Merge branch 'opt_test' into 'master'

Make use of test set on compute_perf optional; Update docs

This MR addresses two issues:

* Users which don't have a test set can now also use `bob_compute_perf`. In this case, only reduced number of statistics and plots are printed
* Fixes to the user guide, mainly concerning how to call the scripts since the simplifications with docopt

See merge request !22

bob/measure/script/compute_perf.py

@@ -6,20 +6,21 @@
 
   1. Computes the threshold using either EER or min. HTER criteria on
      development set scores
-  2. Applies the above threshold on test set scores to compute the HTER
+  2. Applies the above threshold on test set scores to compute the HTER, if a
+     test-score set is provided
   3. Reports error rates on the console
   4. Plots ROC, EPC, DET curves and score distributions to a multi-page PDF
      file (unless --no-plot is passed)
 
 
-Usage: %(prog)s [-v...] [options] <dev-scores> <test-scores>
+Usage: %(prog)s [-v...] [options] <dev-scores> [<test-scores>]
        %(prog)s --help
        %(prog)s --version
 
 
 Arguments:
   <dev-scores>   Path to the file containing the development scores
-  <test-scores>  Path to the file containing the test scores
+  <test-scores>  (optional) Path to the file containing the test scores.
 
 
 Options:
@@ -53,15 +54,15 @@ import os
 import sys
 import numpy
 
-import logging
-__logging_format__='[%(levelname)s] %(message)s'
-logging.basicConfig(format=__logging_format__)
-logger = logging.getLogger('bob')
+import bob.core
+logger = bob.core.log.setup("bob.measure")
 
 
-def print_crit(dev_neg, dev_pos, test_neg, test_pos, crit):
+def print_crit(crit, dev_scores, test_scores=None):
   """Prints a single output line that contains all info for a given criterion"""
 
+  dev_neg, dev_pos = dev_scores
+
   if crit == 'EER':
     from .. import eer_threshold
     thres = eer_threshold(dev_neg, dev_pos)
@@ -73,46 +74,71 @@ def print_crit(dev_neg, dev_pos, test_neg, test_pos, crit):
   dev_far, dev_frr = farfrr(dev_neg, dev_pos, thres)
   dev_hter = (dev_far + dev_frr)/2.0
 
-  test_far, test_frr = farfrr(test_neg, test_pos, thres)
-  test_hter = (test_far + test_frr)/2.0
-
   print("[Min. criterion: %s] Threshold on Development set: %e" % (crit, thres))
 
   dev_ni = dev_neg.shape[0] #number of impostors
   dev_fa = int(round(dev_far*dev_ni)) #number of false accepts
   dev_nc = dev_pos.shape[0] #number of clients
   dev_fr = int(round(dev_frr*dev_nc)) #number of false rejects
-  test_ni = test_neg.shape[0] #number of impostors
-  test_fa = int(round(test_far*test_ni)) #number of false accepts
-  test_nc = test_pos.shape[0] #number of clients
-  test_fr = int(round(test_frr*test_nc)) #number of false rejects
 
   dev_far_str = "%.3f%% (%d/%d)" % (100*dev_far, dev_fa, dev_ni)
-  test_far_str = "%.3f%% (%d/%d)" % (100*test_far, test_fa, test_ni)
   dev_frr_str = "%.3f%% (%d/%d)" % (100*dev_frr, dev_fr, dev_nc)
-  test_frr_str = "%.3f%% (%d/%d)" % (100*test_frr, test_fr, test_nc)
   dev_max_len = max(len(dev_far_str), len(dev_frr_str))
-  test_max_len = max(len(test_far_str), len(test_frr_str))
 
