1.建立Python檔案物件的讀寫模式(r,w模式)與建立Java輸入輸出流；
   FileInputStream inputStream=new FileInputStream(new File("E:\\workspace\\tmpfile\\farrago.txt")); 
   FileOutputStream outputStream=new FileOutputStream(new File("E:\\workspace\\tmpfile\\outagainb.txt"));
 2.序列化Python物件，序列化成字串如toJson和序列化成二進位制；
 3.向檔案物件中寫入序列化物件。
 
 PS：有的序列化包，有方便的方法直接將序列化和寫入過程合二為一。numpy的save和load感覺就是序列化而已，那麼其必要性不大了？

import umsgpack
import pickle
import numpy as np

"""
 1.建立Python檔案物件的讀寫模式(r,w模式)與建立Java輸入輸出流；
   FileInputStream inputStream=new FileInputStream(new File("E:\\workspace\\tmpfile\\farrago.txt")); 
   FileOutputStream outputStream=new FileOutputStream(new File("E:\\workspace\\tmpfile\\outagainb.txt"));
 2.序列化Python物件，序列化成字串如toJson和序列化成二進位制；
 3.向檔案物件中寫入序列化物件。
 
 PS：有的序列化包，有方便的方法直接將序列化和寫入過程合二為一。numpy的save和load感覺就是序列化而已，那麼其必要性不大了？
   
"""


with open('test0.bin', 'wb') as f:
    rs = umsgpack.packb({u"compact": True, u"schema": 0})
    f.write(rs)

with open('test0.bin', 'rb') as f:
    print(umsgpack.unpackb(f.read()))
    print("_+"*20)

with open('test.bin', 'wb') as f:
    print(umsgpack.pack({u"compact": True, u"schema": 0}, f))
    print(umsgpack.pack([1,2,3], f))

with open('test.bin', 'rb') as f:
    print(umsgpack.unpack(f))
    print(umsgpack.unpack(f))
    print("_+"*20)

with open('test2.bin', 'wb') as f:
    rs = pickle.dumps({u"compact": True, u"schema": 0})
    f.write(rs)
    print(pickle.loads(rs))

with open('test2.bin', 'rb') as f:
    print(pickle.load(f))
    print("_+"*20)

with open('test3.bin', 'wb') as f:
    pickle.dump({u"compact": True, u"schema": 0},f)

with open('test3.bin', 'rb') as f:
    print(pickle.load(f))
    print("_+="*20)


with open('test4.bin', 'wb') as f:
    """有沒有覺得numpy的save與load就是個二進位制序列化協議"""
    np.save(f, 'abc')
    np.save(f, np.arange(10))

with open('test4.bin', 'rb') as f:
    print(np.load(f))
    print(np.load(f))
    print("888" * 20)