Python numpy 模块，int8() 实例源码

def run_test_matmul_aa_correlator_kernel(self, ntime, nstand, nchan, misalign=0):
        x_shape = (ntime, nchan, nstand*2)
        perm = [1,0,2]
        x8 = ((np.random.random(size=x_shape+(2,))*2-1)*127).astype(np.int8)
        x = x8.astype(np.float32).view(np.complex64).reshape(x_shape)
        x = x.transpose(perm)
        x = x[..., misalign:]
        b_gold = np.matmul(H(x), x)
        triu = np.triu_indices(x.shape[-1], 1)
        b_gold[..., triu[0], triu[1]] = 0
        x = x8.view(bf.DataType.ci8).reshape(x_shape)
        x = bf.asarray(x, space='cuda')
        x = x.transpose(perm)
        x = x[..., misalign:]
        b = bf.zeros_like(b_gold, space='cuda')
        self.linalg.matmul(1, None, x, 0, b)
        b = b.copy('system')
        np.testing.assert_allclose(b, b_gold, RTOL*10, ATOL)

def test_quantize_from_probs2(size, resolution):
    set_random_seed(make_seed(size, resolution))
    probs = np.exp(np.random.random(size)).astype(np.float32)
    probs2 = probs.reshape((1, size))
    quantized = quantize_from_probs2(probs2, resolution)
    assert quantized.shape == probs2.shape
    assert quantized.dtype == np.int8
    assert np.all(quantized.sum(axis=1) == resolution)

    # Check that quantized result is closer to target than any other value.
    quantized = quantized.reshape((size, ))
    target = resolution * probs / probs.sum()
    distance = np.abs(quantized - target).sum()
    for combo in itertools.combinations(range(size), resolution):
        other = np.zeros(size, np.int8)
        for i in combo:
            other[i] += 1
        assert other.sum() == resolution
        other_distance = np.abs(other - target).sum()
        assert distance <= other_distance

def test_server_logprob_normalized(N, V, C, M):
    model = generate_fake_model(N, V, C, M)
    config = TINY_CONFIG.copy()
    config['model_num_clusters'] = M
    model['config'] = config
    server = TreeCatServer(model)

    # The total probability of all categorical rows should be 1.
    ragged_index = model['suffstats']['ragged_index']
    factors = []
    for v in range(V):
        C = ragged_index[v + 1] - ragged_index[v]
        factors.append([one_hot(c, C) for c in range(C)])
    data = np.array(
        [np.concatenate(columns) for columns in itertools.product(*factors)],
        dtype=np.int8)
    logprobs = server.logprob(data)
    logtotal = np.logaddexp.reduce(logprobs)
    assert logtotal == pytest.approx(0.0, abs=1e-5)

def test_server_median(N, V, C, M):
    model = generate_fake_model(N, V, C, M)
    config = TINY_CONFIG.copy()
    config['model_num_clusters'] = M
    model['config'] = config
    server = TreeCatServer(model)

    # Evaluate on random data.
    counts = np.random.randint(10, size=[V], dtype=np.int8)
    table = generate_dataset(N, V, C)['table']
    median = server.median(counts, table.data)
    assert median.shape == table.data.shape
    assert median.dtype == np.int8
    for v in range(V):
        beg, end = table.ragged_index[v:v + 2]
        totals = median[:, beg:end].sum(axis=1)
        assert np.all(totals == counts[v])

def observed_perplexity(self, counts):
        """Compute perplexity = exp(entropy) of observed variables.

        Perplexity is an information theoretic measure of the number of
        clusters or latent classes. Perplexity is a real number in the range
        [1, M], where M is model_num_clusters.

        Args:
            counts: A [V]-shaped array of multinomial counts.

        Returns:
            A [V]-shaped numpy array of perplexity.
        """
        V, E, M, R = self._VEMR
        if counts is not None:
            counts = np.ones(V, dtype=np.int8)
        assert counts.shape == (V, )
        assert counts.dtype == np.int8
        assert np.all(counts > 0)
        observed_entropy = np.empty(V, dtype=np.float32)
        for v in range(V):
            beg, end = self._ragged_index[v:v + 2]
            probs = np.dot(self._feat_cond[beg:end, :], self._vert_probs[v, :])
            observed_entropy[v] = multinomial_entropy(probs, counts[v])
        return np.exp(observed_entropy)

def quantize_from_probs2(probs, resolution):
    """Quantize multiple non-normalized probs to given resolution.

    Args:
        probs: An [N, M]-shaped numpy array of non-normalized probabilities.

    Returns:
        An [N, M]-shaped array of quantized probabilities such that
        np.all(result.sum(axis=1) == resolution).
    """
    assert len(probs.shape) == 2
    N, M = probs.shape
    probs = probs / probs.sum(axis=1, keepdims=True)
    result = np.zeros(probs.shape, np.int8)
    range_N = np.arange(N, dtype=np.int32)
    for _ in range(resolution):
        sample = probs.argmax(axis=1)
        result[range_N, sample] += 1
        probs[range_N, sample] -= 1.0 / resolution
    return result

def count_observations(ragged_index, data):
    """Count the observations in each cell of a ragged data array.

