load.py 8.37 KB
Newer Older
André Anjos's avatar
André Anjos committed
1 2 3 4 5 6 7 8 9
#!/usr/bin/env python
# vim: set fileencoding=utf-8 :
# Andre Anjos <andre.anjos@idiap.ch>
# Mon 23 May 2011 16:23:05 CEST

"""A set of utilities to load score files with different formats.
"""

import numpy
10 11 12 13 14 15 16 17
import tarfile
import os

def open_file(filename):
  """Opens the given score file for reading.
  Score files might be raw text files, or a tar-file including a single score file inside.

  Parameters:
18 19 20 21

  filename : str or file-like
    The name of the score file to open, or a file-like object open for reading.
    If a file name is given, the according file might be a raw text file or a (compressed) tar file containing a raw text file.
22 23 24 25

  Returns:
    A read-only file-like object as it would be returned by open().
  """
26 27 28 29
  if not isinstance(filename, str) and hasattr(filename, 'read'):
    # It seems that this is an open file
    return filename

30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47
  if not os.path.isfile(filename):
    raise IOError("Score file '%s' does not exist." % filename)
  if not tarfile.is_tarfile(filename):
    return open(filename, 'rt')

  # open the tar file for reading
  tar = tarfile.open(filename, 'r')
  # get the first file in the tar file
  tar_info = tar.next()
  while tar_info is not None and not tar_info.isfile():
    tar_info = tar.next()
  # check that one file was found in the archive
  if tar_info is None:
    raise IOError("The given file is a .tar file, but it does not contain any file.")

  # open the file for reading
  return tar.extractfile(tar_info)

André Anjos's avatar
André Anjos committed
48 49 50 51 52 53

def four_column(filename):
  """Loads a score set from a single file to memory.

  Verifies that all fields are correctly placed and contain valid fields.

54
  Returns a python generator of tuples containing the following fields:
André Anjos's avatar
André Anjos committed
55 56 57 58 59 60 61 62 63 64 65

    [0]
      claimed identity (string)
    [1]
      real identity (string)
    [2]
      test label (string)
    [3]
      score (float)
  """

66 67
  for i, l in enumerate(open_file(filename)):
    if isinstance(l, bytes): l = l.decode('utf-8')
André Anjos's avatar
André Anjos committed
68 69 70 71 72 73 74 75 76
    s = l.strip()
    if len(s) == 0 or s[0] == '#': continue #empty or comment
    field = [k.strip() for k in s.split()]
    if len(field) < 4:
      raise SyntaxError('Line %d of file "%s" is invalid: %s' % (i, filename, l))
    try:
      score = float(field[3])
    except:
      raise SyntaxError('Cannot convert score to float at line %d of file "%s": %s' % (i, filename, l))
77
    yield (field[0], field[1], field[2], score)
André Anjos's avatar
André Anjos committed
78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94


def split_four_column(filename):
  """Loads a score set from a single file to memory and splits the scores
  between positives and negatives. The score file has to respect the 4 column
  format as defined in the method four_column().

  This method avoids loading and allocating memory for the strings present in
  the file. We only keep the scores.

  Returns a python tuple (negatives, positives). The values are 1-D blitz
  arrays of float64.
  """

  # split in positives and negatives
  neg = []
  pos = []
95 96 97 98 99 100
  # read four column list line by line
  for (client_id, probe_id, _, score) in four_column(filename):
    if client_id == probe_id:
      pos.append(score)
    else:
      neg.append(score)
André Anjos's avatar
André Anjos committed
101 102 103 104 105

  return (numpy.array(neg, numpy.float64), numpy.array(pos, numpy.float64))

def cmc_four_column(filename):
  """Loads scores to compute CMC curves from a file in four column format.
106 107 108 109 110 111
  The four column file needs to be in the same format as described in the four_column function,
  and the "test label" (column 3) has to contain the test/probe file name.

  This function returns a list of tuples.
  For each probe file, the tuple consists of a list of negative scores and a list of positive scores.
  Usually, the list of positive scores should contain only one element, but more are allowed.
112

113 114
  The result of this function can directly be passed to, e.g., the bob.measure.cmc function.
  """
André Anjos's avatar
André Anjos committed
115 116 117
  # extract positives and negatives
  pos_dict = {}
  neg_dict = {}
118 119
  # read four column list
  for (client_id, probe_id, probe_name, score_str) in four_column(filename):
André Anjos's avatar
André Anjos committed
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
    try:
      score = float(score_str)
      # check in which dict we have to put the score
      if client_id == probe_id:
        correct_dict = pos_dict
      else:
        correct_dict = neg_dict
      # append score
      if probe_name in correct_dict:
        correct_dict[probe_name].append(score)
      else:
        correct_dict[probe_name] = [score]
    except:
      raise SyntaxError("Cannot convert score '%s' to float" % score_str)

