Re: PrintTextLocations 1.8 vs 2.0
Sorry for the late response. I tested the code with severel documents and it works well :-) Thank you very much. Am 17.03.2016 um 18:07 schrieb Tilman Hausherr: I found a case where the strategy you mentioned didn't work for a TT font, the file 032431. Here's some updated code. private Shape calculateGlyphBounds(Matrix textRenderingMatrix, PDFont font, int code) throws IOException { GeneralPath path = null; AffineTransform at = textRenderingMatrix.createAffineTransform(); at.concatenate(font.getFontMatrix().createAffineTransform()); if (font instanceof PDType3Font) { PDType3Font t3Font = (PDType3Font) font; PDType3CharProc charProc = t3Font.getCharProc(code); if (charProc != null) { PDRectangle glyphBBox = charProc.getGlyphBBox(); if (glyphBBox != null) { path = glyphBBox.toGeneralPath(); } } } else if (font instanceof PDVectorFont) { PDVectorFont vectorFont = (PDVectorFont) font; path = vectorFont.getPath(code); if (font instanceof PDTrueTypeFont) { PDTrueTypeFont ttFont = (PDTrueTypeFont) font; int unitsPerEm = ttFont.getTrueTypeFont().getHeader().getUnitsPerEm(); at.scale(1000d / unitsPerEm, 1000d / unitsPerEm); } if (font instanceof PDType0Font) { PDType0Font t0font = (PDType0Font) font; if (t0font.getDescendantFont() instanceof PDCIDFontType2) { int unitsPerEm = ((PDCIDFontType2) t0font.getDescendantFont()).getTrueTypeFont().getHeader().getUnitsPerEm(); at.scale(1000d / unitsPerEm, 1000d / unitsPerEm); } } } else if (font instanceof PDSimpleFont) { PDSimpleFont simpleFont = (PDSimpleFont) font; // these two lines do not always work, e.g. for the TT fonts in file 032431.pdf // which is why PDVectorFont is tried first. String name = simpleFont.getEncoding().getName(code); path = simpleFont.getPath(name); } else { // shouldn't happen, please open issue in JIRA System.out.println("Unknown font class: " + font.getClass()); } if (path == null) { return null; } return at.createTransformedShape(path.getBounds2D()); } - To unsubscribe, e-mail: users-unsubscr...@pdfbox.apache.org For additional commands, e-mail: users-h...@pdfbox.apache.org - To unsubscribe, e-mail: users-unsubscr...@pdfbox.apache.org For additional commands, e-mail: users-h...@pdfbox.apache.org
Re: PrintTextLocations 1.8 vs 2.0
I found a case where the strategy you mentioned didn't work for a TT font, the file 032431. Here's some updated code. private Shape calculateGlyphBounds(Matrix textRenderingMatrix, PDFont font, int code) throws IOException { GeneralPath path = null; AffineTransform at = textRenderingMatrix.createAffineTransform(); at.concatenate(font.getFontMatrix().createAffineTransform()); if (font instanceof PDType3Font) { PDType3Font t3Font = (PDType3Font) font; PDType3CharProc charProc = t3Font.getCharProc(code); if (charProc != null) { PDRectangle glyphBBox = charProc.getGlyphBBox(); if (glyphBBox != null) { path = glyphBBox.toGeneralPath(); } } } else if (font instanceof PDVectorFont) { PDVectorFont vectorFont = (PDVectorFont) font; path = vectorFont.getPath(code); if (font instanceof PDTrueTypeFont) { PDTrueTypeFont ttFont = (PDTrueTypeFont) font; int unitsPerEm = ttFont.getTrueTypeFont().getHeader().getUnitsPerEm(); at.scale(1000d / unitsPerEm, 1000d / unitsPerEm); } if (font instanceof PDType0Font) { PDType0Font t0font = (PDType0Font) font; if (t0font.getDescendantFont() instanceof PDCIDFontType2) { int unitsPerEm = ((PDCIDFontType2) t0font.getDescendantFont()).getTrueTypeFont().getHeader().getUnitsPerEm(); at.scale(1000d / unitsPerEm, 1000d / unitsPerEm); } } } else if (font instanceof PDSimpleFont) { PDSimpleFont simpleFont = (PDSimpleFont) font; // these two lines do not always work, e.g. for the TT fonts in file 032431.pdf // which is why PDVectorFont is tried first. String name = simpleFont.getEncoding().getName(code); path = simpleFont.getPath(name); } else { // shouldn't happen, please open issue in JIRA System.out.println("Unknown font class: " + font.getClass()); } if (path == null) { return null; } return at.createTransformedShape(path.getBounds2D()); } - To unsubscribe, e-mail: users-unsubscr...@pdfbox.apache.org For additional commands, e-mail: users-h...@pdfbox.apache.org
Re: PrintTextLocations 1.8 vs 2.0