    Args:
        ragged_index: A [V+1]-shaped numpy array as returned by
            make_ragged_index.
        data: A [N, R]-shaped ragged array of multinomial count data, where
            N is the number of rows and R = ragged_index[-1].

    Returns:
        A [N, V]-shaped array whose entries are the number of observations
        in each cell of data.
    """
    N, R = data.shape
    assert R == ragged_index[-1]
    V = len(ragged_index) - 1
    counts = np.zeros([N, V], np.int8)
    for v in range(V):
        beg, end = ragged_index[v:v + 2]
        counts[:, v] = data[:, beg:end].sum(axis=1)
    return counts

def __getitem__(self, index):
        img_name = self.files[self.split][index]
        img_path = self.root + '/' + self.split + '/' + img_name
        lbl_path = self.root + '/' + self.split + 'annot/' + img_name

        img = m.imread(img_path)
        img = np.array(img, dtype=np.uint8)

        lbl = m.imread(lbl_path)
        lbl = np.array(lbl, dtype=np.int8)

        if self.augmentations is not None:
            img, lbl = self.augmentations(img, lbl)

        if self.is_transform:
            img, lbl = self.transform(img, lbl)

        return img, lbl

def _get_dtype_maps():
    """ Get dictionaries to map numpy data types to ITK types and the 
    other way around.
    """

    # Define pairs
    tmp = [ (np.float32, 'MET_FLOAT'),  (np.float64, 'MET_DOUBLE'),
            (np.uint8, 'MET_UCHAR'),    (np.int8, 'MET_CHAR'),
            (np.uint16, 'MET_USHORT'),  (np.int16, 'MET_SHORT'),
            (np.uint32, 'MET_UINT'),    (np.int32, 'MET_INT'),
            (np.uint64, 'MET_ULONG'),   (np.int64, 'MET_LONG') ]

    # Create dictionaries
    map1, map2 = {}, {}
    for np_type, itk_type in tmp:
        map1[np_type.__name__] = itk_type
        map2[itk_type] = np_type.__name__

    # Done
    return map1, map2

def flip_code(code):
    if isinstance(code, (numpy.dtype,type)):
        # since several things map to complex64 we must carefully select
        # the opposite that is an exact match (ticket 1518)
        if code == numpy.int8:
            return gdalconst.GDT_Byte
        if code == numpy.complex64:
            return gdalconst.GDT_CFloat32

        for key, value in codes.items():
            if value == code:
                return key
        return None
    else:
        try:
            return codes[code]
        except KeyError:
            return None

def UnHidePdfintoPng(img,x,y):
    counterx = 0
    countery = 0
    h , w , c = img.shape
    data = []
    while counterx <= x:
        temp1 = (img[counterx][countery][0] & 0x03) << 6;
        temp2 = (img[counterx][countery][ 1]  & 0x03) << 4
        temp3 = (img[counterx][countery ][2]  & 0x03) << 2
        temp4 = (img[counterx][countery ][3]  & 0x03)
        data.append(temp1 | temp2 | temp3 | temp4)
        if counterx == x and countery == y:
            #print ('EOF Found')
            break
        if(countery == w - 1):
            countery = 0
            counterx += 1
        else:
            countery += 1
    data = np.int8(data)
    return data

def inc_region(self, dst, y, x, h, w):
    '''Incremets dst in the specified region. Runs fastest on np.int8, but not much slower on
    np.int16.'''

    dh, dw = dst.shape
    h2 = h // 2
    w2 = w // 2
    py = y - h2 
    px = x - w2 
    y_min = max(0, py)
    y_max = min(dh, y + h2)
    x_min = max(0, px)
    x_max = min(dw, x + w2)
    if y_max - y_min <= 0 or x_max - x_min <= 0:
      return

    dst[y_min:y_max, x_min:x_max] += 1

def __read_spike_fixed(self, numpts=40):
        """
        Read a spike with a fixed waveform length (40 time bins)

        -------------------------------------------
        Returns the time, waveform and trig2 value.

        The returned objects must be converted to a SpikeTrain then
        added to the Block.

        ID: 29079
        """

        # float32 -- spike time stamp in ms since start of SpikeTrain
        time = np.fromfile(self._fsrc, dtype=np.float32, count=1)

        # int8 * 40 -- spike shape -- use numpts for spike_var
        waveform = np.fromfile(self._fsrc, dtype=np.int8,
                               count=numpts).reshape(1, 1, numpts)

        # uint8 -- point of return to noise
        trig2 = np.fromfile(self._fsrc, dtype=np.uint8, count=1)

        return time, waveform, trig2

def __read_spike_fixed(self, numpts=40):
        """
        Read a spike with a fixed waveform length (40 time bins)

        -------------------------------------------
        Returns the time, waveform and trig2 value.

        The returned objects must be converted to a SpikeTrain then
        added to the Block.

