Add support for user specified blocksize and increase default to 8192

cyanea-bt · Dec 6, 2020 · 5bdc0df · 5bdc0df
1 parent 01a79af
commit 5bdc0df
Show file tree

Hide file tree

Showing 2 changed files with 13 additions and 3 deletions.
diff --git a/filehash/filehash.py b/filehash/filehash.py
@@ -149,7 +149,7 @@ class FileHash:
     Class wrapping the hashlib module to facilitate calculating file hashes.
     """
 
-    def __init__(self, hash_algorithm='sha256', chunk_size=4096):
+    def __init__(self, hash_algorithm='sha256', chunk_size=8192):
         """
         Initialize the FileHash class.
 
@@ -164,6 +164,7 @@ def __init__(self, hash_algorithm='sha256', chunk_size=4096):
         if hash_algorithm not in SUPPORTED_ALGORITHMS:
             raise ValueError("Error, unsupported hash/checksum algorithm: {0}".format(hash_algorithm))
         self.chunk_size = chunk_size
+        # print("FileHash:INIT: chunk_size =", self.chunk_size)
         self.hash_algorithm = hash_algorithm
 
     def hash_file(self, filename):

diff --git a/filehash/filehash_cli.py b/filehash/filehash_cli.py
@@ -6,7 +6,7 @@
 
 
 default_hash_algorithm = 'sha256'
-
+default_chunk_size = 8192
 def create_parser():
     parser = argparse.ArgumentParser(
         description="Tool for calculating the checksum / hash of a file or directory."
@@ -20,6 +20,15 @@ def create_parser():
         ),
         default=default_hash_algorithm
     )
+    parser.add_argument(
+        u"-b",
+        u"--blocksize",
+        type=int,
+        help=u"Blocksize/ChunkSize for hash algorithm to use.  Valid values are multiples of 4096.  Defaults to \"{0}\"".format(
+            default_chunk_size
+        ),
+        default=default_chunk_size
+    )
 
     parser_group = parser.add_mutually_exclusive_group(required=True)
     parser_group.add_argument(
@@ -95,7 +104,7 @@ def main():
         parser.print_help()
         sys.exit(1)
 
-    hasher = FileHash(args.algorithm.lower())
+    hasher = FileHash(args.algorithm.lower(), args.blocksize)
     if args.checksums:
         process_checksum_file(args.checksums, hasher)
     elif args.directory: