Source Code for Module jazzparser.utils.scripts

  1  """Utilities for bin scripts. 
  2   
  3  These are used by scripts outside the Jazz Parser packages, in  
  4  <PROJECT_ROOT>/bin. 
  5   
  6  """ 
  7  """ 
  8  ============================== License ======================================== 
  9   Copyright (C) 2008, 2010-12 University of Edinburgh, Mark Granroth-Wilding 
 10    
 11   This file is part of The Jazz Parser. 
 12    
 13   The Jazz Parser is free software: you can redistribute it and/or modify 
 14   it under the terms of the GNU General Public License as published by 
 15   the Free Software Foundation, either version 3 of the License, or 
 16   (at your option) any later version. 
 17    
 18   The Jazz Parser is distributed in the hope that it will be useful, 
 19   but WITHOUT ANY WARRANTY; without even the implied warranty of 
 20   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the 
 21   GNU General Public License for more details. 
 22    
 23   You should have received a copy of the GNU General Public License 
 24   along with The Jazz Parser.  If not, see <http://www.gnu.org/licenses/>. 
 25   
 26  ============================ End license ====================================== 
 27   
 28  """ 
 29  __author__ = "Mark Granroth-Wilding <mark.granroth-wilding@ed.ac.uk>"  
 30   
 31 -def prepare_evaluation_options(usage=None, description=None,  
 32          optparse_options=[], check_args=None, optparse_groups=[]): 
 33      """ 
 34      Various tasks common to the initial part of the evaluation routine 
 35      scripts (C{models/eval.py}). 
 36       
 37      @todo: This is not used any more. Remove it, after checking it's definitely  
 38          not used. 
 39       
 40      @param usage: the optparse usage string 
 41      @param description: the optparse description string 
 42      @type optparse_options: list of tuples 
 43      @param optparse_options: (args,kwargs) pairs to add additional  
 44          options to the optparse parser. 
 45      @type check_args: function 
 46      @param check_args: function to take the command-line arguments and  
 47          check them. This will be called early in the script. Must  
 48          return a tuple of (1) the model name (or model basename) that  
 49          will be used in the partition model names and (2) the input  
 50          filename to get sequences from. 
 51      @type optparse_groups: list of pairs 
 52      @param optparse_groups: specificatios for option groups to add to the  
 53          optparse option parser. The first of each pair is a tuple of  
 54          args to C{OptionGroup}'s init (excluding the first).  
 55          The second is a list of options  
 56          each formatted as C{optparse_options}. 
 57           
 58      @rtype: tuple 
 59      @return: (1) list of (sequences,model_name,partition_index) tuples 
 60          for each partition; (2) list of lists containing the sequence  
 61          ids for each partition; (3) optparse options; (4) optparse  
 62          arguments. 
 63       
 64      """ 
 65      import sys 
 66      from optparse import OptionParser, OptionGroup 
 67      from jazzparser.utils.config import parse_args_with_config 
 68      from jazzparser.utils.loggers import init_logging 
 69      from jazzparser.data.db_mirrors import SequenceIndex 
 70      from jazzparser.utils.data import partition 
 71       
 72      parser = OptionParser(usage=usage, description=description) 
 73      group = OptionGroup(parser, "Input", "Input data and partitioning for evaluation") 
 74      group.add_option("-s", "--sequence", dest="sequence", action="store", help="limit the evaluation to just one sequence, with the given index in the input file") 
 75      group.add_option("--partition", dest="partition", action="store", help="restrict to only one partition of the data. Specify as i/n, where i is the partition number and n the total number of partitions.") 
 76      group.add_option("-p", "--partitions", dest="partitions", type="int", action="store", help="test on all n partitions of the data, using a different model for each. Will look for a model <NAME>i, where <NAME> is the given model name and i the partition number.") 
 77      parser.add_option_group(group) 
 78       
 79      parser.add_option("--debug", dest="debug", action="store_true", help="show debugging output") 
 80       
 81      # Add the options according to their specs 
 82      for args,kwargs in optparse_options: 
 83          parser.add_option(*args, **kwargs) 
 84           
 85      # Add groups and their options 
 86      for group_args,options in optparse_groups: 
 87          # Check whether the group already exists 
 88          same_titles = [g for g in parser.option_groups if g.title == group_args[0]] 
 89          if same_titles: 
 90              group = same_titles[0] 
 91          else: 
 92              group = OptionGroup(parser, *group_args) 
 93              parser.add_option_group(group) 
 94          # Add options to this group 
 95          for args,kwargs in options: 
 96              group.add_option(*args, **kwargs) 
 97      options, arguments = parse_args_with_config(parser) 
 98       
 99      if check_args is None: 
100          raise ValueError, "could not check arguments and get model "\ 
101              "name. check_args must not be None" 
102      model_name,input_filename = check_args(arguments) 
103           
104      if options.debug: 
105          # Set the log level to debug and do the standard logging init 
106          init_logging(logging.DEBUG) 
107      else: 
108          init_logging() 
109           
110      # Load up sequences 
111      seqs = SequenceIndex.from_file(input_filename) 
112           
113      def _get_seq_by_index(index): 
114          seq = seqs.sequence_by_index(index) 
115          if seq is None: 
116              print >>sys.stderr, "There are only %d sequences" % len(seqs) 
117              sys.exit(1) 
118          return seq 
119       
120      ################ Data partitioning #################### 
121      if options.partitions is not None: 
122          # Divide the data up into n partitions and use a different model name for each 
123          total_parts = options.partitions 
124          print >>sys.stderr, "Cross validation: dividing test data into %d partitions" % total_parts 
125          partitions = [(part,"%s%d" % (model_name,i), i) for i,part in enumerate(partition(seqs.sequences, total_parts))] 
126          part_ids = partition(seqs.ids, total_parts) 
127      elif options.partition is not None: 
128          # Just select one partition 
129          # Split up the argument to get two integers 
130          parti,total_parts = options.partition.split("/") 
131          parti,total_parts = int(parti), int(total_parts) 
132          print >>sys.stderr, "Restricting sequences to %d-way partition %d" % (total_parts,parti) 
133          # Get a list of sequence indices to restrict our set to 
134          part_ids = partition(seqs.ids, total_parts)[parti] 
135          partitions = [ [(part,"%s%d" % (model_name,i), i) for i,part in enumerate(partition(seqs.sequences, total_parts))][parti] ] 
136      elif options.sequence is not None: 
137          # Just select one sequence 
138          seq = _get_seq_by_index(int(options.sequence)) 
139          partitions = [( [seq], model_name, 0 )] 
140          part_ids = [seq.id] 
141      else: 
142          # Don't partition the sequences 
143          partitions = [(seqs.sequences, model_name,0)] 
144          part_ids = [None] 
145       
146      return partitions,part_ids,options,arguments 
147