        ID: 29079
        """

        # float32 -- spike time stamp in ms since start of SpikeTrain
        time = np.fromfile(self._fsrc, dtype=np.float32, count=1)

        # int8 * 40 -- spike shape -- use numpts for spike_var
        waveform = np.fromfile(self._fsrc, dtype=np.int8,
                               count=numpts).reshape(1, 1, numpts)

        # uint8 -- point of return to noise
        trig2 = np.fromfile(self._fsrc, dtype=np.uint8, count=1)

        return time, waveform, trig2

def _check_valid_data(self, data):
        """Checks that the incoming data is a 2 x #elements ndarray of ints.

        Parameters
        ----------
        data : :obj:`numpy.ndarray`
            The data to verify.

        Raises
        ------
        ValueError
            If the data is not of the correct shape or type.
        """
        if data.dtype.type != np.int8 and data.dtype.type != np.int16 \
                and data.dtype.type != np.int32 and data.dtype.type != np.int64 \
                and data.dtype.type != np.uint8 and data.dtype.type != np.uint16 \
                and data.dtype.type != np.uint32 and data.dtype.type != np.uint64:
            raise ValueError('Must initialize image coords with a numpy int ndarray')
        if data.shape[0] != 2:
            raise ValueError('Illegal data array passed to image coords. Must have 2 coordinates')
        if len(data.shape) > 2:
            raise ValueError('Illegal data array passed to point cloud. Must have 1 or 2 dimensions')

def _gene_signature(self,wm,size,key):
        '''????????????????????????'''
        wm = cv2.resize(wm,(size,size))        
        wU,_,wV = np.linalg.svd(np.mat(wm))


        sumU = np.sum(np.array(wU),axis=0)
        sumV = np.sum(np.array(wV),axis=0)

        sumU_mid = np.median(sumU)
        sumV_mid = np.median(sumV)

        sumU=np.array([1 if sumU[i] >sumU_mid else 0 for i in range(len(sumU)) ])
        sumV=np.array([1 if sumV[i] >sumV_mid else 0 for i in range(len(sumV)) ])

        uv_xor=np.logical_xor(sumU,sumV)

        np.random.seed(key)
        seq=np.random.randint(2,size=len(uv_xor))

        signature = np.logical_xor(uv_xor, seq)

        sqrts = int(np.sqrt(size))
        return np.array(signature,dtype=np.int8).reshape((sqrts,sqrts))

def _gene_signature(self,wm,key):
        '''????????????????????????'''
        wm = cv2.resize(wm,(256,256))        
        wU,_,wV = np.linalg.svd(np.mat(wm))


        sumU = np.sum(np.array(wU),axis=0)
        sumV = np.sum(np.array(wV),axis=0)

        sumU_mid = np.median(sumU)
        sumV_mid = np.median(sumV)

        sumU=np.array([1 if sumU[i] >sumU_mid else 0 for i in range(len(sumU)) ])
        sumV=np.array([1 if sumV[i] >sumV_mid else 0 for i in range(len(sumV)) ])

        uv_xor=np.logical_xor(sumU,sumV)

        np.random.seed(key)
        seq=np.random.randint(2,size=len(uv_xor))

        signature = np.logical_xor(uv_xor, seq)
        return np.array(signature,dtype=np.int8)

def _gene_signature(self,wU,wV,key):
        '''????????????????????????'''
        sumU = np.sum(wU,axis=0)
        sumV = np.sum(wV,axis=0)

        sumU_mid = np.median(sumU)
        sumV_mid = np.median(sumV)

        sumU=np.array([1 if sumU[i] >sumU_mid else 0 for i in range(len(sumU)) ])
        sumV=np.array([1 if sumV[i] >sumV_mid else 0 for i in range(len(sumV)) ])

        uv_xor=np.logical_xor(sumU,sumV)

        np.random.seed(key)
        seq=np.random.randint(2,size=len(uv_xor))

        signature = np.logical_xor(uv_xor, seq)
        return np.array(signature,dtype=np.int8)

def test_bin_counts(self):
        metadata2 = {'cellType': 'int32ud-500',
                     'extent': self.extent,
                     'crs': '+proj=longlat +datum=WGS84 +no_defs ',
                     'bounds': {
                         'minKey': {'col': 0, 'row': 0},
                         'maxKey': {'col': 0, 'row': 0}},
                     'layoutDefinition': {
                         'extent': self.extent,
                         'tileLayout': {'tileCols': 4, 'tileRows': 4, 'layoutCols': 1, 'layoutRows': 1}}}

        arr2 = np.int8([[[1, 1, 1, 1],
                         [3, 1, 1, 1],
                         [4, 3, 1, 1],
                         [5, 4, 3, 1]]])

        tile2 = Tile(arr2, 'INT', -500)
        rdd2 = BaseTestClass.pysc.parallelize([(self.spatial_key, tile2)])
        tiled2 = TiledRasterLayer.from_numpy_rdd(LayerType.SPATIAL, rdd2,
                                                 metadata2)

