from reportlab.lib.testutils import setOutDir,makeSuiteForClasses, outputfile, printLocation, NearTestCase, rlSkipIf setOutDir(__name__) import unittest, sys from reportlab.pdfgen.canvas import Canvas from reportlab.pdfbase import pdfmetrics from reportlab.pdfbase.ttfonts import TTFont from reportlab.pdfbase import pdfutils from reportlab.platypus.paragraph import Paragraph from reportlab.lib.styles import ParagraphStyle from reportlab.lib.rl_accel import escapePDF from reportlab.lib.utils import isUnicode from reportlab.graphics.shapes import Drawing, String, Ellipse import re import codecs textPat = re.compile(r'\([^(]*\)') #test sentences testCp1252 = b'copyright \xa9 trademark \x99 registered \xae ReportLab! Ol\xe9!' testUni = testCp1252.decode('cp1252') testUTF8 = testUni.encode('utf-8') # expected result is octal-escaped text in the PDF expectedCp1252 = escapePDF(testCp1252) def extractText(pdfOps): """Utility to rip out the PDF text within a block of PDF operators. PDF will show a string draw as something like "(Hello World) Tj" i.e. text is in curved brackets. Crude and dirty, probably fails on escaped brackets. """ found = textPat.findall(pdfOps) #chop off '(' and ')' return [x[1:-1] for x in found] def subsetToUnicode(ttf, subsetCodeStr): """Return unicode string represented by given subsetCode string as found when TrueType font rendered to PDF, ttf must be the font object that was used.""" # This relies on TTFont internals and uses the first document # and subset it finds subset = list(ttf.state.values())[0].subsets[0] chrs = [] for codeStr in subsetCodeStr.split('\\'): if codeStr: chrs.append(chr(subset[int(codeStr[1:], 8)])) return ''.join(chrs) class TextEncodingTestCase(NearTestCase): """Tests of expected Unicode and encoding behaviour """ def setUp(self): self.vera = TTFont("Vera", "Vera.ttf") pdfmetrics.registerFont(self.vera) self.styNormal = ParagraphStyle(name='Helvetica', fontName='Helvetica-Oblique') self.styTrueType = ParagraphStyle(name='TrueType', fontName='Vera') def testStringWidth(self): msg = 'Hello World' self.assertNear(pdfmetrics.stringWidth(msg, 'Courier', 10),66.0) self.assertNear(pdfmetrics.stringWidth(msg, 'Helvetica', 10),51.67) self.assertNear(pdfmetrics.stringWidth(msg, 'Times-Roman', 10),50.27) self.assertNear(pdfmetrics.stringWidth(msg, 'Vera', 10),57.7685546875) uniMsg1 = "Hello World" self.assertNear(pdfmetrics.stringWidth(uniMsg1, 'Courier', 10),66.0) self.assertNear(pdfmetrics.stringWidth(uniMsg1, 'Helvetica', 10),51.67) self.assertNear(pdfmetrics.stringWidth(uniMsg1, 'Times-Roman', 10),50.27) self.assertNear(pdfmetrics.stringWidth(uniMsg1, 'Vera', 10),57.7685546875) # Courier are all 600 ems wide. So if one 'measures as utf8' one will # get a wrong width as extra characters are seen self.assertEqual(len(testCp1252),52) self.assertNear(pdfmetrics.stringWidth(testCp1252, 'Courier', 10, 'cp1252'),312.0) # the test string has 5 more bytes and so "measures too long" if passed to # a single-byte font which treats it as a single-byte string. self.assertEqual(len(testUTF8),57) self.assertNear(pdfmetrics.stringWidth(testUTF8, 'Courier', 