Merge pull request dmlc#165 from viktor-ferenczi/viktor-ferenczi

Python wrapper fixes

wrapper/cxxnet.py

@@ -12,9 +12,9 @@
 # set this line correctly
 if os.name == 'nt':
  # TODO windows
- CXXNET_PATH = os.path.dirname(__file__) + '/libcxxnetwrapper.dll'
+ CXXNET_PATH = os.path.join(os.path.dirname(__file__), 'libcxxnetwrapper.dll')
 else:
- CXXNET_PATH = os.path.dirname(__file__) + '/libcxxnetwrapper.so'
+ CXXNET_PATH = os.path.join(os.path.dirname(__file__), 'libcxxnetwrapper.so')
 
 # load in xgboost library
 cxnlib = ctypes.cdll.LoadLibrary(CXXNET_PATH)
@@ -35,7 +35,8 @@ def ctypes2numpy(cptr, length, dtype=numpy.float32):
  """convert a ctypes pointer array to numpy array """
  #assert isinstance(cptr, ctypes.POINTER(ctypes.c_float))
  res = numpy.zeros(length, dtype=dtype)
- assert ctypes.memmove(res.ctypes.data, cptr, length * res.strides[0])
+ if not ctypes.memmove(res.ctypes.data, cptr, length * res.strides[0]):
+ raise AssertionError('ctypes.memmove failed')
  return res
 
 def ctypes2numpyT(cptr, shape, dtype=numpy.float32, stride = None):
@@ -45,11 +46,13 @@ def ctypes2numpyT(cptr, shape, dtype=numpy.float32, stride = None):
  size *= x
  if stride is None:
  res = numpy.zeros(size, dtype=dtype)
- assert ctypes.memmove(res.ctypes.data, cptr, size * res.strides[0])
+ if not ctypes.memmove(res.ctypes.data, cptr, size * res.strides[0]):
+ raise AssertionError('ctypes.memmove failed')
  else:
  dsize = size / shape[-1] * stride
  res = numpy.zeros(dsize, dtype=dtype)
- assert ctypes.memmove(res.ctypes.data, cptr, dsize * res.strides[0])
+ if not ctypes.memmove(res.ctypes.data, cptr, dsize * res.strides[0]):
+ raise AssertionError('ctypes.memmove failed')
  res = res.reshape((dsize / shape[-1], shape[-1]))
  res = res[:, 0 :shape[-1]]
  return res.reshape(shape)
@@ -166,11 +169,15 @@ def update(self, data, label = None):
  if not isinstance(label, numpy.ndarray):
  raise Exception('Net.update: label need to be ndarray')
  if label.ndim == 1:
- label = label.reshape((label.size(0),1))
+ label = label.reshape(label.shape[0], 1)
  if label.ndim != 2:
  raise Exception('Net.update: label need to be 2 dimension or one dimension ndarray')
  if label.shape[0] != data.shape[0]:
  raise Exception('Net.update: data size mismatch')
+ if data.dtype != numpy.float32:
+ raise Exception('Net.update: data must be of type numpy.float32')
+ if label.dtype != numpy.float32:
+ raise Exception('Net.update: label must be of type numpy.float32')
  cxnlib.CXNNetUpdateBatch(self.handle,
  data.ctypes.data_as(ctypes.POINTER(ctypes.c_float)),
  shape2ctypes(data),
@@ -278,7 +285,7 @@ def get_weight(self, layer_name, tag):
  return None
  return ctypes2numpyT(ret, [oshape[i] for i in range(odim.value)], 'float32')
 
-def train(cfg, data, num_round, param, eval_data = None):
+def train(cfg, data, label, num_round, param, eval_data = None):
  net = Net(cfg = cfg)
  if isinstance(param, dict):
  param = param.items()