Hendrik/xgboost
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Merge pull request #488 from sinhrks/pyfeaturenames

Browse Source 
Support feature names in Python package
This commit is contained in:
Tianqi Chen
2015-09-15 09:56:55 -07:00
parent bda3282f6d 6063d243eb
commit ae43fd7c7a
6 changed files with 214 additions and 43 deletions
Download Patch File Download Diff File Expand all files Collapse all files

112
python-package/xgboost/core.py
View File

@@ -1,5 +1,5 @@
# coding: utf-8
# pylint: disable=too-many-arguments
# pylint: disable=too-many-arguments, too-many-branches
"""Core XGBoost Library."""
from __future__ import absolute_import
@@ -23,8 +23,9 @@ class XGBoostError(Exception):
if sys.version_info[0] == 3:
# pylint: disable=invalid-name
# pylint: disable=invalid-name, redefined-builtin
STRING_TYPES = str,
unicode = str
else:
# pylint: disable=invalid-name
STRING_TYPES = basestring,
@@ -131,7 +132,11 @@ class DMatrix(object):
which is optimized for both memory efficiency and training speed.
You can construct DMatrix from numpy.arrays
"""
def __init__(self, data, label=None, missing=0.0, weight=None, silent=False):
feature_names = None # for previous version's pickle
def __init__(self, data, label=None, missing=0.0,
weight=None, silent=False, feature_names=None):
"""
Data matrix used in XGBoost.
@@ -149,6 +154,8 @@ class DMatrix(object):
Weight for each instance.
silent : boolean, optional
Whether print messages during construction
feature_names : list, optional
Labels for features.
"""
# force into void_p, mac need to pass things in as void_p
if data is None:
@@ -176,6 +183,21 @@ class DMatrix(object):
if weight is not None:
self.set_weight(weight)
# validate feature name
if not feature_names is None:
if not isinstance(feature_names, list):
feature_names = list(feature_names)
if len(feature_names) != len(set(feature_names)):
raise ValueError('feature_names must be unique')
if len(feature_names) != self.num_col():
msg = 'feature_names must have the same length as data'
raise ValueError(msg)
# prohibit to use symbols may affect to parse. e.g. ``[]=.``
if not all(isinstance(f, STRING_TYPES) and f.isalnum()
for f in feature_names):
raise ValueError('all feature_names must be alphanumerics')
self.feature_names = feature_names
def _init_from_csr(self, csr):
"""
Initialize data from a CSR matrix.
@@ -391,6 +413,18 @@ class DMatrix(object):
ctypes.byref(ret)))
return ret.value
def num_col(self):
"""Get the number of columns (features) in the DMatrix.
Returns
-------
number of columns : int
"""
ret = ctypes.c_uint()
_check_call(_LIB.XGDMatrixNumCol(self.handle,
ctypes.byref(ret)))
return ret.value
def slice(self, rindex):
"""Slice the DMatrix and return a new DMatrix that only contains `rindex`.
@@ -404,7 +438,7 @@ class DMatrix(object):
res : DMatrix
A new DMatrix containing only selected indices.
"""
res = DMatrix(None)
res = DMatrix(None, feature_names=self.feature_names)
res.handle = ctypes.c_void_p()
_check_call(_LIB.XGDMatrixSliceDMatrix(self.handle,
c_array(ctypes.c_int, rindex),
@@ -419,6 +453,9 @@ class Booster(object):
Booster is the model of xgboost, that contains low level routines for
training, prediction and evaluation.
"""
feature_names = None
def __init__(self, params=None, cache=(), model_file=None):
# pylint: disable=invalid-name
"""Initialize the Booster.
@@ -435,6 +472,7 @@ class Booster(object):
for d in cache:
if not isinstance(d, DMatrix):
raise TypeError('invalid cache item: {}'.format(type(d).__name__))
self._validate_feature_names(d)
