Issue #1625: BZ2File and bz2.decompress() now support multi-stream files.

nvawda · nvawda · commit 55b4338874ed · 2011-05-27T01:52:15.000+02:00
Initial patch by Nir Aides.
diff --git a/Lib/bz2.py b/Lib/bz2.py
@@ -76,6 +76,10 @@ def __init__(self, filename=None, mode="r", buffering=None,
             mode = "wb"
             mode_code = _MODE_WRITE
             self._compressor = BZ2Compressor()
+        elif mode in ("a", "ab"):
+            mode = "ab"
+            mode_code = _MODE_WRITE
+            self._compressor = BZ2Compressor()
         else:
             raise ValueError("Invalid mode: {!r}".format(mode))
 
@@ -161,14 +165,25 @@ def _check_can_seek(self):
     def _fill_buffer(self):
         if self._buffer:
             return True
-        if self._decompressor.eof:
-            self._mode = _MODE_READ_EOF
-            self._size = self._pos
-            return False
-        rawblock = self._fp.read(_BUFFER_SIZE)
+
+        if self._decompressor.unused_data:
+            rawblock = self._decompressor.unused_data
+        else:
+            rawblock = self._fp.read(_BUFFER_SIZE)
+
         if not rawblock:
-            raise EOFError("Compressed file ended before the "
-                           "end-of-stream marker was reached")
+            if self._decompressor.eof:
+                self._mode = _MODE_READ_EOF
+                self._size = self._pos
+                return False
+            else:
+                raise EOFError("Compressed file ended before the "
+                               "end-of-stream marker was reached")
+
+        # Continue to next stream.
+        if self._decompressor.eof:
+            self._decompressor = BZ2Decompressor()
+
         self._buffer = self._decompressor.decompress(rawblock)
         return True
 
@@ -384,9 +399,15 @@ def decompress(data):
     """
     if len(data) == 0:
         return b""
-    decomp = BZ2Decompressor()
-    result = decomp.decompress(data)
-    if not decomp.eof:
-        raise ValueError("Compressed data ended before the "
-                         "end-of-stream marker was reached")
-    return result
+
+    result = b""
+    while True:
+        decomp = BZ2Decompressor()
+        result += decomp.decompress(data)
+        if not decomp.eof:
+            raise ValueError("Compressed data ended before the "
+                             "end-of-stream marker was reached")
+        if not decomp.unused_data:
+            return result
+        # There is unused data left over. Proceed to next stream.
+        data = decomp.unused_data
diff --git a/Lib/test/test_bz2.py b/Lib/test/test_bz2.py
@@ -84,9 +84,9 @@ def getData(self, crlf=False):
         else:
             return self.DATA
 
-    def createTempFile(self, crlf=False):
+    def createTempFile(self, crlf=False, streams=1):
         with open(self.filename, "wb") as f:
-            f.write(self.getData(crlf))
+            f.write(self.getData(crlf) * streams)
 
     def testRead(self):
         # "Test BZ2File.read()"
@@ -95,6 +95,26 @@ def testRead(self):
             self.assertRaises(TypeError, bz2f.read, None)
             self.assertEqual(bz2f.read(), self.TEXT)
 
+    def testReadMultiStream(self):
+        # "Test BZ2File.read() with a multi stream archive"
+        self.createTempFile(streams=5)
+        with BZ2File(self.filename) as bz2f:
+            self.assertRaises(TypeError, bz2f.read, None)
+            self.assertEqual(bz2f.read(), self.TEXT * 5)
+
+    def testReadMonkeyMultiStream(self):
+        # "Test BZ2File.read() with a multi stream archive in which stream"
+        # "end is alined with internal buffer size"
+        buffer_size = bz2._BUFFER_SIZE
+        bz2._BUFFER_SIZE = len(self.DATA)
+        try:
+            self.createTempFile(streams=5)
+            with BZ2File(self.filename) as bz2f:
+                self.assertRaises(TypeError, bz2f.read, None)
+                self.assertEqual(bz2f.read(), self.TEXT * 5)
+        finally:
+            bz2._BUFFER_SIZE = buffer_size
+
     def testRead0(self):
         # "Test BBZ2File.read(0)"
         self.createTempFile()
@@ -114,6 +134,18 @@ def testReadChunk10(self):
                 text += str
             self.assertEqual(text, self.TEXT)
 
+    def testReadChunk10MultiStream(self):
+        # "Test BZ2File.read() in chunks of 10 bytes with a multi stream archive"
+        self.createTempFile(streams=5)
+        with BZ2File(self.filename) as bz2f:
+            text = b''
+            while 1:
+                str = bz2f.read(10)
+                if not str:
+                    break
+                text += str
+            self.assertEqual(text, self.TEXT * 5)
+
     def testRead100(self):
         # "Test BZ2File.read(100)"
         self.createTempFile()
@@ -151,6 +183,15 @@ def testReadLine(self):
             for line in sio.readlines():
                 self.assertEqual(bz2f.readline(), line)
 
+    def testReadLineMultiStream(self):
+        # "Test BZ2File.readline() with a multi stream archive"
+        self.createTempFile(streams=5)
+        with BZ2File(self.filename) as bz2f:
+            self.assertRaises(TypeError, bz2f.readline, None)
+            sio = BytesIO(self.TEXT * 5)
+            for line in sio.readlines():
+                self.assertEqual(bz2f.readline(), line)
+
     def testReadLines(self):
         # "Test BZ2File.readlines()"
         self.createTempFile()
@@ -159,13 +200,28 @@ def testReadLines(self):
             sio = BytesIO(self.TEXT)
             self.assertEqual(bz2f.readlines(), sio.readlines())
 
+    def testReadLinesMultiStream(self):
+        # "Test BZ2File.readlines() with a multi stream archive"
+        self.createTempFile(streams=5)
+        with BZ2File(self.filename) as bz2f:
+            self.assertRaises(TypeError, bz2f.readlines, None)
+            sio = BytesIO(self.TEXT * 5)
+            self.assertEqual(bz2f.readlines(), sio.readlines())
+
     def testIterator(self):
         # "Test iter(BZ2File)"
         self.createTempFile()
         with BZ2File(self.filename) as bz2f:
             sio = BytesIO(self.TEXT)
             self.assertEqual(list(iter(bz2f)), sio.readlines())
 
+    def testIteratorMultiStream(self):
+        # "Test iter(BZ2File) with a multi stream archive"
+        self.createTempFile(streams=5)
+        with BZ2File(self.filename) as bz2f:
+            sio = BytesIO(self.TEXT * 5)
+            self.assertEqual(list(iter(bz2f)), sio.readlines())
+
     def testClosedIteratorDeadlock(self):
         # "Test that iteration on a closed bz2file releases the lock."
         # http://bugs.python.org/issue3309
@@ -217,6 +273,17 @@ def testWriteMethodsOnReadOnlyFile(self):
             self.assertRaises(IOError, bz2f.write, b"a")
             self.assertRaises(IOError, bz2f.writelines, [b"a"])
 
+    def testAppend(self):
+        # "Test BZ2File.write()"
+        with BZ2File(self.filename, "w") as bz2f:
+            self.assertRaises(TypeError, bz2f.write)
+            bz2f.write(self.TEXT)
+        with BZ2File(self.filename, "a") as bz2f:
+            self.assertRaises(TypeError, bz2f.write)
+            bz2f.write(self.TEXT)
+        with open(self.filename, 'rb') as f:
+            self.assertEqual(self.decompress(f.read()), self.TEXT * 2)
+
     def testSeekForward(self):
         # "Test BZ2File.seek(150, 0)"
         self.createTempFile()
@@ -225,6 +292,14 @@ def testSeekForward(self):
             bz2f.seek(150)
             self.assertEqual(bz2f.read(), self.TEXT[150:])
 
+    def testSeekForwardMultiStream(self):
+        # "Test BZ2File.seek(150, 0) across stream boundaries"
+        self.createTempFile(streams=2)
+        with BZ2File(self.filename) as bz2f:
+            self.assertRaises(TypeError, bz2f.seek)
+            bz2f.seek(len(self.TEXT) + 150)
+            self.assertEqual(bz2f.read(), self.TEXT[150:])
+
     def testSeekBackwards(self):
         # "Test BZ2File.seek(-150, 1)"
         self.createTempFile()
@@ -233,13 +308,30 @@ def testSeekBackwards(self):
             bz2f.seek(-150, 1)
             self.assertEqual(bz2f.read(), self.TEXT[500-150:])
 
