writeLine(normalize(line)); writeParagraphEnd(); } endArticle(); } writePageEnd(); }
From source file:com.repeatability.pdf.PDFTextStripper.java
License:Apache License
/** * Write the string in TextPosition to the output stream. * * @param text The text to write to the stream. * @throws IOException If there is an error when writing the text. *//* w ww .j av a2s . co m*/ protected void writeCharacters(TextPosition text) throws IOException { output.write(text.getUnicode()); }
From source file:com.repeatability.pdf.PDFTextStripper.java
License:Apache License
/** * This will process a TextPosition object and add the text to the list of characters on a page. It takes care of * overlapping text.// w ww . j a v a 2 s .c o m * * @param text The text to process. */ @Override protected void processTextPosition(TextPosition text) { boolean showCharacter = true; if (suppressDuplicateOverlappingText) { showCharacter = false; String textCharacter = text.getUnicode(); float textX = text.getX(); float textY = text.getY(); TreeMap<Float, TreeSet<Float>> sameTextCharacters = characterListMapping.get(textCharacter); if (sameTextCharacters == null) { sameTextCharacters = new TreeMap<Float, TreeSet<Float>>(); characterListMapping.put(textCharacter, sameTextCharacters); } // RDD - Here we compute the value that represents the end of the rendered // text. This value is used to determine whether subsequent text rendered // on the same line overwrites the current text. // // We subtract any positive padding to handle cases where extreme amounts // of padding are applied, then backed off (not sure why this is done, but there // are cases where the padding is on the order of 10x the character width, and // the TJ just backs up to compensate after each character). Also, we subtract // an amount to allow for kerning (a percentage of the width of the last // character). boolean suppressCharacter = false; float tolerance = text.getWidth() / textCharacter.length() / 3.0f; SortedMap<Float, TreeSet<Float>> xMatches = sameTextCharacters.subMap(textX - tolerance, textX + tolerance); for (TreeSet<Float> xMatch : xMatches.values()) { SortedSet<Float> yMatches = xMatch.subSet(textY - tolerance, textY + tolerance); if (!yMatches.isEmpty()) { suppressCharacter = true; break; } } if (!suppressCharacter) { TreeSet<Float> ySet = sameTextCharacters.get(textX); if (ySet == null) { ySet = new TreeSet<Float>(); sameTextCharacters.put(textX, ySet); } ySet.add(textY); showCharacter = true; } } if (showCharacter) { // if we are showing the character then we need to determine which article it belongs to int foundArticleDivisionIndex = -1; int notFoundButFirstLeftAndAboveArticleDivisionIndex = -1; int notFoundButFirstLeftArticleDivisionIndex = -1; int notFoundButFirstAboveArticleDivisionIndex = -1; float x = text.getX(); float y = text.getY(); if (shouldSeparateByBeads) { for (int i = 0; i < beadRectangles.size() && foundArticleDivisionIndex == -1; i++) { PDRectangle rect = beadRectangles.get(i); if (rect != null) { if (rect.contains(x, y)) { foundArticleDivisionIndex = i * 2 + 1; } else if ((x < rect.getLowerLeftX() || y < rect.getUpperRightY()) && notFoundButFirstLeftAndAboveArticleDivisionIndex == -1) { notFoundButFirstLeftAndAboveArticleDivisionIndex = i * 2; } else if (x < rect.getLowerLeftX() && notFoundButFirstLeftArticleDivisionIndex == -1) { notFoundButFirstLeftArticleDivisionIndex = i * 2; } else if (y < rect.getUpperRightY() && notFoundButFirstAboveArticleDivisionIndex == -1) { notFoundButFirstAboveArticleDivisionIndex = i * 2; } } else { foundArticleDivisionIndex = 0; } } } else { foundArticleDivisionIndex = 0; } int articleDivisionIndex; if (foundArticleDivisionIndex != -1) { articleDivisionIndex = foundArticleDivisionIndex; } else if (notFoundButFirstLeftAndAboveArticleDivisionIndex != -1) { articleDivisionIndex = notFoundButFirstLeftAndAboveArticleDivisionIndex; } else if (notFoundButFirstLeftArticleDivisionIndex != -1) { articleDivisionIndex = notFoundButFirstLeftArticleDivisionIndex; } else