dmats = c_array(ctypes.c_void_p, [d.handle for d in cache])
self.handle = ctypes.c_void_p()
_check_call(_LIB.XGBoosterCreate(dmats, len(cache), ctypes.byref(self.handle)))
@@ -519,6 +557,8 @@ class Booster(object):
"""
if not isinstance(dtrain, DMatrix):
raise TypeError('invalid training matrix: {}'.format(type(dtrain).__name__))
self._validate_feature_names(dtrain)
if fobj is None:
_check_call(_LIB.XGBoosterUpdateOneIter(self.handle, iteration, dtrain.handle))
else:
@@ -543,6 +583,8 @@ class Booster(object):
raise ValueError('grad / hess length mismatch: {} / {}'.format(len(grad), len(hess)))
if not isinstance(dtrain, DMatrix):
raise TypeError('invalid training matrix: {}'.format(type(dtrain).__name__))
self._validate_feature_names(dtrain)
_check_call(_LIB.XGBoosterBoostOneIter(self.handle, dtrain.handle,
c_array(ctypes.c_float, grad),
c_array(ctypes.c_float, hess),
@@ -572,6 +614,8 @@ class Booster(object):
raise TypeError('expected DMatrix, got {}'.format(type(d[0]).__name__))
if not isinstance(d[1], STRING_TYPES):
raise TypeError('expected string, got {}'.format(type(d[1]).__name__))
self._validate_feature_names(d[0])
dmats = c_array(ctypes.c_void_p, [d[0].handle for d in evals])
evnames = c_array(ctypes.c_char_p, [c_str(d[1]) for d in evals])
msg = ctypes.c_char_p()
@@ -605,6 +649,7 @@ class Booster(object):
result: str
Evaluation result string.
"""
self._validate_feature_names(data)
return self.eval_set([(data, name)], iteration)
def predict(self, data, output_margin=False, ntree_limit=0, pred_leaf=False):
@@ -642,6 +687,9 @@ class Booster(object):
option_mask |= 0x01
if pred_leaf:
option_mask |= 0x02
self._validate_feature_names(data)
length = ctypes.c_ulong()
preds = ctypes.POINTER(ctypes.c_float)()
_check_call(_LIB.XGBoosterPredict(self.handle, data.handle,
@@ -731,16 +779,46 @@ class Booster(object):
"""
Returns the dump the model as a list of strings.
"""
length = ctypes.c_ulong()
sarr = ctypes.POINTER(ctypes.c_char_p)()
_check_call(_LIB.XGBoosterDumpModel(self.handle,
c_str(fmap),
int(with_stats),
ctypes.byref(length),
ctypes.byref(sarr)))
if self.feature_names is not None and fmap == '':
flen = int(len(self.feature_names))
fname = (ctypes.c_char_p * flen)()
ftype = (ctypes.c_char_p * flen)()
# supports quantitative type only
# {'q': quantitative, 'i': indicator}
if sys.version_info[0] == 3:
features = [bytes(f, 'utf-8') for f in self.feature_names]
types = [bytes('q', 'utf-8')] * flen
else:
features = [f.encode('utf-8') if isinstance(f, unicode) else f
for f in self.feature_names]
types = ['q'] * flen
fname[:] = features
ftype[:] = types
_check_call(_LIB.XGBoosterDumpModelWithFeatures(self.handle,
flen,
fname,
ftype,
int(with_stats),
ctypes.byref(length),
ctypes.byref(sarr)))
else:
_check_call(_LIB.XGBoosterDumpModel(self.handle,
c_str(fmap),
int(with_stats),
ctypes.byref(length),
ctypes.byref(sarr)))
res = []
for i in range(length.value):
res.append(str(sarr[i].decode('ascii')))
try:
res.append(str(sarr[i].decode('ascii')))
except UnicodeDecodeError:
res.append(unicode(sarr[i].decode('utf-8')))
return res
def get_fscore(self, fmap=''):
@@ -765,3 +843,17 @@ class Booster(object):
else:
fmap[fid] += 1
return fmap
def _validate_feature_names(self, data):
"""
Validate Booster and data's feature_names are identical
"""
if self.feature_names is None:
self.feature_names = data.feature_names
else:
# Booster can't accept data with different feature names
if self.feature_names != data.feature_names:
msg = 'feature_names mismatch: {0} {1}'
raise ValueError(msg.format(self.feature_names,
data.feature_names))