10),312.0) self.assertEqual(len(testUni),52) self.assertNear(pdfmetrics.stringWidth(testUni, 'Courier', 10),312.0) # now try a TrueType font. Should be able to accept Unicode or UTF8 self.assertNear(pdfmetrics.stringWidth(testUTF8, 'Vera', 10),279.809570313) self.assertNear(pdfmetrics.stringWidth(testUni, 'Vera', 10),279.809570313) @rlSkipIf(sys.getfilesystemencoding().lower() in ('ascii','ansi_x3.4-1968'),'need better file system encoding') def testUtf8FileName(self): fn=outputfile('test_pdfbase_utf8_filename') if not isUnicode(fn): fn = fn.decode('utf8') fn += u'_portr\xe4t.pdf' c = Canvas(fn) c.drawString(100,700, u'Filename='+fn) c.save() def testUtf8Canvas(self): """Verify canvas declared as utf8 autoconverts. This assumes utf8 input. It converts to the encoding of the underlying font, so both text lines APPEAR the same.""" c = Canvas(outputfile('test_pdfbase_encodings_utf8.pdf')) c.drawString(100,700, testUTF8) # Set a font with UTF8 encoding c.setFont('Vera', 12) # This should pass the UTF8 through unchanged c.drawString(100,600, testUTF8) # and this should convert from Unicode to UTF8 c.drawString(100,500, testUni) # now add a paragraph in Latin-1 in the latin-1 style p = Paragraph(testUTF8, style=self.styNormal, encoding="utf-8") w, h = p.wrap(150, 100) p.drawOn(c, 100, 400) #3 c.rect(100,300,w,h) # now add a paragraph in UTF-8 in the UTF-8 style p2 = Paragraph(testUTF8, style=self.styTrueType, encoding="utf-8") w, h = p2.wrap(150, 100) p2.drawOn(c, 300, 400) #4 c.rect(100,300,w,h) # now add a paragraph in Unicode in the latin-1 style p3 = Paragraph(testUni, style=self.styNormal) w, h = p3.wrap(150, 100) p3.drawOn(c, 100, 300) c.rect(100,300,w,h) # now add a paragraph in Unicode in the UTF-8 style p4 = Paragraph(testUni, style=self.styTrueType) p4.wrap(150, 100) p4.drawOn(c, 300, 300) c.rect(300,300,w,h) # now a graphic d1 = Drawing(400,50) d1.add(Ellipse(200,25,200,12.5, fillColor=None)) d1.add(String(200,25,testUTF8, textAnchor='middle', encoding='utf-8')) d1.drawOn(c, 100, 150) # now a graphic in utf8 d2 = Drawing(400,50) d2.add(Ellipse(200,25,200,12.5, fillColor=None)) d2.add(String(200,25,testUTF8, fontName='Vera', textAnchor='middle', encoding='utf-8')) d2.drawOn(c, 100, 100) # now a graphic in Unicode with T1 font d3 = Drawing(400,50) d3.add(Ellipse(200,25,200,12.5, fillColor=None)) d3.add(String(200,25,testUni, textAnchor='middle')) d3.drawOn(c, 100, 50) # now a graphic in Unicode with TT font d4 = Drawing(400,50) d4.add(Ellipse(200,25,200,12.5, fillColor=None)) d4.add(String(200,25,testUni, fontName='Vera', textAnchor='middle')) d4.drawOn(c, 100, 0) extracted = extractText(c.getCurrentPageContent()) self.assertEqual(extracted[0], expectedCp1252) if False and pdfmetrics.getFont('Vera').shapable: self.assertEqual(extracted[1:6],['copyright \\001 trademark \\002 r', 'egister', 'ed \\003 R', 'eportLab! Ol\\004!', 'copyright \\001 trademark \\002 r']) else: self.assertEqual(extracted[1], extracted[2]) #self.assertEqual(subsetToUnicode(self.vera, extracted[1]), testUni) c.save() class FontEncodingTestCase(unittest.TestCase): """Make documents with custom encodings of Type 1 built-in fonts. Nothing really to do with character encodings; this is about