/usr/share/pyshared/statsmodels/base/data.py is in python-statsmodels 0.4.2-1.2.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 | """
Base tools for handling various kinds of data structures, attaching metadata to
results, and doing data cleaning
"""
import numpy as np
from pandas import DataFrame, Series, TimeSeries
from statsmodels.tools.decorators import (resettable_cache,
cache_readonly, cache_writable)
import statsmodels.tools.data as data_util
class ModelData(object):
"""
Class responsible for handling input data and extracting metadata into the
appropriate form
"""
def __init__(self, endog, exog=None, **kwds):
self._orig_endog = endog
self._orig_exog = exog
self.endog, self.exog = self._convert_endog_exog(endog, exog)
self._check_integrity()
self._cache = resettable_cache()
def _convert_endog_exog(self, endog, exog):
# for consistent outputs if endog is (n,1)
yarr = self._get_yarr(endog)
xarr = None
if exog is not None:
xarr = self._get_xarr(exog)
if xarr.ndim == 1:
xarr = xarr[:, None]
if xarr.ndim != 2:
raise ValueError("exog is not 1d or 2d")
return yarr, xarr
@cache_writable()
def ynames(self):
endog = self._orig_endog
ynames = self._get_names(endog)
if not ynames:
ynames = _make_endog_names(self.endog)
if len(ynames) == 1:
return ynames[0]
else:
return list(ynames)
@cache_writable()
def xnames(self):
exog = self._orig_exog
if exog is not None:
xnames = self._get_names(exog)
if not xnames:
xnames = _make_exog_names(self.exog)
return list(xnames)
return None
@cache_readonly
def row_labels(self):
exog = self._orig_exog
if exog is not None:
row_labels = self._get_row_labels(exog)
else:
endog = self._orig_endog
row_labels = self._get_row_labels(endog)
return row_labels
def _get_row_labels(self, arr):
return None
def _get_names(self, arr):
if isinstance(arr, DataFrame):
return list(arr.columns)
elif isinstance(arr, Series):
if arr.name:
return [arr.name]
else:
return
else:
try:
return arr.dtype.names
except AttributeError:
pass
return None
def _get_yarr(self, endog):
if data_util._is_structured_ndarray(endog):
endog = data_util.struct_to_ndarray(endog)
return np.asarray(endog).squeeze()
def _get_xarr(self, exog):
if data_util._is_structured_ndarray(exog):
exog = data_util.struct_to_ndarray(exog)
return np.asarray(exog)
def _check_integrity(self):
if self.exog is not None:
if len(self.exog) != len(self.endog):
raise ValueError("endog and exog matrices are different sizes")
def wrap_output(self, obj, how='columns'):
if how == 'columns':
return self.attach_columns(obj)
elif how == 'rows':
return self.attach_rows(obj)
elif how == 'cov':
return self.attach_cov(obj)
elif how == 'dates':
return self.attach_dates(obj)
elif how == 'columns_eq':
return self.attach_columns_eq(obj)
elif how == 'cov_eq':
return self.attach_cov_eq(obj)
else:
return obj
def attach_columns(self, result):
return result
def attach_columns_eq(self, result):
return result
def attach_cov(self, result):
return result
def attach_cov_eq(self, result):
return result
def attach_rows(self, result):
return result
def attach_dates(self, result):
return result
class PandasData(ModelData):
"""
Data handling class which knows how to reattach pandas metadata to model
results
"""
def _check_integrity(self):
try:
endog, exog = self._orig_endog, self._orig_exog
# exog can be None and we could be upcasting one or the other
if exog is not None and (hasattr(endog, 'index') and
hasattr(exog, 'index')):
assert self._orig_endog.index.equals(self._orig_exog.index)
except AssertionError:
raise ValueError("The indices for endog and exog are not aligned")
super(PandasData, self)._check_integrity()
def _get_row_labels(self, arr):
try:
return arr.index
except AttributeError, err:
# if we've gotten here it's because endog is pandas and
# exog is not, so just return the row labels from endog
return self._orig_endog.index
def attach_columns(self, result):
if result.squeeze().ndim <= 1:
return Series(result, index=self.xnames)
else: # for e.g., confidence intervals
return DataFrame(result, index=self.xnames)
def attach_columns_eq(self, result):
return DataFrame(result, index=self.xnames, columns=self.ynames)
def attach_cov(self, result):
return DataFrame(result, index=self.xnames, columns=self.xnames)
def attach_cov_eq(self, result):
return DataFrame(result, index=self.ynames, columns=self.ynames)
def attach_rows(self, result):
# assumes if len(row_labels) > len(result) it's bc it was truncated
# at the front, for AR lags, for example
if result.squeeze().ndim == 1:
return Series(result, index=self.row_labels[-len(result):])
else: # this is for VAR results, may not be general enough
return DataFrame(result, index=self.row_labels[-len(result):],
columns=self.ynames)
def attach_dates(self, result):
return TimeSeries(result, index=self.predict_dates)
class TimeSeriesData(ModelData):
"""
Data handling class which returns scikits.timeseries model results
"""
def _get_row_labels(self, arr):
return arr.dates
#def attach_columns(self, result):