Am 16.03.2016 um 09:52 schrieb Peter Prusinowski: thank you for the hints, now I am overwriting showGlyph() and trying to get the value with PDSimpleFont sf = (PDSimpleFont) font; String name = sf.getEncoding().getName(code); sf.getPath(name).getBounds() but I am getting the same height, no matter which font size is set. This happens with type1 and truetype fonts. What am I doing wrong ? Here's some code, use it with the DrawPrintTextLocations example. Please tell if it works, and if possible, upload files where it doesn't. @Override protected void showGlyph(Matrix textRenderingMatrix, PDFont font, int code, String unicode, Vector displacement) throws IOException { super.showGlyph(textRenderingMatrix, font, code, unicode, displacement); Rectangle2D bounds = null; AffineTransform at = textRenderingMatrix.createAffineTransform(); if (font instanceof PDSimpleFont) { PDSimpleFont simpleFont = (PDSimpleFont) font; String name = simpleFont.getEncoding().getName(code); GeneralPath path = simpleFont.getPath(name); bounds = path.getBounds2D(); at.scale(1/1000f, 1/1000f); if (font instanceof PDTrueTypeFont) { PDTrueTypeFont ttFont = (PDTrueTypeFont) font; int unitsPerEm = ttFont.getTrueTypeFont().getHeader().getUnitsPerEm(); at.scale(1000d / unitsPerEm, 1000d / unitsPerEm); } } else if (font instanceof PDVectorFont) { PDVectorFont vectorFont = (PDVectorFont) font; GeneralPath path = vectorFont.getPath(code); bounds = path.getBounds2D(); at.scale(1/1000f, 1/1000f); if (font instanceof PDType0Font) { PDType0Font t0font = (PDType0Font) font; if (t0font.getDescendantFont() instanceof PDCIDFontType2) { int unitsPerEm = ((PDCIDFontType2) t0font.getDescendantFont()).getTrueTypeFont().getHeader().getUnitsPerEm(); at.scale(1000d / unitsPerEm, 1000d / unitsPerEm); } } } else { System.out.println("TODO other: " + font.getClass()); } if (bounds != null) { Shape s = at.createTransformedShape(bounds); // flip y-axis AffineTransform flip = new AffineTransform(); flip.translate(0, getCurrentPage().getBBox().getHeight()); flip.scale(1, -1); s = flip.createTransformedShape(s); AffineTransform transform = g2d.getTransform(); int rotation = getCurrentPage().getRotation(); if (rotation != 0) { PDRectangle mediaBox = getCurrentPage().getMediaBox(); switch (rotation) { case 90: g2d.translate(mediaBox.getHeight(), 0); break; case 270: g2d.translate(0, mediaBox.getWidth()); break; case 180: g2d.translate(mediaBox.getWidth(), mediaBox.getHeight()); break; default: break; } g2d.rotate(Math.toRadians(rotation)); } g2d.setColor(Color.CYAN); g2d.draw(s); if (rotation != 0) { g2d.setTransform(transform); } } } - To unsubscribe, e-mail: users-unsubscr...@pdfbox.apache.org For additional commands, e-mail: users-h...@pdfbox.apache.org
Re: PrintTextLocations 1.8 vs 2.0
Hi, > Peter Prusinowski hat am 16. März 2016 um 09:52 > geschrieben: > > > Good morning, > > thank you for the hints, now I am overwriting showGlyph() and trying to > get the value with > > PDSimpleFont sf = (PDSimpleFont) font; > String name = sf.getEncoding().getName(code); > sf.getPath(name).getBounds() > > but I am getting the same height, no matter which font size is set. This > happens with type1 and truetype fonts. What am I doing wrong ? The font provides always the same unscaled shapes. You have to take the text transformation matrix and the font matrix into account. Have a look at PageDrawer#showFontGlyph to see how to do so. HTH Andreas > > Am 07.03.2016 um 18:16 schrieb Tilman Hausherr: > > Am 07.03.2016 um 11:46 schrieb Peter Prusinowski: > >> Okay, thank you for information. I tried to get the height with > >> getPath(). If its one of the 14 standard fonts, I can get the height > >> with PDType1Font.fontName.getPath(text.getUnicode()).getBounds()). > >> But I dont know how to get the information from other fonts in a > >> generic way. Do you have a hint for me ? > > > > It is not available for all fonts. It is available for all > > PDSimpleFont objects, except for PDType3Font (which doesn't draw just > > vectors). > > > > The best would be to look at the source code, at PageDrawer.java > > > > createGlyph2D() returns a Glyph2D for the font. That one you can use > > for glyph2D.getPathForCharacterCode(code); > > > > See also showFontGlyph(), you can override that one in a subclass. > > > > Have also a look at showGlyph(), this makes a difference between type3 > > fonts and others. See also CustomGraphicsStreamEngine. > > > > Tilman > > > > > > > >> > >> Peter > >> > >> Am 06.03.2016 um 17:40 schrieb Tilman Hausherr: > >>> > >>> In 