hacking the font itself""" def test0(self): "Make custom encodings of standard fonts" # make a custom encoded font. c = Canvas(outputfile('test_pdfbase_encodings.pdf')) c.setPageCompression(0) c.setFont('Helvetica', 12) c.drawString(100, 700, 'The text below should be in a custom encoding in which all vowels become "z"') # invent a new language where vowels are replaced with letter 'z' zenc = pdfmetrics.Encoding('EncodingWithoutVowels', 'WinAnsiEncoding') for ch in 'aeiou': zenc[ord(ch)] = 'z' for ch in 'AEIOU': zenc[ord(ch)] = 'Z' pdfmetrics.registerEncoding(zenc) # now we can make a font based on this encoding # AR hack/workaround: the name of the encoding must be a Python codec! f = pdfmetrics.Font('FontWithoutVowels', 'Helvetica-Oblique', 'EncodingWithoutVowels') pdfmetrics.registerFont(f) c.setFont('FontWithoutVowels', 12) c.drawString(125, 675, "The magic word is squamish ossifrage") # now demonstrate adding a Euro to MacRoman, which lacks one c.setFont('Helvetica', 12) c.drawString(100, 650, "MacRoman encoding lacks a Euro. We'll make a Mac font with the Euro at #219:") # WinAnsi Helvetica pdfmetrics.registerFont(pdfmetrics.Font('Helvetica-WinAnsi', 'Helvetica-Oblique', 'WinAnsiEncoding')) c.setFont('Helvetica-WinAnsi', 12) c.drawString(125, 625, 'WinAnsi with Euro: character 128 = "\200"') pdfmetrics.registerFont(pdfmetrics.Font('MacHelvNoEuro', 'Helvetica-Oblique', 'MacRomanEncoding')) c.setFont('MacHelvNoEuro', 12) c.drawString(125, 600, 'Standard MacRoman, no Euro: Character 219 = "\333"') # oct(219)=0333 # now make our hacked encoding euroMac = pdfmetrics.Encoding('MacWithEuro', 'MacRomanEncoding') euroMac[219] = 'Euro' pdfmetrics.registerEncoding(euroMac) pdfmetrics.registerFont(pdfmetrics.Font('MacHelvWithEuro', 'Helvetica-Oblique', 'MacWithEuro')) c.setFont('MacHelvWithEuro', 12) c.drawString(125, 575, 'Hacked MacRoman with Euro: Character 219 = "\333"') # oct(219)=0333 # now test width setting with and without _rl_accel - harder # make an encoding where 'm' becomes 'i' c.setFont('Helvetica', 12) c.drawString(100, 500, "Recode 'm' to 'i' and check we can measure widths. Boxes should surround letters.") sample = 'Mmmmm. ' * 6 + 'Mmmm' c.setFont('Helvetica-Oblique',12) c.drawString(125, 475, sample) w = c.stringWidth(sample, 'Helvetica-Oblique', 12) c.rect(125, 475, w, 12) narrowEnc = pdfmetrics.Encoding('m-to-i') narrowEnc[ord('m')] = 'i' narrowEnc[ord('M')] = 'I' pdfmetrics.registerEncoding(narrowEnc) pdfmetrics.registerFont(pdfmetrics.Font('narrow', 'Helvetica-Oblique', 'm-to-i')) c.setFont('narrow', 12) c.drawString(125, 450, sample) w = c.stringWidth(sample, 'narrow', 12) c.rect(125, 450, w, 12) c.setFont('Helvetica', 12) c.drawString(100, 400, "Symbol & Dingbats fonts - check we still get valid PDF in StandardEncoding") c.setFont('Symbol', 12) c.drawString(100, 375, 'abcdefghijklmn') c.setFont('ZapfDingbats', 12) c.drawString(300, 375, 'abcdefghijklmn') c.save() def makeSuite(): return makeSuiteForClasses( TextEncodingTestCase, #FontEncodingTestCase - nobbled for now due to old stuff which needs removing. ) #noruntests if __name__ == "__main__": unittest.TextTestRunner().run(makeSuite()) printLocation()