        hist2 = tiled2.get_class_histogram()
        bin_counts = hist2.bin_counts()

        self.assertEqual(bin_counts, [(1, 10), (3, 3), (4, 2), (5, 1)])

def load_texture(image_file, repeat=False):
    """Carga una textura desde un archivo image_file"""
    img = Image.open(image_file)
    data = numpy.array(list(img.getdata()), numpy.int8)

    tex = glGenTextures(1)
    glPixelStorei(GL_UNPACK_ALIGNMENT, 1)
    glBindTexture(GL_TEXTURE_2D, tex)

    if repeat:
        glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_REPEAT)
        glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_REPEAT)
    else:
        glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_WRAP_S, GL_CLAMP)
        glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_WRAP_T, GL_CLAMP)

    glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_LINEAR)
    glTexParameterf(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_LINEAR)
    glTexImage2D(GL_TEXTURE_2D, 0, GL_RGB, img.size[0], img.size[1], 0, GL_RGB, GL_UNSIGNED_BYTE, data)
    glTexEnvf(GL_TEXTURE_ENV, GL_TEXTURE_ENV_MODE, GL_MODULATE)
    return tex

def wavWrite(y, fs, nbits, audioFile):
        """ Write samples to WAV file
        Args:
            samples: (ndarray / 2D ndarray) (floating point) sample vector
                        mono: DIM: nSamples
                        stereo: DIM: nSamples x nChannels

            fs:     (int) Sample rate in Hz
            nBits:  (int) Number of bits
            fnWAV:  (string) WAV file name to write
        """
        if nbits == 8:
            intsamples = (y+1.0) * AudioIO.normFact['int' + str(nbits)]
            fX = np.int8(intsamples)
        elif nbits == 16:
            intsamples = y * AudioIO.normFact['int' + str(nbits)]
            fX = np.int16(intsamples)
        elif nbits > 16:
            fX = y

        write(audioFile, fs, fX)

def test_copy_from(self):
        a = DynamicArray(0, numpy.dtype([
            ('a', numpy.int8),
            ('b', numpy.float32),
            ('c', numpy.float64),
        ]))
        a.extend([
            (4, 3.932, 902.345),
            (7, 1.016, 548.229),
            (2, 0.542, 771.031),
            (8, 5.429, 858.063),
        ])
        b = DynamicArray(0, numpy.dtype([
            ('a', numpy.int8),
            ('c', numpy.float64),
        ]))
        b.copy_from(a.data)
        self.assertEqual(len(b), 4)
        self.assertEqual(b.data.tolist(), [(4, 902.345), (7, 548.229), (2, 771.031), (8, 858.063),])

def test_resize(self):
        a = DynamicArray(0, numpy.dtype(numpy.int8))
        a.extend([0, 1, 4, 9])
        self.assertEqual(len(a), 4)
        self.assertEqual(len(a.data), 4)
        self.assertEqual(a.data.tolist(), [0, 1, 4, 9])

        a.resize(2)
        self.assertEqual(len(a), 2)
        self.assertEqual(len(a.data), 2)
        self.assertEqual(a.data.tolist(), [0, 1])

        request = a.capacity * 2
        a.resize(request)
        self.assertEqual(len(a), request)
        self.assertEqual(len(a.data), request)
        self.assertGreaterEqual(a.capacity, request)
        self.assertEqual(a.data.tolist()[:2], [0, 1])

def test_append(self):
        b = OrderedBuffer(3, numpy.dtype(numpy.int8))

        data = b.append(9)
        self.assertEqual(data, 9)
        self.assertIsInstance(data, numpy.int8)
        self.assertEqual(len(b), 1)
        self.assertEqual([chunk.tolist() for chunk in b.chunks], [[9]])

        data = b.append(1)
        self.assertIsInstance(data, numpy.int8)
        self.assertEqual(len(b), 2)
        self.assertEqual([chunk.tolist() for chunk in b.chunks], [[9, 1]])

        data = b.append(4)
        self.assertIsInstance(data, numpy.int8)
        self.assertEqual(len(b), 3)
        self.assertEqual([chunk.tolist() for chunk in b.chunks], [[9, 1, 4]])

        data = b.append(0)
        self.assertIsInstance(data, numpy.int8)
        self.assertEqual(len(b), 4)
        self.assertEqual([chunk.tolist() for chunk in b.chunks], [[9, 1, 4], [0]])

def test_clear(self):
        b = OrderedBuffer(3, numpy.dtype(numpy.int8))
        b.extend([9, 1])
        b.extend([1, 2, 3, 4, 5])
        b.append(4)
        self.assertEqual(len(b), 8)
        self.assertEqual([chunk.tolist() for chunk in b.chunks], [[9, 1], [1, 2, 3, 4, 5], [4]])

        b.clear()
        self.assertEqual(len(b), 0)
        self.assertEqual([chunk.tolist() for chunk in b.chunks], [])

        b.append(0)
        b.extend([7, 8, 9])
        b.append(1)
        b.extend([2])
        self.assertEqual(len(b), 6)
        self.assertEqual([chunk.tolist() for chunk in b.chunks], [[0], [7, 8, 9], [1, 2]])

def set_prior(self, bernoulli_prior=None, density_conditions=None):
        """
        density_conditions is the max no of ones in each dimension
        [min_row, min_col, max_row, max_col].
        zero means unrestricted
        """
        if density_conditions is None:
            self.density_conditions = np.array([0,0,0,0], dtype=np.int8)
        else:
            assert len(density_conditions) == 4
            self.density_conditions = np.array(density_conditions, dtype=np.int8)

        self.bernoulli_prior = bernoulli_prior
        if bernoulli_prior is None:
            self.logit_bernoulli_prior = 0
        else:
            self.logit_bernoulli_prior = np.log(bernoulli_prior/(1-bernoulli_prior))

def test_float(self):
        # offset for alignment test
        for i in range(4):
            assert_array_equal(self.f[i:] > 0, self.ef[i:])
            assert_array_equal(self.f[i:] - 1 >= 0, self.ef[i:])
            assert_array_equal(self.f[i:] == 0, ~self.ef[i:])
            assert_array_equal(-self.f[i:] < 0, self.ef[i:])
            assert_array_equal(-self.f[i:] + 1 <= 0, self.ef[i:])
            r = self.f[i:] != 0
            assert_array_equal(r, self.ef[i:])
            r2 = self.f[i:] != np.zeros_like(self.f[i:])
            r3 = 0 != self.f[i:]
            assert_array_equal(r, r2)
            assert_array_equal(r, r3)
            # check bool == 0x1
            assert_array_equal(r.view(np.int8), r.astype(np.int8))
            assert_array_equal(r2.view(np.int8), r2.astype(np.int8))
            assert_array_equal(r3.view(np.int8), r3.astype(np.int8))

            # isnan on amd64 takes the same code path
            assert_array_equal(np.isnan(self.nf[i:]), self.ef[i:])

def test_unaligned(self):
        v = (np.zeros(64, dtype=np.int8) + ord('a'))[1:-7]
        d = v.view(np.dtype("S8"))
        # unaligned source
        x = (np.zeros(16, dtype=np.int8) + ord('a'))[1:-7]
        x = x.view(np.dtype("S8"))
        x[...] = np.array("b" * 8, dtype="S")
        b = np.arange(d.size)
        #trivial
        assert_equal(d[b], d)
        d[b] = x
        # nontrivial
        # unaligned index array
        b = np.zeros(d.size + 1).view(np.int8)[1:-(np.intp(0).itemsize - 1)]
        b = b.view(np.intp)[:d.size]
        b[...] = np.arange(d.size)
        assert_equal(d[b.astype(np.int16)], d)
        d[b.astype(np.int16)] = x
        # boolean
        d[b % 2 == 0]
        d[b % 2 == 0] = x[::2]

def test_int(self):
        for st, ut, s in [(np.int8, np.uint8, 8),
                          (np.int16, np.uint16, 16),
                          (np.int32, np.uint32, 32),
                          (np.int64, np.uint64, 64)]:
            for i in range(1, s):
                assert_equal(hash(st(-2**i)), hash(-2**i),
                             err_msg="%r: -2**%d" % (st, i))
                assert_equal(hash(st(2**(i - 1))), hash(2**(i - 1)),
                             err_msg="%r: 2**%d" % (st, i - 1))
                assert_equal(hash(st(2**i - 1)), hash(2**i - 1),
                             err_msg="%r: 2**%d - 1" % (st, i))

                i = max(i - 1, 1)
                assert_equal(hash(ut(2**(i - 1))), hash(2**(i - 1)),
                             err_msg="%r: 2**%d" % (ut, i - 1))
                assert_equal(hash(ut(2**i - 1)), hash(2**i - 1),
                             err_msg="%r: 2**%d - 1" % (ut, i))

def test_big_binary(self):
        """Test workarounds for 32-bit limited fwrite, fseek, and ftell
        calls in windows. These normally would hang doing something like this.
        See http://projects.scipy.org/numpy/ticket/1660"""
        if sys.platform != 'win32':
            return
        try:
            # before workarounds, only up to 2**32-1 worked
            fourgbplus = 2**32 + 2**16
            testbytes = np.arange(8, dtype=np.int8)
            n = len(testbytes)
            flike = tempfile.NamedTemporaryFile()
            f = flike.file
            np.tile(testbytes, fourgbplus // testbytes.nbytes).tofile(f)
            flike.seek(0)
            a = np.fromfile(f, dtype=np.int8)
            flike.close()
            assert_(len(a) == fourgbplus)
            # check only start and end for speed:
            assert_((a[:n] == testbytes).all())
            assert_((a[-n:] == testbytes).all())
        except (MemoryError, ValueError):
            pass

def test_padded_struct_array(self):
        dt1 = np.dtype(
                [('a', 'b'), ('b', 'i'), ('sub', np.dtype('b,i')), ('c', 'i')],
                align=True)
        x1 = np.arange(dt1.itemsize, dtype=np.int8).view(dt1)
        self._check_roundtrip(x1)