1.8, for Standard 14 fonts (yours is) it uses the bounding box of > >>> each glyph. In a string, it uses a maximum which it keeps for the > >>> string, that results in the weird effect that the "d" is slightly > >>> higher. If the string is changed so that another glyph is appended, > >>> the larger height is kept. > >>> > >>> In 2.0 (and in 1.8 for non standard 14 fonts), it uses 1/2 of the > >>> bounding box from the font descriptor. The not-halved bounding box > >>> is usually too high. > >>> > >>> Anyway, the 1.8 logic would work for you for standard 14 fonts, but > >>> not for all other fonts. > >>> > >>> So there is no bug in 1.8 not in 2.0. > >>> > >>> Tilman > >>> > >>> Am 03.03.2016 um 19:05 schrieb Tilman Hausherr: > Am 03.03.2016 um 09:11 schrieb Peter Prusinowski: > > Okay, I am trying to replace some words in documents and use > > text.height to "delete" these words. Here is an example document : > > http://workupload.com/file/G8ipDe8j > > The getHeightDir() is not the best strategy, for the reason I > mentioned yesterday. In your case, you should call getPath() on the > glyphs and get the bounding box. Or just get the font bounding box > (there's a method) height, however that one is often too high, so > there's a risk that you blank the line above. > > But thanks for the file, I'll try to find out why it is different. > The heights in 1.8 are surprising, usually they are never so > "perfect" (as I said yesterday). And for some reason, in 1.8 the > height of the last glyph is slightly different although it is all > in one string. > > 1.8: > String[100.0,92.0 fs=14.0 xscale=14.0 height=10.052001 > space=3.8920004 width=10.108002]H > String[110.108,92.0 fs=14.0 xscale=14.0 height=10.052001 > space=3.8920004 width=7.784004]e > String[117.892006,92.0 fs=14.0 xscale=14.0 height=10.052001 > space=3.8920004 width=3.8919983]l > String[121.784004,92.0 fs=14.0 xscale=14.0 height=10.052001 > space=3.8920004 width=3.8919983]l > String[125.676,92.0 fs=14.0 xscale=14.0 height=10.052001 > space=3.8920004 width=8.553993]o > String[134.23,92.0 fs=14.0 xscale=14.0 height=10.052001 > space=3.8920004 width=3.8919983] > String[138.122,92.0 fs=14.0 xscale=14.0 height=10.052001 > space=3.8920004 width=13.216003]W > String[151.338,92.0 fs=14.0 xscale=14.0 height=10.052001 > space=3.8920004 width=8.554001]o > String[159.892,92.0 fs=14.0 xscale=14.0 height=10.052001 > space=3.8920004 width=5.445999]r > String[165.338,92.0 fs=14.0 xscale=14.0 height=10.052001 > space=3.8920004 width=3.8919983]l > String[169.23,92.0 fs=14.0 xscale=14.0 *height=10.248001* > space=3.8920004 width=8.554001]d <= ??? > > 2.0: > String[100.0,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 > width=10.108002]H > String[110.108,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 > width=7.7839966]e > String[117.892,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 > wi
Re: PrintTextLocations 1.8 vs 2.0
Good morning, thank you for the hints, now I am overwriting showGlyph() and trying to get the value with PDSimpleFont sf = (PDSimpleFont) font; String name = sf.getEncoding().getName(code); sf.getPath(name).getBounds() but I am getting the same height, no matter which font size is set. This happens with type1 and truetype fonts. What am I doing wrong ? Am 07.03.2016 um 18:16 schrieb Tilman Hausherr: Am 07.03.2016 um 11:46 schrieb Peter Prusinowski: Okay, thank you for information. I tried to get the height with getPath(). If its one of the 14 standard fonts, I can get the height with PDType1Font.fontName.getPath(text.getUnicode()).getBounds()). But I dont know how to get the information from other fonts in a generic way. Do you have a hint for me ? It is not available for all fonts. It is available for all PDSimpleFont objects, except for PDType3Font (which doesn't draw just vectors). The best would be to look at the source code, at PageDrawer.java createGlyph2D() returns a Glyph2D for the font. That one you can use for glyph2D.getPathForCharacterCode(code); See also showFontGlyph(), you can override that one in a subclass. Have also a look at showGlyph(), this makes a difference between type3 fonts and others. See also CustomGraphicsStreamEngine. Tilman Peter Am 06.03.2016 um 17:40 schrieb Tilman Hausherr: In 1.8, for Standard 14 fonts (yours is) it uses the bounding box of each glyph. In a string, it uses a maximum which