   def fmt(s, space):
     return ('%' + ('%d' % space) + 's') % s
 
-  print("       | %s | %s" % (fmt("Development", -1*dev_max_len),
-    fmt("Test", -1*test_max_len)))
-  print("-------+-%s-+-%s" % (dev_max_len*"-", (2+test_max_len)*"-"))
-  print("  FAR  | %s | %s" % (fmt(dev_far_str, dev_max_len), fmt(test_far_str,
-    test_max_len)))
-  print("  FRR  | %s | %s" % (fmt(dev_frr_str, dev_max_len), fmt(test_frr_str,
-    test_max_len)))
-  dev_hter_str = "%.3f%%" % (100*dev_hter)
-  test_hter_str = "%.3f%%" % (100*test_hter)
-  print("  HTER | %s | %s" % (fmt(dev_hter_str, -1*dev_max_len),
-    fmt(test_hter_str, -1*test_max_len)))
+  if test_scores is None:
+
+    # prints only dev performance rates
+    print("       | %s" % fmt("Development", -1*dev_max_len))
+    print("-------+-%s" % (dev_max_len*"-"))
+    print("  FAR  | %s" % fmt(dev_far_str, dev_max_len))
+    print("  FRR  | %s" % fmt(dev_frr_str, dev_max_len))
+    dev_hter_str = "%.3f%%" % (100*dev_hter)
+    print("  HTER | %s" % fmt(dev_hter_str, -1*dev_max_len))
+
+  else:
+
+    # computes statistics for the test set based on the threshold a priori
+    test_neg, test_pos = test_scores
+
+    test_far, test_frr = farfrr(test_neg, test_pos, thres)
+    test_hter = (test_far + test_frr)/2.0
+
+    test_ni = test_neg.shape[0] #number of impostors
+    test_fa = int(round(test_far*test_ni)) #number of false accepts
+    test_nc = test_pos.shape[0] #number of clients
+    test_fr = int(round(test_frr*test_nc)) #number of false rejects
+
+    test_far_str = "%.3f%% (%d/%d)" % (100*test_far, test_fa, test_ni)
+    test_frr_str = "%.3f%% (%d/%d)" % (100*test_frr, test_fr, test_nc)
+    test_max_len = max(len(test_far_str), len(test_frr_str))
+
+    # prints both dev and test performance rates
+    print("       | %s | %s" % (fmt("Development", -1*dev_max_len),
+      fmt("Test", -1*test_max_len)))
+    print("-------+-%s-+-%s" % (dev_max_len*"-", (2+test_max_len)*"-"))
+    print("  FAR  | %s | %s" % (fmt(dev_far_str, dev_max_len),
+      fmt(test_far_str, test_max_len)))
+    print("  FRR  | %s | %s" % (fmt(dev_frr_str, dev_max_len),
+      fmt(test_frr_str, test_max_len)))
+    dev_hter_str = "%.3f%%" % (100*dev_hter)
+    test_hter_str = "%.3f%%" % (100*test_hter)
+    print("  HTER | %s | %s" % (fmt(dev_hter_str, -1*dev_max_len),
+      fmt(test_hter_str, -1*test_max_len)))
 
 
-def plots(dev_neg, dev_pos, test_neg, test_pos, crit, points, filename):
+def plots(crit, points, filename, dev_scores, test_scores=None):
   """Saves ROC, DET and EPC curves on the file pointed out by filename."""
 
+  dev_neg, dev_pos = dev_scores
+
+  if test_scores is not None:
+    test_neg, test_pos = test_scores
+  else:
+    test_neg, test_pos = None, None
+
   from .. import plot
 
   import matplotlib
@@ -124,41 +150,54 @@ def plots(dev_neg, dev_pos, test_neg, test_pos, crit, points, filename):
 
   # ROC
   fig = mpl.figure()
-  plot.roc(dev_neg, dev_pos, points, color=(0.3,0.3,0.3),
-      linestyle='--', dashes=(6,2), label='development')
-  plot.roc(test_neg, test_pos, points, color=(0,0,0),
-      linestyle='-', label='test')
+
+  if test_scores is not None:
+    plot.roc(dev_neg, dev_pos, points, color=(0.3,0.3,0.3),
+        linestyle='--', dashes=(6,2), label='development')
+    plot.roc(test_neg, test_pos, points, color=(0,0,0),
+        linestyle='-', label='test')
+  else:
+    plot.roc(dev_neg, dev_pos, points, color=(0,0,0),
+        linestyle='-', label='development')
+
   mpl.axis([0,40,0,40])
   mpl.title("ROC Curve")
   mpl.xlabel('FAR (%)')
   mpl.ylabel('FRR (%)')
   mpl.grid(True, color=(0.3,0.3,0.3))
-  mpl.legend()
+  if test_scores is not None: mpl.legend()
   pp.savefig(fig)
 