        dt2 = np.dtype(
                [('a', 'b'), ('b', 'i'), ('c', 'b', (3,)), ('d', 'i')],
                align=True)
        x2 = np.arange(dt2.itemsize, dtype=np.int8).view(dt2)
        self._check_roundtrip(x2)

        dt3 = np.dtype(
                [('a', 'b'), ('b', 'i'), ('c', 'b'), ('d', 'b'),
                    ('e', 'b'), ('sub', np.dtype('b,i', align=True))])
        x3 = np.arange(dt3.itemsize, dtype=np.int8).view(dt3)
        self._check_roundtrip(x3)

def test_basic(self):
        ba = [1, 2, 10, 11, 6, 5, 4]
        ba2 = [[1, 2, 3, 4], [5, 6, 7, 9], [10, 3, 4, 5]]
        for ctype in [np.int8, np.uint8, np.int16, np.uint16, np.int32,
                      np.uint32, np.float32, np.float64, np.complex64, np.complex128]:
            a = np.array(ba, ctype)
            a2 = np.array(ba2, ctype)

            tgt = np.array([1, 3, 13, 24, 30, 35, 39], ctype)
            assert_array_equal(np.cumsum(a, axis=0), tgt)

            tgt = np.array(
                [[1, 2, 3, 4], [6, 8, 10, 13], [16, 11, 14, 18]], ctype)
            assert_array_equal(np.cumsum(a2, axis=0), tgt)

            tgt = np.array(
                [[1, 3, 6, 10], [5, 11, 18, 27], [10, 13, 17, 22]], ctype)
            assert_array_equal(np.cumsum(a2, axis=1), tgt)

def test_dtypes(self):
        c = array([11, -12, 13], dtype=np.int8)
        v = vander(c)
        expected = np.array([[121, 11, 1],
                             [144, -12, 1],
                             [169, 13, 1]])
        yield (assert_array_equal, v, expected)

        c = array([1.0+1j, 1.0-1j])
        v = vander(c, N=3)
        expected = np.array([[2j, 1+1j, 1],
                             [-2j, 1-1j, 1]])
        # The data is floating point, but the values are small integers,
        # so assert_array_equal *should* be safe here (rather than, say,
        # assert_array_almost_equal).
        yield (assert_array_equal, v, expected)

def test_shuffle(self):
        # Test lists, arrays (of various dtypes), and multidimensional versions
        # of both, c-contiguous or not:
        for conv in [lambda x: np.array([]),
                     lambda x: x,
                     lambda x: np.asarray(x).astype(np.int8),
                     lambda x: np.asarray(x).astype(np.float32),
                     lambda x: np.asarray(x).astype(np.complex64),
                     lambda x: np.asarray(x).astype(object),
                     lambda x: [(i, i) for i in x],
                     lambda x: np.asarray([[i, i] for i in x]),
                     lambda x: np.vstack([x, x]).T,
                     # gh-4270
                     lambda x: np.asarray([(i, i) for i in x],
                                          [("a", object, 1),
                                           ("b", np.int32, 1)])]:
            np.random.seed(self.seed)
            alist = conv([1, 2, 3, 4, 5, 6, 7, 8, 9, 0])
            np.random.shuffle(alist)
            actual = alist
            desired = conv([0, 1, 9, 6, 2, 4, 5, 8, 7, 3])
            np.testing.assert_array_equal(actual, desired)

def __init__(self, model_name, saved_model_file=None, max_training_batches=1, watch=False):
        self.saved_model_file = saved_model_file
        if saved_model_file is not None:
            print('Loading saved model from %s' % saved_model_file)
            self.model = self.load_model(saved_model_file)
        else:
            self.init_model(model_name)
        # Treat as a ring buffer
        self.current_pos = 0
        self.max_pos = 0
        self.states_t0 = np.zeros((BUFFER_SIZE,1,BOARD_HEIGHT,BOARD_WIDTH), dtype=np.int8)
        self.actions = np.zeros([BUFFER_SIZE], dtype=np.int8)
        self.states_t1 = np.zeros((BUFFER_SIZE,1,BOARD_HEIGHT,BOARD_WIDTH), dtype=np.int8)
        self.rewards = np.zeros([BUFFER_SIZE], dtype=np.float32)
        self.n_games = 0
        self.state_printer = WebSocketPrinter()
        self.current_game_length = 0
        self.current_episode_length = 0
        self.n_games = 0
        self.max_training_batches = max_training_batches
        self.n_training_batches = 0
        self.model_name = model_name

def test_load_columnar_pandas_all(self, con, all_types_table):
        pd = pytest.importorskip("pandas")
        import numpy as np