it keeps for the string, that results in the weird effect that the "d" is slightly higher. If the string is changed so that another glyph is appended, the larger height is kept. In 2.0 (and in 1.8 for non standard 14 fonts), it uses 1/2 of the bounding box from the font descriptor. The not-halved bounding box is usually too high. Anyway, the 1.8 logic would work for you for standard 14 fonts, but not for all other fonts. So there is no bug in 1.8 not in 2.0. Tilman Am 03.03.2016 um 19:05 schrieb Tilman Hausherr: Am 03.03.2016 um 09:11 schrieb Peter Prusinowski: Okay, I am trying to replace some words in documents and use text.height to "delete" these words. Here is an example document : http://workupload.com/file/G8ipDe8j The getHeightDir() is not the best strategy, for the reason I mentioned yesterday. In your case, you should call getPath() on the glyphs and get the bounding box. Or just get the font bounding box (there's a method) height, however that one is often too high, so there's a risk that you blank the line above. But thanks for the file, I'll try to find out why it is different. The heights in 1.8 are surprising, usually they are never so "perfect" (as I said yesterday). And for some reason, in 1.8 the height of the last glyph is slightly different although it is all in one string. 1.8: String[100.0,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=10.108002]H String[110.108,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=7.784004]e String[117.892006,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=3.8919983]l String[121.784004,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=3.8919983]l String[125.676,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=8.553993]o String[134.23,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=3.8919983] String[138.122,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=13.216003]W String[151.338,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=8.554001]o String[159.892,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=5.445999]r String[165.338,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=3.8919983]l String[169.23,92.0 fs=14.0 xscale=14.0 *height=10.248001* space=3.8920004 width=8.554001]d <= ??? 2.0: String[100.0,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=10.108002]H String[110.108,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=7.7839966]e String[117.892,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=3.8919983]l String[121.784,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=3.8919983]l String[125.675995,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=8.554001]o String[134.23,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=3.8919983] String[138.122,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=13.216003]W String[151.338,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=8.554001]o String[159.892,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=5.445999]r String[165.338,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=3.8919983]l String[169.23,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=8.554001]d Tilman Peter Am 02.03.2016 um 19:24 schrieb Tilman Hausherr: Am 02.03.2016 um 14:48 schrieb Peter Prusinowski: Hello, I have noticed
Re: PrintTextLocations 1.8 vs 2.0
Am 07.03.2016 um 11:46 schrieb Peter Prusinowski: Okay, thank you for information. I tried to get the height with getPath(). If its one of the 14 standard fonts, I can get the height with PDType1Font.fontName.getPath(text.getUnicode()).getBounds()). But I dont know how to get the information from other fonts in a generic way. Do you have a hint for me ? It is not available for all fonts. It is available for all PDSimpleFont objects, except for PDType3Font (which doesn't draw just vectors). The best would be to look at the source code, at PageDrawer.java createGlyph2D() returns a Glyph2D for the font. That one you can use for glyph2D.getPathForCharacterCode(code); See also showFontGlyph(), you can override that one in a subclass. Have also a look at showGlyph(), this makes a difference between type3 fonts and others. See also CustomGraphicsStreamEngine. Tilman Peter Am 06.03.2016 um 17:40 schrieb Tilman Hausherr: In 1.8, for Standard 14 fonts (yours is) it uses the bounding box of each glyph. In a string, it uses a maximum which it keeps for the string, that results in the weird