   # DET
   fig = mpl.figure()
-  plot.det(dev_neg, dev_pos, points, color=(0.3,0.3,0.3),
-      linestyle='--', dashes=(6,2), label='development')
-  plot.det(test_neg, test_pos, points, color=(0,0,0),
-      linestyle='-', label='test')
+
+  if test_scores is not None:
+    plot.det(dev_neg, dev_pos, points, color=(0.3,0.3,0.3),
+        linestyle='--', dashes=(6,2), label='development')
+    plot.det(test_neg, test_pos, points, color=(0,0,0),
+        linestyle='-', label='test')
+  else:
+    plot.det(dev_neg, dev_pos, points, color=(0,0,0),
+        linestyle='-', label='development')
+
   plot.det_axis([0.01, 40, 0.01, 40])
   mpl.title("DET Curve")
   mpl.xlabel('FAR (%)')
   mpl.ylabel('FRR (%)')
   mpl.grid(True, color=(0.3,0.3,0.3))
-  mpl.legend()
+  if test_scores is not None: mpl.legend()
   pp.savefig(fig)
 
-  # EPC
-  fig = mpl.figure()
-  plot.epc(dev_neg, dev_pos, test_neg, test_pos, points,
-      color=(0,0,0), linestyle='-')
-  mpl.title('EPC Curve')
-  mpl.xlabel('Cost')
-  mpl.ylabel('Min. HTER (%)')
-  mpl.grid(True, color=(0.3,0.3,0.3))
-  pp.savefig(fig)
+  # EPC - requires test set
+  if test_scores is not None:
+    fig = mpl.figure()
+    plot.epc(dev_neg, dev_pos, test_neg, test_pos, points,
+        color=(0,0,0), linestyle='-')
+    mpl.title('EPC Curve')
+    mpl.xlabel('Cost')
+    mpl.ylabel('Min. HTER (%)')
+    mpl.grid(True, color=(0.3,0.3,0.3))
+    pp.savefig(fig)
 
   # Distribution for dev and test scores on the same page
   if crit == 'EER':
@@ -168,9 +207,15 @@ def plots(dev_neg, dev_pos, test_neg, test_pos, crit, points, filename):
     from .. import min_hter_threshold
     thres = min_hter_threshold(dev_neg, dev_pos)
 
-  mpl.subplot(2,1,1)
+  fig = mpl.figure()
+
+  if test_scores is not None:
+    mpl.subplot(2,1,1)
+    all_scores = numpy.hstack((dev_neg, test_neg, dev_pos, test_pos))
+  else:
+    all_scores = numpy.hstack((dev_neg, dev_pos))
+
   nbins=20
-  all_scores = numpy.hstack((dev_neg, test_neg, dev_pos, test_pos))
   score_range = all_scores.min(), all_scores.max()
   mpl.hist(dev_neg, label='Impostors', normed=True, color='red', alpha=0.5,
       bins=nbins)
@@ -179,25 +224,33 @@ def plots(dev_neg, dev_pos, test_neg, test_pos, crit, points, filename):
   mpl.xlim(*score_range)
   _, _, ymax, ymin = mpl.axis()
   mpl.vlines(thres, ymin, ymax, color='black', label='EER', linestyle='dashed')
-  mpl.ylabel('Dev. Scores (normalized)')
-  ax = mpl.gca()
-  ax.axes.get_xaxis().set_ticklabels([])
-  mpl.legend(loc='upper center', ncol=3, bbox_to_anchor=(0.5, -0.01),
-      fontsize=10)
+
+  if test_scores is not None:
+    ax = mpl.gca()
+    ax.axes.get_xaxis().set_ticklabels([])
+    mpl.legend(loc='upper center', ncol=3, bbox_to_anchor=(0.5, -0.01),
+        fontsize=10)
+    mpl.ylabel('Dev. Scores (normalized)')
+  else:
+    mpl.ylabel('Normalized Count')
+    mpl.legend(loc='best', fancybox=True, framealpha=0.5)
   mpl.title('Score Distributions')
   mpl.grid(True, alpha=0.5)
 
