Changeset 687

Timestamp:

01/18/06 11:50:23 (3 years ago)

Author:

laotzu

Message:

Add support for PDF 1.5 cross-reference streams and object streams.

Files:

• pypdf/trunk/pyPdf/pdf.py (modified) (9 diffs)

pypdf/trunk/pyPdf/pdf.py

@@ -38 +38 @@
-
-import re
+import struct
+import zlib
+try:
+    from cStringIO import StringIO
+except ImportError:
+    from StringIO import StringIO
-
-
@@ -158 +164 @@
-class PdfFileReader(object):
-    def __init__(self, stream):
+        self.flattenedPages = None
+        self.resolvedObjects = {}
-        self.read(stream)
-        self.stream = stream
-        self.flattenedPages = None
-        self.resolvedObjects = {}
-
-    def getNumPages(self):
@@ -207 +213 @@
-        if retval != None:
-            return retval
+        if indirectReference.generation == 0 and \
+           self.xref_objStm.has_key(indirectReference.idnum):
+            # indirect reference to object in object stream
+            # read the entire object stream into memory
+            stmnum,idx = self.xref_objStm[indirectReference.idnum]
+            objStm = self.getObject(IndirectObject(stmnum, 0, self))
+            assert objStm['/Type'] == '/ObjStm'
+            assert idx < objStm['/N']
+            streamData = StringIO(decodeStreamData(objStm))
+            for i in range(objStm['/N']):
+                objnum = NumberObject.readFromStream(streamData)
+                readNonWhitespace(streamData)
+                streamData.seek(-1, 1)
+                offset = NumberObject.readFromStream(streamData)
+                readNonWhitespace(streamData)
+                streamData.seek(-1, 1)
+                t = streamData.tell()
+                streamData.seek(objStm['/First']+offset, 0)
+                obj = readObject(streamData, self)
+                self.resolvedObjects[0][objnum] = obj
+                streamData.seek(t, 0)
+            return self.resolvedObjects[0][indirectReference.idnum]
-        start = self.xref[indirectReference.generation][indirectReference.idnum]
-        self.stream.seek(start, 0)
-
-
-
-
-
+
+
+
-        retval = readObject(self.stream, self)
-
-
-
+
-        return retval
+
+    def readObjectHeader(self, stream):
+        idnum = readUntilWhitespace(stream)
+        generation = readUntilWhitespace(stream)
+        obj = stream.read(3)
+        readNonWhitespace(stream)
+        stream.seek(-1, 1)
+        return int(idnum), int(generation)
+
+    def cacheIndirectObject(self, generation, idnum, obj):
+        if not self.resolvedObjects.has_key(generation):
+            self.resolvedObjects[generation] = {}
+        self.resolvedObjects[generation][idnum] = obj
-
-    def read(self, stream):
@@ -234 +271 @@
-        # read all cross reference tables and their trailers
-        self.xref = {}
+        self.xref_objStm = {}
-        self.trailer = {}
-        while 1:
-            # load the xref table
-            stream.seek(startxref, 0)
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
-            else:
-
-
-
-
-
-
-
-
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
-
-    def readNextEndLine(self, stream):
@@ -318 +402 @@
-            return NumberObject.readFromStream(stream)
-        peek = stream.read(20)
-20
+len(peek)
-        if re.match(r"(\d+)\s(\d+)\sR", peek) != None:
-            return IndirectObject.readFromStream(stream, pdf)
@@ -336 +420 @@
-
-    def readFromStream(stream):
-readUntilWhitespace(stream
+stream.read(4
-        if word == "true":
-            return BooleanObject(True)
-
+
+
-            return BooleanObject(False)
-        assert False
@@ -398 +483 @@
-                break
-            generation += tok
-
+
+
+
+
+
+
-        return IndirectObject(int(idnum), int(generation), pdf)
-    readFromStream = staticmethod(readFromStream)
@@ -550 +640 @@
-            stream.seek(-1, 1)
-            value = readObject(stream, pdf)
+            if retval.has_key(key):
+                # multiple definitions of key not handled yet
+                assert False
-            retval[key] = value
-        pos = stream.tell()
@@ -607 +700 @@
-    return tok
-
+def decodeStreamData(stream):
+    if stream.get("/Filter",None) == "/FlateDecode":
+        data = zlib.decompress(stream["__streamdata__"])
+    else:
+        # unsupported Filter
+        assert False
+    predictor = stream.get("/DecodeParms", {}).get("/Predictor", 1)
+    if predictor != 1:
+        columns = stream["/DecodeParms"]["/Columns"]
+        if predictor >= 10:
+            newdata = ""
+            # PNG prediction can vary from row to row
+            rowlength = columns + 1
+            assert len(data) % rowlength == 0
+            prev_rowdata = "\x00"*rowlength
+            for row in range(len(data) / rowlength):
+                rowdata = list(data[(row*rowlength):((row+1)*rowlength)])
+                filterByte = ord(rowdata[0])
+                if filterByte == 0:
+                    pass
+                elif filterByte == 1:
+                    for i in range(2, rowlength):
+                        rowdata[i] = chr((ord(rowdata[i]) + ord(rowdata[i-1])) % 256)
+                elif filterByte == 2:
+                    for i in range(1, rowlength):
+                        rowdata[i] = chr((ord(rowdata[i]) + ord(prev_rowdata[i])) % 256)
+                else:
+                    # unsupported PNG filter
+                    assert False
+                prev_rowdata = rowdata
+                newdata += ''.join(rowdata[1:])
+            data = newdata
+        else:
+            # unsupported predictor
+            assert False
+    return data
+
+def convertToInt(d, size):
+    if size <= 4:
+        d = "\x00\x00\x00\x00" + d
+        d = d[-4:]
+        return struct.unpack(">l", d)[0]
+    elif size <= 8:
+        d = "\x00\x00\x00\x00\x00\x00\x00\x00" + d
+        d = d[-8:]
+        return struct.unpack(">q", d)[0]
+    else:
+        # size too big
+        assert False
+
-
-if __name__ == "__main__":
-    input = PdfFileReader(file("cc-cc.pdf", "rb"))
-    output = PdfFileWriter()
-
-
+
+
+
+
+
+
+
+
+