effect that the "d" is slightly higher. If the string is changed so that another glyph is appended, the larger height is kept. In 2.0 (and in 1.8 for non standard 14 fonts), it uses 1/2 of the bounding box from the font descriptor. The not-halved bounding box is usually too high. Anyway, the 1.8 logic would work for you for standard 14 fonts, but not for all other fonts. So there is no bug in 1.8 not in 2.0. Tilman Am 03.03.2016 um 19:05 schrieb Tilman Hausherr: Am 03.03.2016 um 09:11 schrieb Peter Prusinowski: Okay, I am trying to replace some words in documents and use text.height to "delete" these words. Here is an example document : http://workupload.com/file/G8ipDe8j The getHeightDir() is not the best strategy, for the reason I mentioned yesterday. In your case, you should call getPath() on the glyphs and get the bounding box. Or just get the font bounding box (there's a method) height, however that one is often too high, so there's a risk that you blank the line above. But thanks for the file, I'll try to find out why it is different. The heights in 1.8 are surprising, usually they are never so "perfect" (as I said yesterday). And for some reason, in 1.8 the height of the last glyph is slightly different although it is all in one string. 1.8: String[100.0,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=10.108002]H String[110.108,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=7.784004]e String[117.892006,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=3.8919983]l String[121.784004,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=3.8919983]l String[125.676,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=8.553993]o String[134.23,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=3.8919983] String[138.122,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=13.216003]W String[151.338,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=8.554001]o String[159.892,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=5.445999]r String[165.338,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=3.8919983]l String[169.23,92.0 fs=14.0 xscale=14.0 *height=10.248001* space=3.8920004 width=8.554001]d <= ??? 2.0: String[100.0,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=10.108002]H String[110.108,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=7.7839966]e String[117.892,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=3.8919983]l String[121.784,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=3.8919983]l String[125.675995,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=8.554001]o String[134.23,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=3.8919983] String[138.122,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=13.216003]W String[151.338,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=8.554001]o String[159.892,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=5.445999]r String[165.338,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=3.8919983]l String[169.23,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=8.554001]d Tilman Peter Am 02.03.2016 um 19:24 schrieb Tilman Hausherr: Am 02.03.2016 um 14:48 schrieb Peter Prusinowski: Hello, I have noticed that the PrintTextLocations example in 1.8 and 2.0 gives different results for text.getHeightDir(). In 1.8 the value seems to be right, but in 2.0 it is too small. I tried with some PDFBox created documents. Is this a bug ? Maybe, maybe not. The height is a heuristic value to help with text extraction, which is sometimes computed differently in 2.0, and it is usually about the height of an "a". Please upload the PDF. Tilman ---
Re: PrintTextLocations 1.8 vs 2.0
Okay, thank you for information. I tried to get the height with getPath(). If its one of the 14 standard fonts, I can get the height with PDType1Font.fontName.getPath(text.getUnicode()).getBounds()). But I dont know how to get the information from other fonts in a generic way. Do you have a hint for me ? Peter Am 06.03.2016 um 17:40 schrieb Tilman Hausherr: In 1.8, for Standard 14 fonts (yours is) it uses the bounding box of each glyph. In a string, it uses a maximum which it keeps for the string, that results in the weird effect that the "d" is slightly higher. If the string is changed so that another glyph is appended, the larger height is kept. In 2.0 (and in 1.8 for non standard 14 fonts), it uses 