replaced tabs with spaces in order to fix codemod failure (#1999)

* replaced tabs with spaces in order to fix codemod failure

* fix bandit security warning

* more bandit security fix

* run pre-commit to fix format

Author: rshraga

packaging/install_torchdata.sh

@@ -34,13 +34,13 @@ if [ "$package_type" = "wheel" ]; then
   TORCHDATA_VERSION="$(pip show torchdata | grep ^Version: | sed 's/Version:  *//' | sed 's/+.\+//')"
 else
   TORCHDATA_VERSION="$(conda search --json 'torchdata[channel=pytorch-'"${channel}"] | \
-	  python -c "import json, os, re, sys; \
-	    cuver = 'cpu'; \
-	    pyver = os.environ.get('PYTHON_VERSION').replace('.', ''); \
-	    print(re.sub(r'\\+.*$', '',
-	      [x['version'] for x in json.load(sys.stdin)['torchdata'] \
-		      if 'py' + pyver in x['fn']][-1]))"
-	    )"
+    python -c "import json, os, re, sys; \
+      cuver = 'cpu'; \
+      pyver = os.environ.get('PYTHON_VERSION').replace('.', ''); \
+      print(re.sub(r'\\+.*$', '',
+        [x['version'] for x in json.load(sys.stdin)['torchdata'] \
+          if 'py' + pyver in x['fn']][-1]))"
+      )"
   echo "export CONDA_TORCHDATA_CONSTRAINT='- torchdata==${TORCHDATA_VERSION}'" >> "${BUILD_ENV_FILE}"
 fi
 

test/torchtext_unittest/datasets/test_cnndm.py

@@ -29,7 +29,7 @@ def _get_mock_dataset(root_dir):
             stories = []
             for i in range(5):
                 url = "_".join([source, split, str(i)])
-                h = hashlib.sha1()
+                h = hashlib.new("sha1", usedforsecurity=False)
                 h.update(url.encode())
                 filename = h.hexdigest() + ".story"
                 txt_file = os.path.join(source_dir, filename)
@@ -74,7 +74,7 @@ def _mock_split_list(source, split):
         story_fnames = []
         for i in range(5):
             url = "_".join([source, split, str(i)])
-            h = hashlib.sha1()
+            h = hashlib.new("sha1", usedforsecurity=False)
             h.update(url.encode())
             filename = h.hexdigest() + ".story"
             story_fnames.append(filename)

torchtext/datasets/cnndm.py

@@ -75,7 +75,7 @@ def _hash_urls(s: tuple):
     Code is inspired from https://github.com/abisee/cnn-dailymail/blob/master/make_datafiles.py
     """
     url = s[1]
-    h = hashlib.sha1()
+    h = hashlib.new("sha1", usedforsecurity=False)
     h.update(url)
     url_hash = h.hexdigest()
     story_fname = url_hash + ".story"

torchtext/utils.py

@@ -47,10 +47,8 @@ def validate_file(file_obj, hash_value, hash_type="sha256"):
         bool: return True if its a valid file, else False.
     """
 
-    if hash_type == "sha256":
-        hash_func = hashlib.sha256()
-    elif hash_type == "md5":
-        hash_func = hashlib.md5()
+    if hash_type in ("sha256", "md5"):
+        hash_func = hashlib.new(hash_type, usedforsecurity=False)
     else:
         raise ValueError