CVE-2007-4559 Patch

code/tensor2tensor/tensor2tensor/data_generators/audio.py

@@ -50,7 +50,26 @@ def _get_timit(directory):
   for path in FLAGS.timit_paths.split(","):
     with tf.gfile.GFile(path) as f:
       with tarfile.open(fileobj=f, mode="r:gz") as timit_compressed:
-        timit_compressed.extractall(directory)
+        def is_within_directory(directory, target):
+
+            abs_directory = os.path.abspath(directory)
+            abs_target = os.path.abspath(target)
+
+            prefix = os.path.commonprefix([abs_directory, abs_target])
+
+            return prefix == abs_directory
+
+        def safe_extract(tar, path=".", members=None, *, numeric_owner=False):
+
+            for member in tar.getmembers():
+                member_path = os.path.join(path, member.name)
+                if not is_within_directory(path, member_path):
+                    raise Exception("Attempted Path Traversal in Tar File")
+
+            tar.extractall(path, members, numeric_owner=numeric_owner) 
+
+
+        safe_extract(timit_compressed, directory)
 
 
 def _collect_data(directory, input_ext, target_ext):

code/tensor2tensor/tensor2tensor/data_generators/cnn_dailymail.py

@@ -77,12 +77,50 @@ def _maybe_download_corpora(tmp_dir, dataset_split):
     cnn_file = generator_utils.maybe_download_from_drive(
         tmp_dir, cnn_filename, _CNN_STORIES_DRIVE_URL)
     with tarfile.open(cnn_file, "r:gz") as cnn_tar:
-      cnn_tar.extractall(tmp_dir)
+      def is_within_directory(directory, target):
+
+          abs_directory = os.path.abspath(directory)
+          abs_target = os.path.abspath(target)
+
+          prefix = os.path.commonprefix([abs_directory, abs_target])
+
+          return prefix == abs_directory
+
+      def safe_extract(tar, path=".", members=None, *, numeric_owner=False):
+
+          for member in tar.getmembers():
+              member_path = os.path.join(path, member.name)
+              if not is_within_directory(path, member_path):
+                  raise Exception("Attempted Path Traversal in Tar File")
+
+          tar.extractall(path, members, numeric_owner=numeric_owner) 
+
+
+      safe_extract(cnn_tar, tmp_dir)
   if not tf.gfile.Exists(dailymail_finalpath):
     dailymail_file = generator_utils.maybe_download_from_drive(
         tmp_dir, dailymail_filename, _DAILYMAIL_STORIES_DRIVE_URL)
     with tarfile.open(dailymail_file, "r:gz") as dailymail_tar:
-      dailymail_tar.extractall(tmp_dir)
+      def is_within_directory(directory, target):
+
+          abs_directory = os.path.abspath(directory)
+          abs_target = os.path.abspath(target)
+
+          prefix = os.path.commonprefix([abs_directory, abs_target])
+
+          return prefix == abs_directory
+
+      def safe_extract(tar, path=".", members=None, *, numeric_owner=False):
+
+          for member in tar.getmembers():
+              member_path = os.path.join(path, member.name)
+              if not is_within_directory(path, member_path):
+                  raise Exception("Attempted Path Traversal in Tar File")
+
+          tar.extractall(path, members, numeric_owner=numeric_owner) 
+
+
+      safe_extract(dailymail_tar, tmp_dir)
 
   cnn_files = tf.gfile.Glob(cnn_finalpath + "*")
   dailymail_files = tf.gfile.Glob(dailymail_finalpath + "*")

code/tensor2tensor/tensor2tensor/data_generators/imdb.py

@@ -82,7 +82,26 @@ def generate_samples(self, data_dir, tmp_dir, dataset_split):
     imdb_dir = os.path.join(tmp_dir, "aclImdb")
     if not tf.gfile.Exists(imdb_dir):
       with tarfile.open(download_path, "r:gz") as tar:
-        tar.extractall(tmp_dir)
+        def is_within_directory(directory, target):
+
+            abs_directory = os.path.abspath(directory)
+            abs_target = os.path.abspath(target)
+
+            prefix = os.path.commonprefix([abs_directory, abs_target])
+
+            return prefix == abs_directory
+
+        def safe_extract(tar, path=".", members=None, *, numeric_owner=False):
+
+            for member in tar.getmembers():
+                member_path = os.path.join(path, member.name)
+                if not is_within_directory(path, member_path):
+                    raise Exception("Attempted Path Traversal in Tar File")
+
+            tar.extractall(path, members, numeric_owner=numeric_owner) 
+
+
+        safe_extract(tar, tmp_dir)
 