        data = pd.DataFrame({
            "boolean_": [True, False],
            "smallint_": np.array([0, 1], dtype=np.int8),
            "int_": np.array([0, 1], dtype=np.int32),
            "bigint_": np.array([0, 1], dtype=np.int64),
            "float_": np.array([0, 1], dtype=np.float32),
            "double_": np.array([0, 1], dtype=np.float64),
            "varchar_": ["a", "b"],
            "text_": ['a', 'b'],
            "time_": [datetime.time(0, 11, 59), datetime.time(13)],
            "timestamp_": [pd.Timestamp("2016"), pd.Timestamp("2017")],
            "date_": [datetime.date(2016, 1, 1), datetime.date(2017, 1, 1)],
        }, columns=['boolean_', 'smallint_', 'int_', 'bigint_', 'float_',
                    'double_', 'varchar_', 'text_', 'time_', 'timestamp_',
                    'date_'])
        con.load_table_columnar(all_types_table, data, preserve_index=False)

def test_load_table_creates(self, con, not_a_table):
        pd = pytest.importorskip("pandas")
        import numpy as np

        data = pd.DataFrame({
            "boolean_": [True, False],
            "smallint_cast": np.array([0, 1], dtype=np.int8),
            "smallint_": np.array([0, 1], dtype=np.int16),
            "int_": np.array([0, 1], dtype=np.int32),
            "bigint_": np.array([0, 1], dtype=np.int64),
            "float_": np.array([0, 1], dtype=np.float32),
            "double_": np.array([0, 1], dtype=np.float64),
            "varchar_": ["a", "b"],
            "text_": ['a', 'b'],
            "time_": [datetime.time(0, 11, 59), datetime.time(13)],
            "timestamp_": [pd.Timestamp("2016"), pd.Timestamp("2017")],
            "date_": [datetime.date(2016, 1, 1), datetime.date(2017, 1, 1)],
        }, columns=['boolean_', 'smallint_', 'int_', 'bigint_', 'float_',
                    'double_', 'varchar_', 'text_', 'time_', 'timestamp_',
                    'date_'])
        con.load_table(not_a_table, data, create=True)

def wavWrite(y, fs, nbits, audioFile):
        """ Write samples to WAV file
        Args:
            samples: (ndarray / 2D ndarray) (floating point) sample vector
                        mono: DIM: nSamples
                        stereo: DIM: nSamples x nChannels

            fs:     (int) Sample rate in Hz
            nBits:  (int) Number of bits
            fnWAV:  (string) WAV file name to write
        """
        if nbits == 8:
            intsamples = (y+1.0) * AudioIO.normFact['int' + str(nbits)]
            fX = np.int8(intsamples)
        elif nbits == 16:
            intsamples = y * AudioIO.normFact['int' + str(nbits)]
            fX = np.int16(intsamples)
        elif nbits > 16:
            fX = y

        write(audioFile, fs, fX)

def one_hot_comparison(hot_axes, axes, C):
    """
    TODO.

    Arguments:
      hot_axes: TODO
      axes: TODO
    """
    u = rng.random_integers(0, C.length - 1, axes, dtype=np.int8)
    u_p = ng.placeholder(axes, dtype=u.dtype)
    v = np.zeros(hot_axes.lengths, dtype=np.float32)
    udxiter = np.nditer(u, flags=['multi_index'])
    for uiter in udxiter:
        vindex = [int(uiter)]
        vindex.extend(udxiter.multi_index)
        v[tuple(vindex)] = 1

    with executor(ng.one_hot(u_p, axis=C), u_p) as ex:
        v_t = ex(u)
        ng.testing.assert_allclose(v_t, v)

def interpret_header(self):
        """redefine variables from header dictionary"""
        self.nifs = self.header['nifs']
        self.nchans = self.header['nchans']
        self.nbits = self.header['nbits']
        signed = 'signed' in self.header and self.header['signed'] is True
        if self.nbits >= 8:
            if signed:
                self.dtype = {8: np.int8,
                              16: np.int16,
                              32: np.float32,
                              64: np.float64}[self.nbits]
            else:
                self.dtype = {8: np.uint8,
                              16: np.uint16,
                              32: np.float32,
                              64: np.float64}[self.nbits]
        else:
            self.dtype = np.int8 if signed else np.uint8

def numpy2bifrost(dtype):
    if   dtype == np.int8:       return _bf.BF_DTYPE_I8
    elif dtype == np.int16:      return _bf.BF_DTYPE_I16
    elif dtype == np.int32:      return _bf.BF_DTYPE_I32
    elif dtype == np.uint8:      return _bf.BF_DTYPE_U8
    elif dtype == np.uint16:     return _bf.BF_DTYPE_U16
    elif dtype == np.uint32:     return _bf.BF_DTYPE_U32
    elif dtype == np.float16:    return _bf.BF_DTYPE_F16
    elif dtype == np.float32:    return _bf.BF_DTYPE_F32
    elif dtype == np.float64:    return _bf.BF_DTYPE_F64
    elif dtype == np.float128:   return _bf.BF_DTYPE_F128
    elif dtype == ci8:           return _bf.BF_DTYPE_CI8
    elif dtype == ci16:          return _bf.BF_DTYPE_CI16
    elif dtype == ci32:          return _bf.BF_DTYPE_CI32
    elif dtype == cf16:          return _bf.BF_DTYPE_CF16
    elif dtype == np.complex64:  return _bf.BF_DTYPE_CF32
    elif dtype == np.complex128: return _bf.BF_DTYPE_CF64
    elif dtype == np.complex256: return _bf.BF_DTYPE_CF128
    else: raise ValueError("Unsupported dtype: " + str(dtype))