-  mpl.subplot(2,1,2)
-  mpl.hist(test_neg, label='Impostors', normed=True, color='red', alpha=0.5,
-      bins=nbins)
-  mpl.hist(test_pos, label='Genuine', normed=True, color='blue', alpha=0.5,
-      bins=nbins)
-  mpl.ylabel('Test Scores (normalized)')
-  mpl.xlabel('Score value')
-  mpl.xlim(*score_range)
-  _, _, ymax, ymin = mpl.axis()
-  mpl.vlines(thres, ymin, ymax, color='black', label='EER', linestyle='dashed')
-  mpl.grid(True, alpha=0.5)
+  if test_scores is not None:
+    mpl.subplot(2,1,2)
+    mpl.hist(test_neg, label='Impostors', normed=True, color='red', alpha=0.5,
+        bins=nbins)
+    mpl.hist(test_pos, label='Genuine', normed=True, color='blue', alpha=0.5,
+        bins=nbins)
+    mpl.ylabel('Test Scores (normalized)')
+    mpl.xlabel('Score value')
+    mpl.xlim(*score_range)
+    _, _, ymax, ymin = mpl.axis()
+    mpl.vlines(thres, ymin, ymax, color='black', label='EER',
+        linestyle='dashed')
+    mpl.grid(True, alpha=0.5)
+
   pp.savefig(fig)
 
   pp.close()
@@ -225,8 +278,8 @@ def main(user_input=None):
       )
 
   # Sets-up logging
-  if args['--verbose'] == 1: logging.getLogger().setLevel(logging.INFO)
-  elif args['--verbose'] >= 2: logging.getLogger().setLevel(logging.DEBUG)
+  verbosity = int(args['--verbose'])
+  bob.core.log.set_verbosity_level(logger, verbosity)
 
   # Checks number of points option
   try:
@@ -240,15 +293,18 @@ def main(user_input=None):
         'than zero')
 
   from ..load import load_score, get_negatives_positives
-  dev_neg, dev_pos = get_negatives_positives(load_score(args['<dev-scores>']))
-  test_neg, test_pos = get_negatives_positives(load_score(args['<test-scores>']))
+  dev_scores = get_negatives_positives(load_score(args['<dev-scores>']))
+
+  if args['<test-scores>'] is not None:
+    test_scores = get_negatives_positives(load_score(args['<test-scores>']))
+  else:
+    test_scores = None
 
-  print_crit(dev_neg, dev_pos, test_neg, test_pos, 'EER')
-  print_crit(dev_neg, dev_pos, test_neg, test_pos, 'Min. HTER')
+  print_crit('EER', dev_scores, test_scores)
+  print_crit('Min. HTER', dev_scores, test_scores)
 
   if not args['--no-plot']:
-    plots(dev_neg, dev_pos, test_neg, test_pos, 'EER', args['--points'],
-        args['--output'])
+    plots('EER', args['--points'], args['--output'], dev_scores, test_scores)
     print("[Plots] Performance curves => '%s'" % args['--output'])
 
   return 0

bob/measure/test_scripts.py

@@ -46,6 +46,22 @@ def test_compute_perf():
   nose.tools.eq_(main(cmdline), 0)
 
 
+def test_compute_perf_only_dev():
+
+  # sanity checks
+  assert os.path.exists(DEV_SCORES)
+
+  tmp_output = tempfile.NamedTemporaryFile(prefix=__name__, suffix='.pdf')
+
+  cmdline = [
+      DEV_SCORES,
+      '--output=' + tmp_output.name,
+      ]
+
+  from .script.compute_perf import main
+  nose.tools.eq_(main(cmdline), 0)
+
+
 def test_eval_threshold():
 