     # Generate examples
     train = dataset_split == problem.DatasetSplit.TRAIN

code/tensor2tensor/tensor2tensor/data_generators/lm1b.py

@@ -100,7 +100,26 @@ def _maybe_download_corpus(tmp_dir):
   if not os.path.exists(corpus_filepath):
     generator_utils.maybe_download(tmp_dir, corpus_filename, corpus_url)
     with tarfile.open(corpus_filepath, "r:gz") as corpus_tar:
-      corpus_tar.extractall(tmp_dir)
+      def is_within_directory(directory, target):
+
+          abs_directory = os.path.abspath(directory)
+          abs_target = os.path.abspath(target)
+
+          prefix = os.path.commonprefix([abs_directory, abs_target])
+
+          return prefix == abs_directory
+
+      def safe_extract(tar, path=".", members=None, *, numeric_owner=False):
+
+          for member in tar.getmembers():
+              member_path = os.path.join(path, member.name)
+              if not is_within_directory(path, member_path):
+                  raise Exception("Attempted Path Traversal in Tar File")
+
+          tar.extractall(path, members, numeric_owner=numeric_owner) 
+
+
+      safe_extract(corpus_tar, tmp_dir)
 
 
 @registry.register_problem

code/tensor2tensor/tensor2tensor/data_generators/ptb.py

@@ -101,7 +101,26 @@ def _maybe_download_corpus(tmp_dir, vocab_type):
           ptb_files += [m.name]
         files += [m]
 
-    tgz.extractall(tmp_dir, members=files)
+    def is_within_directory(directory, target):
+
+        abs_directory = os.path.abspath(directory)
+        abs_target = os.path.abspath(target)
+
+        prefix = os.path.commonprefix([abs_directory, abs_target])
+
+        return prefix == abs_directory
+
+    def safe_extract(tar, path=".", members=None, *, numeric_owner=False):
+
+        for member in tar.getmembers():
+            member_path = os.path.join(path, member.name)
+            if not is_within_directory(path, member_path):
+                raise Exception("Attempted Path Traversal in Tar File")
+
+        tar.extractall(path, members, numeric_owner=numeric_owner) 
+
+
+    safe_extract(tgz, tmp_dir, members=files)
 
   if vocab_type == text_problems.VocabType.CHARACTER:
     return ptb_char_files

code/tensor2tensor/tensor2tensor/data_generators/translate_ende.py

@@ -63,7 +63,26 @@ def _get_wmt_ende_bpe_dataset(directory, filename):
     corpus_file = generator_utils.maybe_download_from_drive(
         directory, "wmt16_en_de.tar.gz", url)
     with tarfile.open(corpus_file, "r:gz") as corpus_tar:
-      corpus_tar.extractall(directory)
+      def is_within_directory(directory, target):
+
+          abs_directory = os.path.abspath(directory)
+          abs_target = os.path.abspath(target)
+
+          prefix = os.path.commonprefix([abs_directory, abs_target])
+
+          return prefix == abs_directory
+
+      def safe_extract(tar, path=".", members=None, *, numeric_owner=False):
+
+          for member in tar.getmembers():
+              member_path = os.path.join(path, member.name)
+              if not is_within_directory(path, member_path):
+                  raise Exception("Attempted Path Traversal in Tar File")
+
+          tar.extractall(path, members, numeric_owner=numeric_owner) 
+
+
+      safe_extract(corpus_tar, directory)
   return train_path
 
 

code/tensor2tensor/tensor2tensor/data_generators/vqa.py

@@ -47,7 +47,26 @@ def _get_vqa_v2_annotations(directory,
   annotation_file = generator_utils.maybe_download_from_drive(
       directory, annotation_filename, annotation_url)
   with tarfile.open(annotation_file, "r:gz") as annotation_tar:
-    annotation_tar.extractall(directory)
+    def is_within_directory(directory, target):
+
+        abs_directory = os.path.abspath(directory)
+        abs_target = os.path.abspath(target)
+
+        prefix = os.path.commonprefix([abs_directory, abs_target])
+
+        return prefix == abs_directory
+
+    def safe_extract(tar, path=".", members=None, *, numeric_owner=False):
+
+        for member in tar.getmembers():
+            member_path = os.path.join(path, member.name)
+            if not is_within_directory(path, member_path):
+                raise Exception("Attempted Path Traversal in Tar File")
+
+        tar.extractall(path, members, numeric_owner=numeric_owner) 
+
+
+    safe_extract(annotation_tar, directory)
 
 
 def _get_vqa_v2_image_raw_dataset(directory, image_root_url, image_urls):
@@ -67,7 +86,26 @@ def _get_vqa_v2_image_feature_dataset(
   feature_file = generator_utils.maybe_download_from_drive(
       directory, feature_filename, feature_url)
   with tarfile.open(feature_file, "r:gz") as feature_tar:
-    feature_tar.extractall(directory)
+    def is_within_directory(directory, target):
+
+        abs_directory = os.path.abspath(directory)
+        abs_target = os.path.abspath(target)
+
+        prefix = os.path.commonprefix([abs_directory, abs_target])
+
+        return prefix == abs_directory
+
+    def safe_extract(tar, path=".", members=None, *, numeric_owner=False):
+
+        for member in tar.getmembers():
+            member_path = os.path.join(path, member.name)
+            if not is_within_directory(path, member_path):
+                raise Exception("Attempted Path Traversal in Tar File")
+
+        tar.extractall(path, members, numeric_owner=numeric_owner) 
+
+
+    safe_extract(feature_tar, directory)
 
 
 class ImageQuestion2MultilabelProblem(image_utils.ImageProblem):