CCS †裏† Advent Calendar 2020の20日目の記事ように書いたgitのパーサー

ccsgit.py

import zlib
import os
import hashlib


def ls_obj():
    object_dirs = os.listdir(path='.git/objects')
    for object_dir in object_dirs:
        if len(object_dir) == 2:
            print(object_dir + (os.listdir(path='.git/objects/' + object_dir)[0]))


def ls_ref():
    ref_files = os.listdir(path='.git/refs/heads')
    for ref_file in ref_files:
        print(ref_file)


def cat_ref(name: str):
    with open('.git/refs/heads/' + name, 'r') as ref_file:
        print(ref_file.read())


def cat_head():
    with open('.git/HEAD', 'r') as head_file:
        print(head_file.read())



class GitObject:
    @property
    def object_type(self) -> str:
        return "unknown"

    def print(self):
        print("not implemented")


class Blob(GitObject):

    def __init__(self, data: bytes):
        self.data = data.decode()
        pass

    def print(self):
        print(self.data)

    @property
    def object_type(self) -> str:
        return "blob"


class TreeElement:
    def __init__(self, mode: str, filename: str, sha1_hash: int):
        self.mode = mode
        self.filename = filename
        self.sha1_hash = sha1_hash


def get_tree_elements(data: bytes):
    if len(data) == 0:
        return []
    else:
        mode, tmp = data.split(b' ', 1)
        filename, tmp2 = tmp.split(b'\0', 1)
        sha1_hash = tmp2[:20]
        body = tmp2[20:]
        return [TreeElement(mode.decode(), filename.decode(), sha1_hash.hex())] + get_tree_elements(body)


class Tree(GitObject):

    def __init__(self, data: bytes):
        self.elements = get_tree_elements(data)

    @property
    def object_type(self) -> str:
        return "tree"

    def print(self):
        for element in self.elements:
            print("{} {} {}".format(element.mode, element.filename, element.sha1_hash))


class Commit(GitObject):

    def __init__(self, data):
        self.data = data.decode()
        pass

    def print(self):
        print(self.data)

    @property
    def object_type(self) -> str:
        return "commit"


# メタクラス使うともっとスマートにできるが...
def create_git_object(obj_type: str, data: bytes):
    if obj_type == "blob":
        return Blob(data)
    elif obj_type == "tree":
        return Tree(data)
    elif obj_type == "commit":
        return Commit(data)
    else:
        return GitObject()


def cat_obj(hash_val):
    with open('.git/objects/' + hash_val[:2] + "/" + hash_val[2:], 'rb') as gitobj:
        res = zlib.decompress(gitobj.read())
        print("hash:{}".format(hashlib.sha1(res).hexdigest()))

        header, body = res.split(b'\0', 1)
        obj_type, obj_size = header.decode().split()
        obj = create_git_object(obj_type, body)

        print("type:{}".format(obj.object_type))
        obj.print()


def calc_padding(n: int) -> int:
    floor = (n - 2) // 8
    target = (floor + 1) * 8 + 2
    return target - n


def cat_index():
    with open('.git/index', 'rb') as index_file:
        index_file_data = index_file.read()
        header_sig = index_file_data[0:4]
        header_ver = index_file_data[4:8]
        header_index_number = int.from_bytes(index_file_data[8:12], byteorder='big')
        print("version:{}".format(int.from_bytes(header_ver, byteorder='big')))

        offset = 12

        print("---")

        for _ in range(header_index_number):
            ctime_s = index_file_data[offset:offset + 4]
            ctime_n = index_file_data[offset + 4:offset + 8]
            mtime_s = index_file_data[offset + 8:offset + 12]
            mtime_n = index_file_data[offset + 12:offset + 16]
            dev = index_file_data[offset + 16:offset + 20]
            inode = index_file_data[offset + 20:offset + 24]
            mode = int.from_bytes(index_file_data[offset + 24:offset + 28], byteorder='big')
            uid = int.from_bytes(index_file_data[offset + 28:offset + 32], byteorder='big')
            gid = index_file_data[offset + 32:offset + 36]
            file_size = int.from_bytes(index_file_data[offset + 36:offset + 40], byteorder='big')
            sha1 = index_file_data[offset + 40:offset + 60].hex()

            assume_valid_flag = int.from_bytes(index_file_data[offset + 60:offset + 62], byteorder='big') & 0x800
            extended_flag = int.from_bytes(index_file_data[offset + 60:offset + 62], byteorder='big') & 0x800
            stage = int.from_bytes(index_file_data[offset + 60:offset + 62], byteorder='big') & 0x300
            name_length = int.from_bytes(index_file_data[offset + 60:offset + 62], byteorder='big') & 0x0FFF

            name = index_file_data[offset + 62:offset + 62 + name_length].decode()

            padding = calc_padding(name_length)
            offset = offset + 62 + name_length + padding
            print("{} {} {}".format(oct(mode), name, sha1))