1/2 of the bounding box from the font descriptor. The not-halved bounding box is usually too high. Anyway, the 1.8 logic would work for you for standard 14 fonts, but not for all other fonts. So there is no bug in 1.8 not in 2.0. Tilman Am 03.03.2016 um 19:05 schrieb Tilman Hausherr: Am 03.03.2016 um 09:11 schrieb Peter Prusinowski: Okay, I am trying to replace some words in documents and use text.height to "delete" these words. Here is an example document : http://workupload.com/file/G8ipDe8j The getHeightDir() is not the best strategy, for the reason I mentioned yesterday. In your case, you should call getPath() on the glyphs and get the bounding box. Or just get the font bounding box (there's a method) height, however that one is often too high, so there's a risk that you blank the line above. But thanks for the file, I'll try to find out why it is different. The heights in 1.8 are surprising, usually they are never so "perfect" (as I said yesterday). And for some reason, in 1.8 the height of the last glyph is slightly different although it is all in one string. 1.8: String[100.0,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=10.108002]H String[110.108,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=7.784004]e String[117.892006,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=3.8919983]l String[121.784004,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=3.8919983]l String[125.676,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=8.553993]o String[134.23,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=3.8919983] String[138.122,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=13.216003]W String[151.338,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=8.554001]o String[159.892,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=5.445999]r String[165.338,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=3.8919983]l String[169.23,92.0 fs=14.0 xscale=14.0 *height=10.248001* space=3.8920004 width=8.554001]d <= ??? 2.0: String[100.0,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=10.108002]H String[110.108,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=7.7839966]e String[117.892,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=3.8919983]l String[121.784,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=3.8919983]l String[125.675995,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=8.554001]o String[134.23,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=3.8919983] String[138.122,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=13.216003]W String[151.338,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=8.554001]o String[159.892,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=5.445999]r String[165.338,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=3.8919983]l String[169.23,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=8.554001]d Tilman Peter Am 02.03.2016 um 19:24 schrieb Tilman Hausherr: Am 02.03.2016 um 14:48 schrieb Peter Prusinowski: Hello, I have noticed that the PrintTextLocations example in 1.8 and 2.0 gives different results for text.getHeightDir(). In 1.8 the value seems to be right, but in 2.0 it is too small. I tried with some PDFBox created documents. Is this a bug ? Maybe, maybe not. The height is a heuristic value to help with text extraction, which is sometimes computed differently in 2.0, and it is usually about the height of an "a". Please upload the PDF. Tilman - To unsubscribe, e-mail: users-unsubscr...@pdfbox.apache.org For additional commands, e-mail: users-h...@pdfbox.apache.org - To unsubscribe, e-mail: users-unsubscr...@pdfbox.apache.org For additional commands, e-mail: users-h...@pdfbox.apache.org - To unsubscribe, e-mail: users-unsubscr...@pdfbox.apache.org For additional commands, e-mail: users-h...@pdfbox.apache.org ---
Re: PrintTextLocations 1.8 vs 2.0