# return recarray?
#def attach_cov(self, result):
# return recarray?
def attach_rows(self, result):
from scikits.timeseries import time_series
return time_series(result, dates = self.row_labels[-len(result):])
def attach_dates(self, result):
from scikits.timeseries import time_series
return time_series(result, dates = self.predict_dates)
_la = None
def _lazy_import_larry():
global _la
import la
_la = la
class LarryData(ModelData):
"""
Data handling class which knows how to reattach pandas metadata to model
results
"""
def __init__(self, endog, exog=None, **kwds):
_lazy_import_larry()
super(LarryData, self).__init__(endog, exog=exog, **kwds)
def _get_yarr(self, endog):
try:
return endog.x
except AttributeError:
return np.asarray(endog).squeeze()
def _get_xarr(self, exog):
try:
return exog.x
except AttributeError:
return np.asarray(exog)
def _get_names(self, exog):
try:
return exog.label[1]
except Exception:
pass
return None
def _get_row_labels(self, arr):
return arr.label[0]
def attach_columns(self, result):
if result.ndim == 1:
return _la.larry(result, [self.xnames])
else:
shape = results.shape
return _la.larray(result, [self.xnames, range(shape[1])])
def attach_columns_eq(self, result):
return _la.larray(result, [self.xnames], [self.xnames])
def attach_cov(self, result):
return _la.larry(result, [self.xnames], [self.xnames])
def attach_cov_eq(self, result):
return _la.larray(result, [self.ynames], [self.ynames])
def attach_rows(self, result):
return _la.larry(result, [self.row_labels[-len(result):]])
def attach_dates(self, result):
return _la.larray(result, [self.predict_dates])
def _make_endog_names(endog):
if endog.ndim == 1 or endog.shape[1] == 1:
ynames = ['y']
else: # for VAR
ynames = ['y%d' % (i+1) for i in range(endog.shape[1])]
return ynames
def _make_exog_names(exog):
exog_var = exog.var(0)
if (exog_var == 0).any():
# assumes one constant in first or last position
# avoid exception if more than one constant
const_idx = exog_var.argmin()
exog_names = ['x%d' % i for i in range(1,exog.shape[1])]
exog_names.insert(const_idx, 'const')
else:
exog_names = ['x%d' % i for i in range(1,exog.shape[1]+1)]
return exog_names
def handle_data(endog, exog):
"""
Given inputs
"""
# deal with lists and tuples up-front
if isinstance(endog, (list, tuple)):
endog = np.asarray(endog)
if isinstance(exog, (list, tuple)):
exog = np.asarray(exog)
if data_util._is_using_pandas(endog, exog):
klass = PandasData
elif data_util._is_using_larry(endog, exog):
klass = LarryData
elif data_util._is_using_timeseries(endog, exog):
klass = TimeSeriesData
# keep this check last
elif data_util._is_using_ndarray(endog, exog):
klass = ModelData
else:
raise ValueError('unrecognized data structures: %s / %s' %
(type(endog), type(exog)))
return klass(endog, exog=exog)
|