def roundtrip_truncated(arr):
    f = BytesIO()
    format.write_array(f, arr)
    #BytesIO is one byte short
    f2 = BytesIO(f.getvalue()[0:-1])
    arr2 = format.read_array(f2)
    return arr2

def _unpickle_array(bytes):
    arr = read_array(BytesIO(bytes))

    # All datetimes should be stored as M8[ns].  When unpickling with
    # numpy1.6, it will read these as M8[us].  So this ensures all
    # datetime64 types are read as MS[ns]
    if is_datetime64_dtype(arr):
        arr = arr.view(_NS_DTYPE)

    return arr

def test_version_2_0():
    f = BytesIO()
    # requires more than 2 byte for header
    dt = [(("%d" % i) * 100, float) for i in range(500)]
    d = np.ones(1000, dtype=dt)

    format.write_array(f, d, version=(2, 0))
    with warnings.catch_warnings(record=True) as w:
        warnings.filterwarnings('always', '', UserWarning)
        format.write_array(f, d)
        assert_(w[0].category is UserWarning)

    f.seek(0)
    n = format.read_array(f)
    assert_array_equal(d, n)

    # 1.0 requested but data cannot be saved this way
    assert_raises(ValueError, format.write_array, f, d, (1, 0))

def roundtrip_randsize(arr):
    f = BytesIO()
    format.write_array(f, arr)
    f2 = BytesIOSRandomSize(f.getvalue())
    arr2 = format.read_array(f2)
    return arr2

def roundtrip(arr):
    f = BytesIO()
    format.write_array(f, arr)
    f2 = BytesIO(f.getvalue())
    arr2 = format.read_array(f2)
    return arr2

def _unpickle_array(bytes):
    arr = read_array(BytesIO(bytes))
    return arr

def _unpickle_array(bytes):
    arr = read_array(StringIO(bytes))
    return arr

def roundtrip(arr):
    f = BytesIO()
    format.write_array(f, arr)
    f2 = BytesIO(f.getvalue())
    arr2 = format.read_array(f2, allow_pickle=True)
    return arr2

        # Load the info file and get the task and metric
        info_file = ls(os.path.join(input_dir, basename + '*_public.info'))[0]
        info = get_info(info_file)
        score_name = info['task'][0:-15] + info['metric'][0:-7].upper()
        predict_name = basename
        try:
            # Get the last prediction from the res subdirectory (must end with
            # '.predict')
            predict_file = ls(os.path.join(
                prediction_dir, basename + '_' + set_name + '*.predict'))[-1]
            if (predict_file == []):
                raise IOError('Missing prediction file {}'.format(basename))
            predict_name = predict_file[-predict_file[::-1].index(filesep):-
                                        predict_file[::-1].index('.') - 1]
            # Read the solution and prediction values into numpy arrays
            solution = read_array(solution_file)
            prediction = read_array(predict_file)
            if (solution.shape != prediction.shape):
                raise ValueError(
                    'Bad prediction shape {}'.format(prediction.shape))

            try:
                # Compute the score prescribed by the info file (for regression
                # scores, no normalization)
                if info['metric'] == 'r2_metric' or info[
                        'metric'] == 'a_metric':
                    # Remove NaN and Inf for regression
                    solution = sanitize_array(solution)
                    prediction = sanitize_array(prediction)
                    score = eval(info['metric'] + '(solution, prediction, "' +
                                 info['task'] + '")')

	def toArray(s):
	    f=StringIO(s)
	    arr=format.read_array(f)
	    return arr

def from_string(s): 
    f = StringIO(s) 
    arr = format.read_array(f) 
    return arr