if (notFoundButFirstAboveArticleDivisionIndex != -1) { articleDivisionIndex = notFoundButFirstAboveArticleDivisionIndex; } else { articleDivisionIndex = charactersByArticle.size() - 1; } List<TextPosition> textList = charactersByArticle.get(articleDivisionIndex); // In the wild, some PDF encoded documents put diacritics (accents on // top of characters) into a separate Tj element. When displaying them // graphically, the two chunks get overlayed. With text output though, // we need to do the overlay. This code recombines the diacritic with // its associated character if the two are consecutive. if (textList.isEmpty()) { textList.add(text); } else { // test if we overlap the previous entry. // Note that we are making an assumption that we need to only look back // one TextPosition to find what we are overlapping. // This may not always be true. */ TextPosition previousTextPosition = textList.get(textList.size() - 1); if (text.isDiacritic() && previousTextPosition.contains(text)) { previousTextPosition.mergeDiacritic(text); } // If the previous TextPosition was the diacritic, merge it into this // one and remove it from the list. else if (previousTextPosition.isDiacritic() && text.contains(previousTextPosition)) { text.mergeDiacritic(previousTextPosition); textList.remove(textList.size() - 1); textList.add(text); } else { textList.add(text); } } } }
From source file:com.repeatability.pdf.PDFTextStripper.java
License:Apache License
/** * returns the list item Pattern object that matches the text at the specified PositionWrapper or null if the text * does not match such a pattern. The list of Patterns tested against is given by the {@link #getListItemPatterns()} * method. To add to the list, simply override that method (if sub-classing) or explicitly supply your own list * using {@link #setListItemPatterns(List)}. * /*from w w w .ja v a 2 s . c om*/ * @param pw position * @return the matching pattern */ private Pattern matchListItemPattern(PositionWrapper pw) { TextPosition tp = pw.getTextPosition(); String txt = tp.getUnicode(); return matchPattern(txt, getListItemPatterns()); }
From source file:com.repeatability.pdf.PDFTextStripper.java
License:Apache License
/** * Used within {@link #normalize(List, boolean, boolean)} to handle a {@link TextPosition}. * /*from w w w . j a v a2 s. c om*/ * @return The StringBuilder that must be used when calling this method. */ // kwa // private StringBuilder normalizeAdd(List<WordWithTextPositions> normalized, // StringBuilder lineBuilder, List<TextPosition> wordPositions, LineItem item) protected StringBuilder normalizeAdd(List<WordWithTextPositions> normalized, StringBuilder lineBuilder, List<TextPosition> wordPositions, LineItem item) { if (item.isWordSeparator()) { normalized.add(createWord(lineBuilder.toString(), new ArrayList<TextPosition>(wordPositions))); lineBuilder = new StringBuilder(); wordPositions.clear(); } else { TextPosition text = item.getTextPosition(); if (text.getUnicode().length() != 1) { ; // System.out.println("Ha!"); } lineBuilder.append(text.getUnicode()); wordPositions.add(text); } return lineBuilder; }
From source file:com.tekstosense.segmenter.data.Text.java
License:Open Source License
public static Text newFor(TextPosition tp, PDGraphicsState gs, String text) { Text t = new Text(); t.x = tp.getXDirAdj();//w ww.j a va 2s . com t.baseline = tp.getYDirAdj(); t.font = tp.getFont(); t.strokeColor = gs.getStrokingColor(); t.nonStrokeColor = gs.getNonStrokingColor(); t.run = tp.getUnicode(); t.width = tp.getWidth(); t.height = tp.getHeight(); t.pointSize = tp.getFontSizeInPt(); t.fontSize = tp.getYScale(); t.tempRun = t.run; // Bump the width by the word spacing for each space in tp. /* for (int i=0; i<tp.getCharacter().length(); i++) { Character c = tp.getCharacter().charAt(i); if (c.equals(" ")) { t.width -= tp.getWidthOfSpace(); t.width += tp.getWordSpacing(); } } */ return t; }
From source file:com.yiyihealth.tools.test.DrawPrintTextLocations.java