   # sanity checks

doc/guide.rst

@@ -64,8 +64,8 @@ scenarios, the threshold :math:`\tau` has to be set a priori: this is typically
 done using a development set (also called cross-validation set). Nevertheless,
 the optimal threshold can be different depending on the relative importance
 given to the FAR and the FRR. Hence, in the EPC framework, the cost
-:math:`\beta \in [0;1]` is defined as the trade-off between the FAR and FRR. The
-optimal threshold :math:`\tau^*` is then computed using different values of
+:math:`\beta \in [0;1]` is defined as the trade-off between the FAR and FRR.
+The optimal threshold :math:`\tau^*` is then computed using different values of
 :math:`\beta`, corresponding to different operating points:
 
 .. math::
@@ -455,7 +455,7 @@ Error Rate) on a set, after setting up |project|, just do:
 
 .. code-block:: sh
 
-  $ bob_eval_threshold.py --scores=development-scores-4col.txt
+  $ bob_eval_threshold.py development-scores-4col.txt
   Threshold: -0.004787956164
   FAR : 6.731% (35/520)
   FRR : 6.667% (26/390)
@@ -470,18 +470,18 @@ application ``bob_apply_threshold.py``:
 
 .. code-block:: sh
 
-  $ bob_apply_threshold.py --scores=test-scores-4col.txt --threshold=-0.0047879
+  $ bob_apply_threshold.py -0.0047879 test-scores-4col.txt
   FAR : 2.115% (11/520)
   FRR : 7.179% (28/390)
   HTER: 4.647%
 
 In this case, only the error figures are presented. You can conduct the
-evaluation and plotting of development and test set data using our combined
+evaluation and plotting of development (and test set data) using our combined
 ``bob_compute_perf.py`` script. You pass both sets and it does the rest:
 
 .. code-block:: sh
 
-  $ bob_compute_perf.py --devel=development-scores-4col.txt --test=test-scores-4col.txt
+  $ bob_compute_perf.py development-scores-4col.txt test-scores-4col.txt
   [Min. criterion: EER] Threshold on Development set: -4.787956e-03
          | Development     | Test
   -------+-----------------+------------------
@@ -499,8 +499,8 @@ evaluation and plotting of development and test set data using our combined
 Inside that script we evaluate 2 different thresholds based on the EER and the
 minimum HTER on the development set and apply the output to the test set. As
 can be seen from the toy-example above, the system generalizes reasonably well.
-A single PDF file is generated containing an EPC as well as ROC and DET plots of such a
-system.
+A single PDF file is generated containing an EPC as well as ROC and DET plots
+of such a system.
 
 Use the ``--help`` option on the above-cited scripts to find-out about more
 options.
@@ -509,71 +509,105 @@ options.
 Score file conversion
 ---------------------
 
-Sometimes, it is required to export the score files generated by Bob to a different format, e.g., to be able to generate a plot comparing Bob's systems with other systems.
-In this package, we provide source code to convert between different types of score files.
+Sometimes, it is required to export the score files generated by Bob to a
+different format, e.g., to be able to generate a plot comparing Bob's systems
+with other systems.  In this package, we provide source code to convert between
+different types of score files.
 
 Bob to OpenBR
 =============
 
-One of the supported formats is the matrix format that the National Institute of Standards and Technology (NIST) uses, and which is supported by OpenBR_.
-The scores are stored in two binary matrices, where the first matrix (usually with a ``.mtx`` filename extension) contains the raw scores, while a second mask matrix (extension ``.mask``) contains information, which scores are positives, and which are negatives.
+One of the supported formats is the matrix format that the National Institute
+of Standards and Technology (NIST) uses, and which is supported by OpenBR_.
+The scores are stored in two binary matrices, where the first matrix (usually
+with a ``.mtx`` filename extension) contains the raw scores, while a second
+mask matrix (extension ``.mask``) contains information, which scores are
+positives, and which are negatives.
 