In 1.8, for Standard 14 fonts (yours is) it uses the bounding box of each glyph. In a string, it uses a maximum which it keeps for the string, that results in the weird effect that the "d" is slightly higher. If the string is changed so that another glyph is appended, the larger height is kept. In 2.0 (and in 1.8 for non standard 14 fonts), it uses 1/2 of the bounding box from the font descriptor. The not-halved bounding box is usually too high. Anyway, the 1.8 logic would work for you for standard 14 fonts, but not for all other fonts. So there is no bug in 1.8 not in 2.0. Tilman Am 03.03.2016 um 19:05 schrieb Tilman Hausherr: Am 03.03.2016 um 09:11 schrieb Peter Prusinowski: Okay, I am trying to replace some words in documents and use text.height to "delete" these words. Here is an example document : http://workupload.com/file/G8ipDe8j The getHeightDir() is not the best strategy, for the reason I mentioned yesterday. In your case, you should call getPath() on the glyphs and get the bounding box. Or just get the font bounding box (there's a method) height, however that one is often too high, so there's a risk that you blank the line above. But thanks for the file, I'll try to find out why it is different. The heights in 1.8 are surprising, usually they are never so "perfect" (as I said yesterday). And for some reason, in 1.8 the height of the last glyph is slightly different although it is all in one string. 1.8: String[100.0,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=10.108002]H String[110.108,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=7.784004]e String[117.892006,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=3.8919983]l String[121.784004,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=3.8919983]l String[125.676,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=8.553993]o String[134.23,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=3.8919983] String[138.122,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=13.216003]W String[151.338,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=8.554001]o String[159.892,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=5.445999]r String[165.338,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=3.8919983]l String[169.23,92.0 fs=14.0 xscale=14.0 *height=10.248001* space=3.8920004 width=8.554001]d <= ??? 2.0: String[100.0,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=10.108002]H String[110.108,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=7.7839966]e String[117.892,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=3.8919983]l String[121.784,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=3.8919983]l String[125.675995,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=8.554001]o String[134.23,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=3.8919983] String[138.122,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=13.216003]W String[151.338,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=8.554001]o String[159.892,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=5.445999]r String[165.338,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=3.8919983]l String[169.23,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=8.554001]d Tilman Peter Am 02.03.2016 um 19:24 schrieb Tilman Hausherr: Am 02.03.2016 um 14:48 schrieb Peter Prusinowski: Hello, I have noticed that the PrintTextLocations example in 1.8 and 2.0 gives different results for text.getHeightDir(). In 1.8 the value seems to be right, but in 2.0 it is too small. I tried with some PDFBox created documents. Is this a bug ? Maybe, maybe not. The height is a heuristic value to help with text extraction, which is sometimes computed differently in 2.0, and it is usually about the height of an "a". Please upload the PDF. Tilman - To unsubscribe, e-mail: users-unsubscr...@pdfbox.apache.org For additional commands, e-mail: users-h...@pdfbox.apache.org - To unsubscribe, e-mail: users-unsubscr...@pdfbox.apache.org For additional commands, e-mail: users-h...@pdfbox.apache.org - To unsubscribe, e-mail: users-unsubscr...@pdfbox.apache.org For additional commands, e-mail: users-h...@pdfbox.apache.org
Re: PrintTextLocations 1.8 vs 2.0
Am 03.03.2016 um 09:11 schrieb Peter Prusinowski: Okay, I am trying to replace some words in documents and use text.height to "delete" these words. Here is an example document : http://workupload.com/file/G8ipDe8j The getHeightDir() is not the best strategy, for the reason I mentioned yesterday. In your case, you should call getPath() on the glyphs and get the bounding box. Or just get the font bounding box (there's a method) height, however that one is often too high, so there's a risk that you blank the line above. But thanks for the file, I'll try to find out why it is different. The heights in 1.8 are surprising, usually they are never so "perfect" (as I said yesterday). And for some reason, in 1.8 the height of the last glyph is slightly different although it is all in one string. 