Documentation for Hasher Module

module
torrentfile.hasher

Piece/File Hashers for Bittorrent meta file contents.

Classes
  • Hasher Piece hasher for Bittorrent V1 files.
  • HasherV2 Calculate the root hash and piece layers for file contents.
  • HasherHybrid Calculate root and piece hashes for creating hybrid torrent file.
Functions
  • merkle_root(blocks) Calculate the merkle root for a seq of sha256 hash digests.

torrentfile.hasher.merkle_root(blocks)

Calculate the merkle root for a seq of sha256 hash digests.

Source code in torrentfile\hasher.py
def merkle_root(blocks):
    """Calculate the merkle root for a seq of sha256 hash digests."""
    while len(blocks) > 1:
        blocks = [sha256(x + y).digest() for x, y in zip(*[iter(blocks)] * 2)]
    return blocks[0]

torrentfile.hasher.Hasher

Piece hasher for Bittorrent V1 files.

Takes a sorted list of all file paths, calculates sha1 hash for fixed size pieces of file data from each file seemlessly until the last piece which may be smaller than others.

Parameters:

Name Type Description Default
paths `list`

List of files.

required
piece_length `int`

Size of chuncks to split the data into.

required
total `int`

Sum of all files in file list.

required

__init__(self, paths, piece_length) special

Generate hashes of piece length data from filelist contents.

Source code in torrentfile\hasher.py
def __init__(self, paths, piece_length):
    """Generate hashes of piece length data from filelist contents."""
    self.piece_length = piece_length
    self.paths = paths
    self.total = sum([os.path.getsize(i) for i in self.paths])
    self.index = 0
    self.current = open(self.paths[0], "rb")
    logging.debug(
        "Hashing v1 torrent file. Size: %s Piece Length: %s",
        humanize_bytes(self.total),
        humanize_bytes(self.piece_length),
    )

__iter__(self) special

Iterate through feed pieces.

Returns:

Type Description
`iterator`

Iterator for leaves/hash pieces.

Source code in torrentfile\hasher.py
def __iter__(self):
    """Iterate through feed pieces.

    Returns
    -------
    self : `iterator`
        Iterator for leaves/hash pieces.
    """
    return self

__next__(self) special

Generate piece-length pieces of data from input file list.

Source code in torrentfile\hasher.py
def __next__(self):
    """Generate piece-length pieces of data from input file list."""
    while True:
        piece = bytearray(self.piece_length)
        size = self.current.readinto(piece)
        if size == 0:
            if not self.next_file():
                raise StopIteration
        elif size < self.piece_length:
            return self._handle_partial(piece[:size])
        else:
            return sha1(piece).digest()  # nosec

next_file(self)

Seemlessly transition to next file in file list.

Source code in torrentfile\hasher.py
def next_file(self):
    """Seemlessly transition to next file in file list."""
    self.index += 1
    if self.index < len(self.paths):
        self.current.close()
        self.current = open(self.paths[self.index], "rb")
        return True
    return False

torrentfile.hasher.HasherV2

Calculate the root hash and piece layers for file contents.

Iterates over 16KiB blocks of data from given file, hashes the data, then creates a hash tree from the individual block hashes until size of hashed data equals the piece-length. Then continues the hash tree until root hash is calculated.

Parameters:

Name Type Description Default
path `str`

Path to file.

required
piece_length `int`

Size of layer hashes pieces.

required

__init__(self, path, piece_length) special

Calculate and store hash information for specific file.

Source code in torrentfile\hasher.py
def __init__(self, path, piece_length):
    """Calculate and store hash information for specific file."""
    self.path = path
    self.root = None
    self.piece_layer = None
    self.layer_hashes = []
    self.piece_length = piece_length
    self.num_blocks = piece_length // BLOCK_SIZE
    logging.debug(
        "Hashing partial v2 torrent file. Piece Length: %s Path: %s",
        humanize_bytes(self.piece_length),
        str(self.path),
    )

    with open(self.path, "rb") as fd:
        self.process_file(fd)

process_file(self, fd)

Calculate hashes over 16KiB chuncks of file content.

Parameters:

Name Type Description Default
fd `str`

Opened file in read mode.

required
Source code in torrentfile\hasher.py
def process_file(self, fd):
    """Calculate hashes over 16KiB chuncks of file content.

    Parameters
    ----------
    fd : `str`
        Opened file in read mode.
    """
    while True:
        total = 0
        blocks = []
        leaf = bytearray(BLOCK_SIZE)
        # generate leaves of merkle tree

        for _ in range(self.num_blocks):
            size = fd.readinto(leaf)
            total += size
            if not size:
                break
            blocks.append(sha256(leaf[:size]).digest())

        # blocks is empty mean eof
        if not blocks:
            break
        if len(blocks) != self.num_blocks:
            # when size of file doesn't fill the last block
            if not self.layer_hashes:
                # when the there is only one block for file

                next_pow_2 = 1 << int(math.log2(total) + 1)
                remaining = ((next_pow_2 - total) // BLOCK_SIZE) + 1

            else:
                # when the file contains multiple pieces
                remaining = self.num_blocks - size
            # pad the the rest with zeroes to fill remaining space.
            padding = [bytes(HASH_SIZE) for _ in range(remaining)]
            blocks.extend(padding)
        # calculate the root hash for the merkle tree up to piece-length

        layer_hash = merkle_root(blocks)
        self.layer_hashes.append(layer_hash)
    self._calculate_root()

torrentfile.hasher.HasherHybrid

Calculate root and piece hashes for creating hybrid torrent file.

Create merkle tree layers from sha256 hashed 16KiB blocks of contents. With a branching factor of 2, merge layer hashes until blocks equal piece_length bytes for the piece layer, and then the root hash.

Parameters:

Name Type Description Default
path `str`

path to target file.

required
piece_length `int`

piece length for data chunks.

required

__init__(self, path, piece_length) special

Construct Hasher class instances for each file in torrent.

Source code in torrentfile\hasher.py
def __init__(self, path, piece_length):
    """Construct Hasher class instances for each file in torrent."""
    self.path = path
    self.piece_length = piece_length
    self.pieces = []
    self.layer_hashes = []
    self.piece_layer = None
    self.root = None
    self.padding_piece = None
    self.padding_file = None
    self.amount = piece_length // BLOCK_SIZE
    logging.debug(
        "Hashing partial Hybrid torrent file. Piece Length: %s Path: %s",
        humanize_bytes(self.piece_length),
        str(self.path),
    )
    with open(path, "rb") as data:
        self._process_file(data)