dull_impl_of_bitcask.py

# -*- coding: utf-8 -*-

import os
from datetime import datetime
from binascii import crc32
import cPickle as pickle
from glob import glob

TABLE = {}
current_active = max([int(f.replace('.sst', '')) for f in glob('*.sst')] or [1])
THRESHOLD = 10000

def io_read(filename, start, length):
    with open(filename, 'rb') as f:
        f.seek(start)
        return f.read(length)

def io_write(crc, ts, key_size, value_size, key, value):
    global current_active
    active_filename = '%s.sst' % current_active
    try:
        if os.path.getsize(active_filename) >= THRESHOLD:
            current_active += 1
            active_filename = '%s.sst' % current_active
    except:
        pass
    with open(active_filename, 'a') as f:
        data = pickle.dumps((crc, ts, key_size, value_size, key, value))
        start = f.tell()
        length = len(data)
        f.write(data)
        return active_filename, start, length

def get(key):
    info = TABLE.get(key)
    if not info:
        raise Exception('not found.')
    pickled_data = io_read(*info)
    data = pickle.loads(pickled_data)
    # check tuple, crc, the same key?
    return data[5]


def put(key, value):
    key_size = len(key)
    value_size = len(value)
    ts = datetime.now()
    crc = crc32('{0}{1}{2}{3}{4}'.format(
        ts, key_size, value_size, key, value))
    info = io_write(crc, ts, key_size, value_size, key, value)
    TABLE[key] = info
    return True

def delete(key):
    pass

def list_keys():
    pass

if __name__ == '__main__':
    put('test', 'hello world')
    print get('test')