issue #2959: text string decoding, support for Info dict, updated tests

Tests/test_file_pdf.py

@@ -7,16 +7,13 @@ import tempfile
 
 class TestFilePdf(PillowTestCase):
 
-    def helper_save_as_pdf(self, mode, save_all=False):
+    def helper_save_as_pdf(self, mode, **kwargs):
         # Arrange
         im = hopper(mode)
         outfile = self.tempfile("temp_" + mode + ".pdf")
 
         # Act
-        if save_all:
-            im.save(outfile, save_all=True)
-        else:
-            im.save(outfile)
+        im.save(outfile, **kwargs)
 
         # Assert
         self.assertTrue(os.path.isfile(outfile))
@@ -134,18 +131,18 @@ class TestFilePdf(PillowTestCase):
 
     def test_pdf_append(self):
         # make a PDF file
-        pdf_filename = self.helper_save_as_pdf("RGB")
+        pdf_filename = self.helper_save_as_pdf("RGB", producer="pdfParser")
         # open it, check pages and info
         pdf = pdfParser.PdfParser(pdf_filename)
         self.assertEqual(len(pdf.pages), 1)
-        self.assertEqual(len(pdf.info), 0)
+        self.assertEqual(len(pdf.info), 1)
+        self.assertEqual(pdfParser.decode_text(pdf.info[b"Producer"]), "pdfParser")
         # append some info
-        pdf.info[b"Title"] = b"abc"
-        pdf.info[b"Author"] = b"def"
+        pdf.info[b"Title"] = pdfParser.encode_text("abc")
+        pdf.info[b"Author"] = pdfParser.encode_text("def")
         pdf.info[b"Subject"] = pdfParser.encode_text("ghi")
-        pdf.info[b"Keywords"] = b"jkl"
-        pdf.info[b"Creator"] = b"hopper()"
-        pdf.info[b"Producer"] = b"pdfParser"
+        pdf.info[b"Keywords"] = pdfParser.encode_text("jkl")
+        pdf.info[b"Creator"] = pdfParser.encode_text("hopper()")
         with open(pdf_filename, "r+b") as f:
             f.seek(0, os.SEEK_END)
             pdf.write_xref_and_trailer(f)
@@ -153,7 +150,7 @@ class TestFilePdf(PillowTestCase):
         pdf = pdfParser.PdfParser(pdf_filename)
         self.assertEqual(len(pdf.pages), 1)
         self.assertEqual(len(pdf.info), 6)
-        self.assertEqual(pdf.info[b"Title"], b"abc")
+        self.assertEqual(pdfParser.decode_text(pdf.info[b"Title"]), "abc")
         # append two images
         mode_CMYK = hopper("CMYK")
         mode_P = hopper("P")
@@ -162,7 +159,8 @@ class TestFilePdf(PillowTestCase):
         pdf = pdfParser.PdfParser(pdf_filename)
         self.assertEqual(len(pdf.pages), 3)
         self.assertEqual(len(pdf.info), 6)
-        self.assertEqual(pdf.info[b"Title"], b"abc")
+        self.assertEqual(pdfParser.decode_text(pdf.info[b"Title"]), "abc")
+        self.assertEqual(pdfParser.decode_text(pdf.info[b"Producer"]), "pdfParser")
 
     def test_pdf_parser(self):
         pdfParser.selftest()

src/PIL/PdfImagePlugin.py

@@ -48,12 +48,32 @@ def _save_all(im, fp, filename):
 def _save(im, fp, filename, save_all=False):
     resolution = im.encoderinfo.get("resolution", 72.0)
     is_appending = im.encoderinfo.get("append", False)
+    title = im.encoderinfo.get("title", None)
+    author = im.encoderinfo.get("author", None)
+    subject = im.encoderinfo.get("subject", None)
+    keywords = im.encoderinfo.get("keywords", None)
+    creator = im.encoderinfo.get("creator", None)
+    producer = im.encoderinfo.get("producer", None)
+
     if is_appending:
         existing_pdf = pdfParser.PdfParser(f=fp, filename=filename)
         fp.seek(0, io.SEEK_END)
     else:
         existing_pdf = pdfParser.PdfParser()
 