def numpy2string(dtype):
    if   dtype == np.int8:       return 'i8'
    elif dtype == np.int16:      return 'i16'
    elif dtype == np.int32:      return 'i32'
    elif dtype == np.int64:      return 'i64'
    elif dtype == np.uint8:      return 'u8'
    elif dtype == np.uint16:     return 'u16'
    elif dtype == np.uint32:     return 'u32'
    elif dtype == np.uint64:     return 'u64'
    elif dtype == np.float16:    return 'f16'
    elif dtype == np.float32:    return 'f32'
    elif dtype == np.float64:    return 'f64'
    elif dtype == np.float128:   return 'f128'
    elif dtype == np.complex64:  return 'cf32'
    elif dtype == np.complex128: return 'cf64'
    elif dtype == np.complex256: return 'cf128'
    else: raise TypeError("Unsupported dtype: " + str(dtype))

def run_test_matmul_aa_ci8_shape(self, shape, transpose=False):
        # **TODO: This currently never triggers the transpose path in the backend
        shape_complex = shape[:-1] + (shape[-1] * 2,)
        # Note: The xGPU-like correlation kernel does not support input values of -128 (only [-127:127])
        a8 = ((np.random.random(size=shape_complex) * 2 - 1) * 127).astype(np.int8)
        a_gold = a8.astype(np.float32).view(np.complex64)
        if transpose:
            a_gold = H(a_gold)
        # Note: np.matmul seems to be slow and inaccurate when there are batch dims
        c_gold = np.matmul(a_gold, H(a_gold))
        triu = np.triu_indices(shape[-2] if not transpose else shape[-1], 1)
        c_gold[..., triu[0], triu[1]] = 0
        a = a8.view(bf.DataType.ci8)
        a = bf.asarray(a, space='cuda')
        if transpose:
            a = H(a)
        c = bf.zeros_like(c_gold, space='cuda')
        self.linalg.matmul(1, a, None, 0, c)
        c = c.copy('system')
        np.testing.assert_allclose(c, c_gold, RTOL, ATOL)

def run_benchmark_matmul_aa_correlator_kernel(self, ntime, nstand, nchan):
        x_shape = (ntime, nchan, nstand*2)
        perm = [1,0,2]
        x8 = ((np.random.random(size=x_shape+(2,))*2-1)*127).astype(np.int8)
        x = x8.astype(np.float32).view(np.complex64).reshape(x_shape)
        x = x.transpose(perm)
        b_gold = np.matmul(H(x[:,[0],:]), x[:,[0],:])
        triu = np.triu_indices(x_shape[-1], 1)
        b_gold[..., triu[0], triu[1]] = 0
        x = x8.view(bf.DataType.ci8).reshape(x_shape)
        x = bf.asarray(x, space='cuda')
        x = x.transpose(perm)
        b = bf.zeros_like(b_gold, space='cuda')
        bf.device.stream_synchronize();
        t0 = time.time()
        nrep = 200
        for _ in xrange(nrep):
            self.linalg.matmul(1, None, x, 0, b)
        bf.device.stream_synchronize();
        dt = time.time() - t0
        nflop = nrep * nchan * ntime * nstand*(nstand+1)/2 * 2*2 * 8
        print nstand, '\t', nflop / dt / 1e9, 'GFLOP/s'
        print '\t\t', nrep*ntime*nchan / dt / 1e6, 'MHz'

def perform(self, node, inputs_storage, output_storage):
        """Peform the transformation from output to feature space.

        Defines the Python implementation of the op. It is in charge of doing 
        the processing to go from output space (statematrix) to feature space.

        Parameters
        ----------
        node : 
            Reference to an Apply node which was previously obtained via 
            the Op‘s make_node() method.
        inputs_storage : array_like
            A list of references to data which can be operated on using 
            non-symbolic statements
        output_storage : array_like
            A list of storage cells where the output is to be stored
        """
        state, time = inputs_storage
        output_storage[0][0] = np.array(self.d.f.note_state_single_to_input_form(state, time), dtype='int8')

def load_board(string):
    reverse_map = {
        'X': go.BLACK,
        'O': go.WHITE,
        '.': go.EMPTY,
        '#': go.FILL,
        '*': go.KO,
        '?': go.UNKNOWN
    }

    string = re.sub(r'[^XO\.#]+', '', string)
    assert len(string) == go.N ** 2, "Board to load didn't have right dimensions"
    board = np.zeros([go.N, go.N], dtype=np.int8)
    for i, char in enumerate(string):
        np.ravel(board)[i] = reverse_map[char]
    return board