1.8: String[100.0,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=10.108002]H String[110.108,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=7.784004]e String[117.892006,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=3.8919983]l String[121.784004,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=3.8919983]l String[125.676,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=8.553993]o String[134.23,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=3.8919983] String[138.122,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=13.216003]W String[151.338,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=8.554001]o String[159.892,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=5.445999]r String[165.338,92.0 fs=14.0 xscale=14.0 height=10.052001 space=3.8920004 width=3.8919983]l String[169.23,92.0 fs=14.0 xscale=14.0 *height=10.248001* space=3.8920004 width=8.554001]d <= ??? 2.0: String[100.0,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=10.108002]H String[110.108,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=7.7839966]e String[117.892,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=3.8919983]l String[121.784,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=3.8919983]l String[125.675995,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=8.554001]o String[134.23,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=3.8919983] String[138.122,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=13.216003]W String[151.338,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=8.554001]o String[159.892,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=5.445999]r String[165.338,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=3.8919983]l String[169.23,92.0 fs=14.0 xscale=14.0 height=8.33 space=3.8920004 width=8.554001]d Tilman Peter Am 02.03.2016 um 19:24 schrieb Tilman Hausherr: Am 02.03.2016 um 14:48 schrieb Peter Prusinowski: Hello, I have noticed that the PrintTextLocations example in 1.8 and 2.0 gives different results for text.getHeightDir(). In 1.8 the value seems to be right, but in 2.0 it is too small. I tried with some PDFBox created documents. Is this a bug ? Maybe, maybe not. The height is a heuristic value to help with text extraction, which is sometimes computed differently in 2.0, and it is usually about the height of an "a". Please upload the PDF. Tilman - To unsubscribe, e-mail: users-unsubscr...@pdfbox.apache.org For additional commands, e-mail: users-h...@pdfbox.apache.org - To unsubscribe, e-mail: users-unsubscr...@pdfbox.apache.org For additional commands, e-mail: users-h...@pdfbox.apache.org
Re: PrintTextLocations 1.8 vs 2.0
Okay, I am trying to replace some words in documents and use text.height to "delete" these words. Here is an example document : http://workupload.com/file/G8ipDe8j Peter Am 02.03.2016 um 19:24 schrieb Tilman Hausherr: Am 02.03.2016 um 14:48 schrieb Peter Prusinowski: Hello, I have noticed that the PrintTextLocations example in 1.8 and 2.0 gives different results for text.getHeightDir(). In 1.8 the value seems to be right, but in 2.0 it is too small. I tried with some PDFBox created documents. Is this a bug ? Maybe, maybe not. The height is a heuristic value to help with text extraction, which is sometimes computed differently in 2.0, and it is usually about the height of an "a". Please upload the PDF. Tilman - To unsubscribe, e-mail: users-unsubscr...@pdfbox.apache.org For additional commands, e-mail: users-h...@pdfbox.apache.org - To unsubscribe, e-mail: users-unsubscr...@pdfbox.apache.org For additional commands, e-mail: users-h...@pdfbox.apache.org
Re: PrintTextLocations 1.8 vs 2.0
Am 02.03.2016 um 14:48 schrieb Peter Prusinowski: Hello, I have noticed that the PrintTextLocations example in 1.8 and 2.0 gives different results for text.getHeightDir(). In 1.8 the value seems to be right, but in 2.0 it is too small. I tried with some PDFBox created documents. Is this a bug ? Maybe, maybe not. The height is a heuristic value to help with text extraction, which is sometimes computed differently in 2.0, and it is usually about the height of an "a". Please upload the PDF. Tilman - To unsubscribe, e-mail: users-unsubscr...@pdfbox.apache.org For additional commands, e-mail: users-h...@pdfbox.apache.org