+    if title:
+        existing_pdf.info[b"Title"] = pdfParser.encode_text(title)
+    if author:
+        existing_pdf.info[b"Author"] = pdfParser.encode_text(author)
+    if subject:
+        existing_pdf.info[b"Subject"] = pdfParser.encode_text(subject)
+    if keywords:
+        existing_pdf.info[b"Keywords"] = pdfParser.encode_text(keywords)
+    if creator:
+        existing_pdf.info[b"Creator"] = pdfParser.encode_text(creator)
+    if producer:
+        existing_pdf.info[b"Producer"] = pdfParser.encode_text(producer)
+
     #
     # make sure image data is available
     im.load()

src/PIL/pdfParser.py

@@ -18,10 +18,63 @@ else:  # Python 3.x
         return s.encode("us-ascii")
 
 
+# see 7.9.2.2 Text String Type on page 86 and D.3 PDFDocEncoding Character Set on page 656
 def encode_text(s):
     return codecs.BOM_UTF16_BE + s.encode("utf_16_be")
 
 
+PDFDocEncoding = {
+    0x16: u"\u0017",
+    0x18: u"\u02D8",
+    0x19: u"\u02C7",
+    0x1A: u"\u02C6",
+    0x1B: u"\u02D9",
+    0x1C: u"\u02DD",
+    0x1D: u"\u02DB",
+    0x1E: u"\u02DA",
+    0x1F: u"\u02DC",
+    0x80: u"\u2022",
+    0x81: u"\u2020",
+    0x82: u"\u2021",
+    0x83: u"\u2026",
+    0x84: u"\u2014",
+    0x85: u"\u2013",
+    0x86: u"\u0192",
+    0x87: u"\u2044",
+    0x88: u"\u2039",
+    0x89: u"\u203A",
+    0x8A: u"\u2212",
+    0x8B: u"\u2030",
+    0x8C: u"\u201E",
+    0x8D: u"\u201C",
+    0x8E: u"\u201D",
+    0x8F: u"\u2018",
+    0x90: u"\u2019",
+    0x91: u"\u201A",
+    0x92: u"\u2122",
+    0x93: u"\uFB01",
+    0x94: u"\uFB02",
+    0x95: u"\u0141",
+    0x96: u"\u0152",
+    0x97: u"\u0160",
+    0x98: u"\u0178",
+    0x99: u"\u017D",
+    0x9A: u"\u0131",
+    0x9B: u"\u0142",
+    0x9C: u"\u0153",
+    0x9D: u"\u0161",
+    0x9E: u"\u017E",
+    0xA0: u"\u20AC",
+    }
+def decode_text(b):
+    if b[:len(codecs.BOM_UTF16_BE)] == codecs.BOM_UTF16_BE:
+        return b[len(codecs.BOM_UTF16_BE):].decode("utf_16_be")
+    elif str == bytes:  # Python 2.x
+        return u"".join(PDFDocEncoding.get(ord(byte), byte) for byte in b)
+    else:
+        return "".join(PDFDocEncoding.get(byte, chr(byte)) for byte in b)
+
+
 class PdfFormatError(RuntimeError):
     pass
 
@@ -667,6 +720,10 @@ class PdfParser:
 
 
 def selftest():
+    assert encode_text("abc") == b"\xFE\xFF\x00a\x00b\x00c"
+    assert decode_text(b"\xFE\xFF\x00a\x00b\x00c") == "abc"
+    assert decode_text(b"abc") == "abc"
+    assert decode_text(b"\x1B a \x1C") == u"\u02D9 a \u02DD"
     assert PdfParser.interpret_name(b"Name#23Hash") == b"Name#Hash"
     assert PdfParser.interpret_name(b"Name#23Hash", as_text=True) == "Name#Hash"
     assert IndirectReference(1,2) == IndirectReference(1,2)