AES GCM Python implementation

Question

I've recently started looking into encryption in python and I came across with pycryptodome library. I made this simple program that encrypt/decrypt a file using AES-GCM. Could someone take a look if I implemented the encryption and authentication correctly? For sake of simplicity, I'm posting only the encryption and makeKey functions.

import os
import struct
import secrets
import hahslib
from Cryptodome.Cipher import AES
def makeKey(password, salt): 
 key = hashlib.pbkdf2_hmac('sha512',password.encode('ascii'), salt, 1000000)
 return key[:32]
def encrypt(password, inputFileName):
 #creates the needed variables for encryption
 salt = secrets.token_bytes(16)
 key = makeKey(password, salt)
 nonce = secrets.token_bytes(AES.block_size)
 cipher = AES.new(key, AES.MODE_GCM, nonce=nonce)
 #put the file size, salt and nonce in outputfile
 fileSize = os.path.getsize(inputFileName)
 with open("output", "wb") as outputFile:
 outputFile.write(struct.pack('<Q', fileSize))
 outputFile.write(salt)
 outputFile.write(nonce)
 #beginning of encryption itself
 chunkSize = 64 * 1024
 with open(inputFileName, "rb") as inputFile:
 while True:
 chunk = inputFile.read(chunkSize)
 if len(chunk) == 0:
 outputFile.write(cipher.digest())
 break
 outputFile.write(cipher.encrypt(chunk))

Answer 1

Handling files - the Python way

Without being able to test run your code at the momemt, there is one thing I can already recommend to you, which is to make use of Python's powerful with statement. Using with will make sure that files will be closed no matter what happens (e.g. unexpected exceptions). with also works with more than one file as described in this SO post.

Also consider adding the output filename as parameter to your function. You can even give your momentary value of "output" as default value.

---

Minor notes on style

I'm not sure if you're aware of the official Style Guide for Python, often just called PEP8. The official recommendation is to use snake_case for variable and function names. This would create a more unified code appearance when working with other Python libraries which often stick to that convention. As always, exceptions apply.

---

Below is your code including these changes.

def make_key(password, salt):
 ...
def encrypt(password, input_filename, output_filename="output"):
 #creates the needed variables for encryption
 salt = secrets.token_bytes(16)
 key = make_key(password, salt)
 nonce = secrets.token_bytes(AES.block_size)
 cipher = AES.new(key, AES.MODE_GCM, nonce=nonce)
 # put the file size, salt and nonce in outputfile
 filesize = os.path.getsize(input_filename)
 with open(output_filename, "wb") as output_file,\
 open(input_filename, "rb") as input_file:
 output_file.write(struct.pack('<Q', filesize))
 output_file.write(salt)
 output_file.write(nonce)
 #beginning of encryption itself
 chunkSize = 64 * 1024
 while True:
 chunk = input_file.read(chunkSize)
 if len(chunk) == 0:
 output_file.write(cipher.digest())
 break
 output_file.write(cipher.encrypt(chunk))
 print("File encrypted successfully!")