+    def testSeekBackwardsMultiStream(self):
+        # "Test BZ2File.seek(-150, 1) across stream boundaries"
+        self.createTempFile(streams=2)
+        with BZ2File(self.filename) as bz2f:
+            readto = len(self.TEXT) + 100
+            while readto > 0:
+                readto -= len(bz2f.read(readto))
+            bz2f.seek(-150, 1)
+            self.assertEqual(bz2f.read(), self.TEXT[100-150:] + self.TEXT)
+
     def testSeekBackwardsFromEnd(self):
         # "Test BZ2File.seek(-150, 2)"
         self.createTempFile()
         with BZ2File(self.filename) as bz2f:
             bz2f.seek(-150, 2)
             self.assertEqual(bz2f.read(), self.TEXT[len(self.TEXT)-150:])
 
+    def testSeekBackwardsFromEndMultiStream(self):
+        # "Test BZ2File.seek(-1000, 2) across stream boundaries"
+        self.createTempFile(streams=2)
+        with BZ2File(self.filename) as bz2f:
+            bz2f.seek(-1000, 2)
+            self.assertEqual(bz2f.read(), (self.TEXT * 2)[-1000:])
+
     def testSeekPostEnd(self):
         # "Test BZ2File.seek(150000)"
         self.createTempFile()
@@ -248,6 +340,14 @@ def testSeekPostEnd(self):
             self.assertEqual(bz2f.tell(), len(self.TEXT))
             self.assertEqual(bz2f.read(), b"")
 
+    def testSeekPostEndMultiStream(self):
+        # "Test BZ2File.seek(150000)"
+        self.createTempFile(streams=5)
+        with BZ2File(self.filename) as bz2f:
+            bz2f.seek(150000)
+            self.assertEqual(bz2f.tell(), len(self.TEXT) * 5)
+            self.assertEqual(bz2f.read(), b"")
+
     def testSeekPostEndTwice(self):
         # "Test BZ2File.seek(150000) twice"
         self.createTempFile()
@@ -257,6 +357,15 @@ def testSeekPostEndTwice(self):
             self.assertEqual(bz2f.tell(), len(self.TEXT))
             self.assertEqual(bz2f.read(), b"")
 
+    def testSeekPostEndTwiceMultiStream(self):
+        # "Test BZ2File.seek(150000) twice with a multi stream archive"
+        self.createTempFile(streams=5)
+        with BZ2File(self.filename) as bz2f:
+            bz2f.seek(150000)
+            bz2f.seek(150000)
+            self.assertEqual(bz2f.tell(), len(self.TEXT) * 5)
+            self.assertEqual(bz2f.read(), b"")
+
     def testSeekPreStart(self):
         # "Test BZ2File.seek(-150, 0)"
         self.createTempFile()
@@ -265,6 +374,14 @@ def testSeekPreStart(self):
             self.assertEqual(bz2f.tell(), 0)
             self.assertEqual(bz2f.read(), self.TEXT)
 
+    def testSeekPreStartMultiStream(self):
+        # "Test BZ2File.seek(-150, 0) with a multi stream archive"
+        self.createTempFile(streams=2)
+        with BZ2File(self.filename) as bz2f:
+            bz2f.seek(-150)
+            self.assertEqual(bz2f.tell(), 0)
+            self.assertEqual(bz2f.read(), self.TEXT * 2)
+
     def testFileno(self):
         # "Test BZ2File.fileno()"
         self.createTempFile()
@@ -510,6 +627,11 @@ def testDecompressIncomplete(self):
         # "Test decompress() function with incomplete data"
         self.assertRaises(ValueError, bz2.decompress, self.DATA[:-10])
 
+    def testDecompressMultiStream(self):
+        # "Test decompress() function for data with multiple streams"
+        text = bz2.decompress(self.DATA * 5)
+        self.assertEqual(text, self.TEXT * 5)
+
 def test_main():
     support.run_unittest(
         BZ2FileTest,
diff --git a/Misc/NEWS b/Misc/NEWS
@@ -161,6 +161,9 @@ Core and Builtins
 Library
 -------
 
+- Issue #1625: BZ2File and bz2.decompress() now support multi-stream files.
+  Initial patch by Nir Aides.
+
 - Issue #12175: BufferedReader.read(-1) now calls raw.readall() if available.
 
 - Issue #12175: FileIO.readall() now only reads the file position and size