-To convert from Bob's four column or five column score file to a pair of these matrices, you can use the :py:func:`bob.measure.openbr.write_matrix` function.
-In the simplest way, this function takes a score file ``'five-column-sore-file'`` and writes the pair ``'openbr.mtx', 'openbr.mask'`` of OpenBR compatible files:
+To convert from Bob's four column or five column score file to a pair of these
+matrices, you can use the :py:func:`bob.measure.openbr.write_matrix` function.
+In the simplest way, this function takes a score file
+``'five-column-sore-file'`` and writes the pair ``'openbr.mtx', 'openbr.mask'``
+of OpenBR compatible files:
 
 .. code-block:: py
 
    >>> bob.measure.openbr.write_matrix('five-column-sore-file', 'openbr.mtx', 'openbr.mask', score_file_format = '5column')
 
-In this way, the score file will be parsed and the matrices will be written in the same order that is obtained from the score file.
-
-For most of the applications, this should be sufficient, but as the identity information is lost in the matrix files, no deeper analysis is possible anymore when just using the matrices.
-To enforce an order of the models and probes inside the matrices, you can use the ``model_names`` and ``probe_names`` parameters of :py:func:`bob.measure.openbr.write_matrix`:
-
-* The ``probe_names`` parameter lists the ``path`` elements stored in the score files, which are the fourth column in a ``5column`` file, and the third column in a ``4column`` file, see :py:func:`bob.measure.load.five_column` and :py:func:`bob.measure.load.four_column`.
-
-* The ``model_names`` parameter is a bit more complicated.
-  In a ``5column`` format score file, the model names are defined by the second column of that file, see :py:func:`bob.measure.load.five_column`.
-  In a ``4column`` format score file, the model information is not contained, but only the client information of the model.
-  Hence, for the ``4column`` format, the ``model_names`` actually lists the client ids found in the first column, see :py:func:`bob.measure.load.four_column`.
+In this way, the score file will be parsed and the matrices will be written in
+the same order that is obtained from the score file.
+
+For most of the applications, this should be sufficient, but as the identity
+information is lost in the matrix files, no deeper analysis is possible anymore
+when just using the matrices.  To enforce an order of the models and probes
+inside the matrices, you can use the ``model_names`` and ``probe_names``
+parameters of :py:func:`bob.measure.openbr.write_matrix`:
+
+* The ``probe_names`` parameter lists the ``path`` elements stored in the score
+  files, which are the fourth column in a ``5column`` file, and the third
+  column in a ``4column`` file, see :py:func:`bob.measure.load.five_column` and
+  :py:func:`bob.measure.load.four_column`.
+* The ``model_names`` parameter is a bit more complicated.  In a ``5column``
+  format score file, the model names are defined by the second column of that
+  file, see :py:func:`bob.measure.load.five_column`.  In a ``4column`` format
+  score file, the model information is not contained, but only the client
+  information of the model.  Hence, for the ``4column`` format, the
+  ``model_names`` actually lists the client ids found in the first column, see
+  :py:func:`bob.measure.load.four_column`.
 