  # convert to lists of tuples of ndarrays
  retval = []
  import logging
  logger = logging.getLogger('bob')
  for probe_name in sorted(pos_dict.keys()):
    if probe_name in neg_dict:
      retval.append((numpy.array(neg_dict[probe_name], numpy.float64), numpy.array(pos_dict[probe_name], numpy.float64)))
    else:
      logger.warn('For probe name "%s" there are only positive scores. This probe name is ignored.' % probe_name)
  # test if there are probes for which only negatives exist
  for probe_name in sorted(neg_dict.keys()):
    if not probe_name in pos_dict.keys():
       logger.warn('For probe name "%s" there are only negative scores. This probe name is ignored.' % probe_name)

  return retval

def five_column(filename):
  """Loads a score set from a single file to memory.

  Verifies that all fields are correctly placed and contain valid fields.

156
  Returns a python generator of tuples containing the following fields:
André Anjos's avatar
André Anjos committed
157 158 159 160 161 162 163 164 165 166 167 168 169

    [0]
      claimed identity (string)
    [1]
      model label (string)
    [2]
      real identity (string)
    [3]
      test label (string)
    [4]
      score (float)
  """

170
  for i, l in enumerate(open_file(filename)):
André Anjos's avatar
André Anjos committed
171 172 173 174 175 176 177 178 179
    s = l.strip()
    if len(s) == 0 or s[0] == '#': continue #empty or comment
    field = [k.strip() for k in s.split()]
    if len(field) < 5:
      raise SyntaxError('Line %d of file "%s" is invalid: %s' % (i, filename, l))
    try:
      score = float(field[4])
    except:
      raise SyntaxError('Cannot convert score to float at line %d of file "%s": %s' % (i, filename, l))
180
    yield (field[0], field[1], field[2], field[3], score)
André Anjos's avatar
André Anjos committed
181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196

def split_five_column(filename):
  """Loads a score set from a single file to memory and splits the scores
  between positives and negatives. The score file has to respect the 5 column
  format as defined in the method five_column().

  This method avoids loading and allocating memory for the strings present in
  the file. We only keep the scores.

  Returns a python tuple (negatives, positives). The values are 1-D blitz
  arrays of float64.
  """

  # split in positives and negatives
  neg = []
  pos = []
197 198 199 200 201 202
  # read five column list
  for (client_id, _, probe_id, _, score) in five_column(filename):
    if client_id == probe_id:
      pos.append(score)
    else:
      neg.append(score)
André Anjos's avatar
André Anjos committed
203 204 205 206 207

  return (numpy.array(neg, numpy.float64), numpy.array(pos, numpy.float64))

def cmc_five_column(filename):
  """Loads scores to compute CMC curves from a file in five column format.
208 209
  The four column file needs to be in the same format as described in the five_column function,
  and the "test label" (column 4) has to contain the test/probe file name.
André Anjos's avatar
André Anjos committed
210

211 212 213
  This function returns a list of tuples.
  For each probe file, the tuple consists of a list of negative scores and a list of positive scores.
  Usually, the list of positive scores should contain only one element, but more are allowed.
214

215
  The result of this function can directly be passed to, e.g., the bob.measure.cmc function.
André Anjos's avatar
André Anjos committed
216
  """
217
  # extract positives and negatives
André Anjos's avatar
André Anjos committed
218 219
  pos_dict = {}
  neg_dict = {}
220 221 222 223 224 225 226 227 228 229 230 231
  # read four column list
  for (client_id, _, probe_id, probe_name, score) in five_column(filename):
    # check in which dict we have to put the score
    if client_id == probe_id:
      correct_dict = pos_dict
    else:
      correct_dict = neg_dict
    # append score
    if probe_name in correct_dict:
      correct_dict[probe_name].append(score)
    else:
      correct_dict[probe_name] = [score]
André Anjos's avatar
André Anjos committed
232 233 234 235 236 237 238 239 240 241 242 243 244 245 246

  # convert to lists of tuples of ndarrays
  retval = []
  import logging
  logger = logging.getLogger('bob')
  for probe_name in sorted(pos_dict.keys()):
    if probe_name in neg_dict:
      retval.append((numpy.array(neg_dict[probe_name], numpy.float64), numpy.array(pos_dict[probe_name], numpy.float64)))
    else:
      logger.warn('For probe name "%s" there are only positive scores. This probe name is ignored.' % probe_name)
  # test if there are probes for which only negatives exist
  for probe_name in sorted(neg_dict.keys()):
    if not probe_name in pos_dict.keys():
       logger.warn('For probe name "%s" there are only negative scores. This probe name is ignored.' % probe_name)
  return retval