MVP implementation of encoder/decoder

mnemonic_key.py

@@ -1,41 +1,148 @@
+#!/usr/bin/env python3
 import hashlib
 import math
+import sys
+import argparse
 
 
-def access_bit(data, num):
+BITS_PER_WORD = 11
-    base = int(num // 8)
-    shift = int(num % 8)
-    return (data[base] & (1 << shift)) >> shift
 
 
-with open("./input.gpg", "rb") as f:
+def parse_args():
-    bs = bytearray(f.read())
+    parser = argparse.ArgumentParser("Encode and decode files as a mnemonic")
-
+    parser.add_argument("wordlist", type=str, help="The wordlist to use")
-with open("./english.txt", "r") as wordlist:
+    parser.add_argument("--decode", action="store_true")
-    words = [word.strip() for word in wordlist.readlines() if word.strip()]
+    parser.add_argument(
-
+        "--input", type=str, help="The input file when encoding"
-digest = hashlib.sha256(bs).digest()
+    )
+    parser.add_argument(
+        "--output", type=str, help="The file to write to when decoding"
+    )
+    parser.add_argument(
+        "--length", type=int, help="Length in bytes of the decoded output"
+    )
+    return parser.parse_args()
 
 
-bits = [access_bit(bs, i) for i in range(len(bs) * 8)]
+def bits_to_int(bits):
-checksum_bits = [access_bit(digest, i) for i in range(len(digest) * 8)]
+    """Convert passed bits into int
 
-n_bits = len(bits)
+    Least significant bit first
-nearest_mulitple_of_11 = math.floor((n_bits/11) + 1) * 11
+    """
-bits_missing = nearest_mulitple_of_11 - n_bits
+    b = 0
-bits += checksum_bits[0:bits_missing]
+    for i, bit in enumerate(bits):
+        b += bit << i
+
+    return b
 
 
-mnemonic = []
+def byte_to_bits(byte):
-for i in range(0, len(bits), 11):
+    """Convert byte into bit array
-    word_bits = bits[i:i+11]
-    word_int = 0
-    for j, bit in enumerate(word_bits):
-        word_int += bit << j
-    word = words[word_int] + "                    "
-    mnemonic.append(word[0:10])
 
-mnemonic += [""] * 5
+    Least significant bit first
-mnemonics = [mnemonic[i:i+5] for i in range(0, len(mnemonic), 5)]
+    """
-for m in mnemonics:
+    return [byte >> i & 1 for i in range(8)]
-    print("".join(m))
+
+
+def create_mnemonic(bites, words, bits_per_word=BITS_PER_WORD):
+    """Create mnemonic from bytes
+
+    Create mnemonic phrase from an input byte array.  Each byte
+    is convert into a bit array (least significant bit first) and
+    all such bit arrays are concatenated in the order of the input
+    bytes.  BITS_PER_WORD many bits are consumed from the beginning
+    of the array and converted into an integer (least significant
+    bit first) which is used as an index to look up a word in the
+    given wordlist.  A list of so looked up words is returned.
+
+    If necessary, the concatenated bit array is padded with the
+    beginning bits of the sha256 hash of the input byte array
+    to get to the next multiple of the word size.
+
+    :param bites: The bytes to convert.
+    :param words: The word list to use, must have 2**n many words
+    :param bits_per_word: The number of bits to consume per word.  The
+        word list should be 2**bits_per_word long
+    :retrun: Mnemonic phrase
+    """
+    digest = hashlib.sha256(bites).digest()
+
+    bits = []
+    for b in bites:
+        bits += byte_to_bits(b)
+
+    checksum_bits = []
+    for b in digest:
+        checksum_bits += byte_to_bits(b)
+
+    n_bits = len(bits)
+    smallest_n_bits = math.floor((n_bits/BITS_PER_WORD) + 1) * BITS_PER_WORD
+    bits_missing = smallest_n_bits - n_bits
+    bits += checksum_bits[0:bits_missing]
+
+    mnemonic = []
+    for i in range(0, len(bits), 11):
+        word_int = bits_to_int(bits[i:i+11])
+        mnemonic.append(words[word_int])
+
+    return mnemonic
+
+
+def parse_mnemonic(mnemonic, words):
+    """Parse mnemonic into bytearray using wordlist
+
+    For each word in the mnemonic, find it's 0 indexed position
+    in the wordlist, convert the the position into a bit array
+    (lest significant bit first) and concatenate all such bit
+    arrays. Pad it with 7 * [0] to ensure the last bits fit into the
+    last byte.  Convert the bit array into a byte array (least
+    significant bit first)
+
+    :param mnemonic: A list of words from the mnemonic
+    :param words: The (ordered) word list
+    :return: Decoded bytes
+    """
+    bits = []
+    for word in mnemonic:
+        i = words.index(word)
+        bits += [i >> j & 1 for j in range(11)]
+
+    n_bits = len(bits)
+    # Add padding bits to ensure the last chunck has 8 bits
+    bits += [0] * 7
+    bites = []
+    for i in range(0, n_bits, 8):
+        bites.append(bits_to_int(bits[i:i+8]))
+
+    return bytearray(bites)
+
+
+def run(word_file, encode, in_file, out_file, length):
+    with open(word_file, "r") as wordlist:
+        words = [word.strip() for word in wordlist.readlines() if word.strip()]
+
+    if encode:
+        with open(in_file, "rb") as in_file:
+            bites = bytearray(in_file.read())
+        mnemonic = create_mnemonic(bites, words)
+        print("\n".join(mnemonic))
+
+    else:
+        mnemonic = []
+        for line in sys.stdin.readlines():
+            mnemonic += line.split()
+
+        bites = parse_mnemonic(mnemonic, words)
+        with open(out_file, "wb") as out_file:
+            out_file.write(bites[:length])
+
+
+if __name__ == "__main__":
+    args = parse_args()
+    run(
+        args.wordlist,
+        not args.decode,
+        args.input,
+        args.output,
+        args.length if args.decode else None,
+    )