   .. warning::
-     The model information is lost, but required to write the matrix files.
-     In the ``4column`` format, we use client ids instead of the model information.
-     Hence, when several models exist per client, this function will not work as expected.
-
-Additionally, there are fields in the matrix files, which define the gallery and probe list files that were used to generate the matrix.
-These file names can be selected with the ``gallery_file_name`` and ``probe_file_name`` keyword parameters of :py:func:`bob.measure.openbr.write_matrix`.
-
-Finally, OpenBR defines a specific ``'search'`` score file format, which is designed to be used to compute CMC curves.
-The score matrix contains descendingly sorted and possibly truncated list of scores, i.e., for each probe, a sorted list of all scores for the models is generated.
-To generate these special score file format, you can specify the ``search`` parameter.
-It specifies the number of highest scores per probe that should be kept.
-If the ``search`` parameter is set to a negative value, all scores will be kept.
-If the ``search`` parameter is higher as the actual number of models, ``NaN`` scores will be appended, and the according mask values will be set to ``0`` (i.e., to be ignored).
+     The model information is lost, but required to write the matrix files.  In
+     the ``4column`` format, we use client ids instead of the model
+     information.  Hence, when several models exist per client, this function
+     will not work as expected.
+
+Additionally, there are fields in the matrix files, which define the gallery
+and probe list files that were used to generate the matrix.  These file names
+can be selected with the ``gallery_file_name`` and ``probe_file_name`` keyword
+parameters of :py:func:`bob.measure.openbr.write_matrix`.
+
+Finally, OpenBR defines a specific ``'search'`` score file format, which is
+designed to be used to compute CMC curves.  The score matrix contains
+descendingly sorted and possibly truncated list of scores, i.e., for each
+probe, a sorted list of all scores for the models is generated.  To generate
+these special score file format, you can specify the ``search`` parameter.  It
+specifies the number of highest scores per probe that should be kept.  If the
+``search`` parameter is set to a negative value, all scores will be kept.  If
+the ``search`` parameter is higher as the actual number of models, ``NaN``
+scores will be appended, and the according mask values will be set to ``0``
+(i.e., to be ignored).
 
 
 OpenBR to Bob
 =============
 
-On the other hand, you might also want to generate a Bob-compatible (four or five column) score file based on a pair of OpenBR matrix and mask files.
-This is possible by using the :py:func:`bob.measure.openbr.write_score_file` function.
-At the basic, it takes the given pair of matrix and mask files, as well as the desired output score file:
+On the other hand, you might also want to generate a Bob-compatible (four or
+five column) score file based on a pair of OpenBR matrix and mask files.  This
+is possible by using the :py:func:`bob.measure.openbr.write_score_file`
+function.  At the basic, it takes the given pair of matrix and mask files, as
+well as the desired output score file:
 
 .. code-block:: py
 
    >>> bob.measure.openbr.write_score_file('openbr.mtx', 'openbr.mask', 'four-column-sore-file')
 
-This score file is sufficient to compute a CMC curve (see `CMC`_), however it does not contain relevant client ids or paths for models and probes.
+This score file is sufficient to compute a CMC curve (see `CMC`_), however it
+does not contain relevant client ids or paths for models and probes.
 Particularly, it assumes that each client has exactly one associated model.
 
-To add/correct these information, you can use additional parameters to :py:func:`bob.measure.openbr.write_score_file`.
-Client ids of models and probes can be added using the ``models_ids`` and ``probes_ids`` keyword arguments.
-The length of these lists must be identical to the number of models and probes as given in the matrix files, **and they must be in the same order as used to compute the OpenBR matrix**.
-This includes that the same same-client and different-client pairs as indicated by the OpenBR mask will be generated, which will be checked inside the function.
-
-To add model and probe path information, the ``model_names`` and ``probe_names`` parameters, which need to have the same size and order as the ``models_ids`` and ``probes_ids``.
-These information are simply stored in the score file, and no further check is applied.
+To add/correct these information, you can use additional parameters to
+:py:func:`bob.measure.openbr.write_score_file`.  Client ids of models and
+probes can be added using the ``models_ids`` and ``probes_ids`` keyword
+arguments.  The length of these lists must be identical to the number of models
+and probes as given in the matrix files, **and they must be in the same order
+as used to compute the OpenBR matrix**.  This includes that the same
+same-client and different-client pairs as indicated by the OpenBR mask will be
+generated, which will be checked inside the function.
+
+To add model and probe path information, the ``model_names`` and
+``probe_names`` parameters, which need to have the same size and order as the
+``models_ids`` and ``probes_ids``.  These information are simply stored in the
+score file, and no further check is applied.
 
 .. note:: The ``model_names`` parameter is used only when writing score files in ``score_file_format='5column'``, in the ``'4column'`` format, this parameter is ignored.