Fix the RLE implementation (#19)

* speedup runlength decoding * fix typing * add PR to changelog * test: sort of kind of test the rle fix --------- Co-authored-by: Benedikt Fuchs <[email protected]> Co-authored-by: Benedikt Fuchs <[email protected]>
dhdaines · Nov 26, 2024 · 03ca1b1 · 03ca1b1
1 parent 7972d04
commit 03ca1b1
Show file tree

Hide file tree

Showing 4 changed files with 19 additions and 15 deletions.
diff --git a/playa/color.py b/playa/color.py
@@ -1,4 +1,4 @@
-from typing import Dict, List, NamedTuple, Union, Tuple
+from typing import Dict, NamedTuple, Union, Tuple
 
 from playa.exceptions import PDFInterpreterError
 from playa.parser import LIT, PDFObject, PSLiteral

diff --git a/playa/runlength.py b/playa/runlength.py
@@ -5,6 +5,8 @@
 #  * public domain *
 #
 
+from typing import List
+
 
 def rldecode(data: bytes) -> bytes:
     """RunLength decoder (Adobe version) implementation based on PDF Reference
@@ -19,21 +21,18 @@ def rldecode(data: bytes) -> bytes:
         (2 to 128) times during decompression. A length value of 128
         denotes EOD.
     """
-    decoded = b""
-    i = 0
-    while i < len(data):
-        length = data[i]
+    decoded_array: List[int] = []
+    data_iter = iter(data)
+
+    while True:
+        length = next(data_iter, 128)
         if length == 128:
             break
 
-        if length >= 0 and length < 128:
-            for j in range(i + 1, (i + 1) + (length + 1)):
-                decoded += bytes((data[j],))
-            i = (i + 1) + (length + 1)
+        if 0 <= length < 128:
+            decoded_array.extend((next(data_iter) for _ in range(length + 1)))
 
         if length > 128:
-            run = bytes((data[i + 1],)) * (257 - length)
-            decoded += run
-            i = (i + 1) + 1
-
-    return decoded
+            run = [next(data_iter)] * (257 - length)
+            decoded_array.extend(run)
+    return bytes(decoded_array)
diff --git a/tests/test_lazy_api.py b/tests/test_lazy_api.py
@@ -7,7 +7,6 @@
 import pytest
 
 import playa
-from playa.parser import LIT
 from playa.color import PREDEFINED_COLORSPACE, Color
 from playa.exceptions import PDFEncryptionError
 

diff --git a/tests/test_pdftypes.py b/tests/test_pdftypes.py
@@ -3,6 +3,7 @@
 """
 
 from playa.data_structures import NameTree, NumberTree
+from playa.runlength import rldecode
 
 NUMTREE1 = {
     "Kids": [
@@ -75,3 +76,8 @@ def test_name_tree():
         (b"xylophone", 123),
         (b"zzyzx", {"x": "y"}),
     ]
+
+
+def test_rle():
+    large_white_image_encoded = bytes([129, 255] * (3 * 3000 * 4000 // 128))
+    _ = rldecode(large_white_image_encoded)