License:Apache License
/** * Override the default functionality of PDFTextStripper. *//*from ww w. j a v a 2s. c o m*/ @Override protected void writeString(String string, List<TextPosition> textPositions) throws IOException { for (TextPosition text : textPositions) { System.out.println("String[" + text.getXDirAdj() + "," + text.getYDirAdj() + " fs=" + text.getFontSize() + " xscale=" + text.getXScale() + " height=" + text.getHeightDir() + " space=" + text.getWidthOfSpace() + " width=" + text.getWidthDirAdj() + "]" + text.getUnicode()); // in red: // show rectangles with the "height" (not a real height, but used for text extraction // heuristics, it is 1/2 of the bounding box height and starts at y=0) Rectangle2D.Float rect = new Rectangle2D.Float(text.getXDirAdj(), (text.getYDirAdj() - text.getHeightDir()), text.getWidthDirAdj(), text.getHeightDir()); g2d.setColor(Color.red); g2d.draw(rect); // in blue: // show rectangle with the real vertical bounds, based on the font bounding box y values // usually, the height is identical to what you see when marking text in Adobe Reader PDFont font = text.getFont(); BoundingBox bbox = font.getBoundingBox(); // advance width, bbox height (glyph space) float xadvance = font.getWidth(text.getCharacterCodes()[0]); // todo: should iterate all chars rect = new Rectangle2D.Float(0, bbox.getLowerLeftY(), xadvance, bbox.getHeight()); // glyph space -> user space // note: text.getTextMatrix() is *not* the Text Matrix, it's the Text Rendering Matrix AffineTransform at = text.getTextMatrix().createAffineTransform(); if (font instanceof PDType3Font) { // bbox and font matrix are unscaled at.concatenate(font.getFontMatrix().createAffineTransform()); } else { // bbox and font matrix are already scaled to 1000 at.scale(1 / 1000f, 1 / 1000f); } Shape s = at.createTransformedShape(rect); s = flipAT.createTransformedShape(s); s = rotateAT.createTransformedShape(s); g2d.setColor(Color.blue); g2d.draw(s); } }
From source file:edu.ist.psu.sagnik.research.pdfbox2playground.javatest.DrawPrintTextLocations.java
License:Apache License
/** * Override the default functionality of PDFTextStripper. *//*w ww . j a v a2 s . co m*/ @Override protected void writeString(String string, List<TextPosition> textPositions) throws IOException { for (TextPosition text : textPositions) { System.out.println("String[" + text.getXDirAdj() + "," + text.getYDirAdj() + " fs=" + text.getFontSize() + " xscale=" + text.getXScale() + " height=" + text.getHeightDir() + " space=" + text.getWidthOfSpace() + " width=" + text.getWidthDirAdj() + "]" + text.getUnicode()); // in red: // show rectangles with the "height" (not a real height, but used for text extraction // heuristics, it is 1/2 of the bounding box height and starts at y=0) Rectangle2D.Float rect = new Rectangle2D.Float(text.getXDirAdj(), (text.getYDirAdj() - text.getHeightDir()), text.getWidthDirAdj(), text.getHeightDir()); g2d.setColor(Color.red); g2d.draw(rect); // in blue: // show rectangle with the real vertical bounds, based on the font bounding box y values // usually, the height is identical to what you see when marking text in Adobe Reader PDFont font = text.getFont(); BoundingBox bbox = font.getBoundingBox(); // advance width, bbox height (glyph space) float xadvance = font.getWidth(text.getCharacterCodes()[0]); // todo: should iterate all chars rect = new Rectangle2D.Float(0, bbox.getLowerLeftY(), xadvance, bbox.getHeight()); // glyph space -> user space // note: text.getTextMatrix() is *not* the Text Matrix, it's the Text Rendering Matrix AffineTransform at = text.getTextMatrix().createAffineTransform(); if (font instanceof PDType3Font) { // bbox and font matrix are unscaled at.concatenate(font.getFontMatrix().createAffineTransform()); } else { // bbox and font matrix are already scaled to 1000 at.scale(1 / 1000f, 1 / 1000f); } Shape s = at.createTransformedShape(rect); s = flipAT.createTransformedShape(s); s = rotateAT.createTransformedShape(s); g2d.setColor(Color.blue); g2d.draw(s); } }