v5.1.

CAT_pack/CAT

@@ -12,43 +12,49 @@ import summarise
 
 
 def usage():
-    message = ('usage: CAT (prepare | contigs | bin | bins | add_names | '
-               'summarise) [-v / --version] [-h / --help]\n'
-               'CAT: error: one of the arguments prepare contigs bin bins '
-               'add_names summarise is required')
+    message = (
+            'usage: CAT (prepare | contigs | bin | bins | add_names | '
+            'summarise) [-v / --version] [-h / --help]\n'
+            'CAT: error: one of the arguments prepare contigs bin bins '
+            'add_names summarise is required')
 
     sys.stdout.write('{0}\n'.format(message))
+
+    return
 
 
 def version():
-    message = ('CAT v{0} ({1}) by {2}.'
-               ''.format(about.__version__, about.__date__, about.__author__))
+    message = ('CAT v{0} ({1}) by {2}.'.format(
+        about.__version__, about.__date__, about.__author__))
 
     sys.stdout.write('{0}\n'.format(message))
+
+    return
 
 
 def help():
-    message = ('usage: CAT (prepare | contigs | bin | bins | add_names | '
-               'summarise) [-v / --version] [-h / --help]\n\n'
-               'Run Contig Annotation Tool (CAT) or '
-               'Bin Annotation Tool (BAT).\n\n'
-               'Required choice:\n'
-               '  prepare\t\tDownload database files and construct '
-               'databases.\n'
-               '  contigs\t\tRun CAT.\n'
-               '  bin\t\t\tRun BAT on a single bin.\n'
-               '  bins\t\t\tRun BAT on a set of bins.\n'
-               '  add_names\t\tAdd taxonomic names to CAT or BAT output '
-               'files.\n'
-               '  summarise\t\tSummarise a named CAT or BAT classification '
-               'file.\n\n'
-               'Optional arguments:\n'
-               '  -v, --version\t\tPrint version information and exit.\n'
-               '  -h, --help\t\tShow this help message and exit.')
+    message = (
+            'usage: CAT (prepare | contigs | bin | bins | add_names | '
+            'summarise) [-v / --version] [-h / --help]\n\n'
+            'Run Contig Annotation Tool (CAT) or '
+            'Bin Annotation Tool (BAT).\n\n'
+            'Required choice:\n'
+            '  prepare\t\tDownload database files and construct databases.\n'
+            '  contigs\t\tRun CAT.\n'
+            '  bin\t\t\tRun BAT on a single bin.\n'
+            '  bins\t\t\tRun BAT on a set of bins.\n'
+            '  add_names\t\tAdd taxonomic names to CAT or BAT output files.\n'
+            '  summarise\t\tSummarise a named CAT or BAT classification file.'
+            '\n\n'
+            'Optional arguments:\n'
+            '  -v, --version\t\tPrint version information and exit.\n'
+            '  -h, --help\t\tShow this help message and exit.')
 
     sys.stdout.write('{0}\n'.format(message))
-
-
+
+    return
+
+
 def main():
     if len(sys.argv) == 1:
         usage()
@@ -71,6 +77,8 @@ def main():
     else:
         usage()
 
+    return
+
 
 if __name__ == '__main__':
     main()

CAT_pack/about.py

@@ -1,5 +1,5 @@
 #!/usr/bin/env python3
 
 __author__ = 'F. A. Bastiaan von Meijenfeldt'
-__version__ = '5.0.5.1'
-__date__ = '18 June, 2020'
+__version__ = '5.1'
+__date__ = '22 June, 2020'

CAT_pack/add_names.py

@@ -1,7 +1,6 @@
 #!/usr/bin/env python3
 
 import argparse
-import os
 import sys
 
 import about
@@ -11,115 +10,121 @@
 
 
 def parse_arguments():
-    parser = argparse.ArgumentParser(prog='CAT add_names',
-                                     description='Add taxonomic names to CAT '
-                                                 'or BAT output files.',
-                                     usage='CAT add_names -i -o -t '
-                                           '[options] [-h / --help]',
-                                     add_help=False)
+    parser = argparse.ArgumentParser(
+            prog='CAT add_names',
+            description='Add taxonomic names to CAT or BAT output files.',
+            usage='CAT add_names -i -o -t [options] [-h / --help]',
+            add_help=False)
 
     required = parser.add_argument_group('Required arguments')
 
-    required.add_argument('-i',
-                          '--input_file',
-                          dest='input_file',
-                          metavar='',
-                          required=True,
-                          type=str,
-                          help='Path to input file. Can be either '
-                               'classification output file or ORF2LCA output '
-                               'file.')
-    required.add_argument('-o',
-                          '--output_file',
-                          dest='output_file',
-                          metavar='',
-                          required=True,
-                          type=str,
-                          help='Path to output file.')
-    required.add_argument('-t',
-                          '--taxonomy_folder',
-                          dest='taxonomy_folder',
-                          metavar='',
-                          required=True,
-                          type=str,
-                          help='Path to folder that contains taxonomy files.')
+    required.add_argument(
+            '-i',
+            '--input_file',
+            dest='input_file',
+            metavar='',
+            required=True,
+            type=str,
+            action=shared.PathAction,
+            help=('Path to input file. Can be either classification output '
+                'file or ORF2LCA output file.'))
+    required.add_argument(
+            '-o',
+            '--output_file',
+            dest='output_file',
+            metavar='',
+            required=True,
+            type=str,
+            action=shared.PathAction,
+            help='Path to output file.')
+    required.add_argument(
+            '-t',
+            '--taxonomy_folder',
+            dest='taxonomy_folder',
+            metavar='',
+            required=True,
+            type=str,
+            action=shared.PathAction,
+            help='Path to folder that contains taxonomy files.')
 
     optional = parser.add_argument_group('Optional arguments')
 
-    optional.add_argument('--only_official',
-                          dest='only_official',
-                          required=False,
-                          action='store_true',
-                          help='Only output official level names.')
-    optional.add_argument('--exclude_scores',
-                          dest='exclude_scores',
-                          required=False,
-                          action='store_true',
-                          help='Do not include bit-score support scores in '
-                               'the lineage.')
-    optional.add_argument('--force',
-                          dest='force',
-                          required=False,
-                          action='store_true',
-                          help='Force overwrite existing files.')
-    optional.add_argument('-q',
-                          '--quiet',
-                          dest='quiet',
-                          required=False,
-                          action='store_true',
-                          help='Suppress verbosity.')
-    optional.add_argument('-h',
-                          '--help',
-                          action='help',
-                          help='Show this help message and exit.')
+    optional.add_argument(
+            '--only_official',
+            dest='only_official',
+            required=False,
+            action='store_true',
+            help=('Only output official rank names (i.e., superkingdom, '
+                'phylum, class, order, family, genus, species).'))
+    optional.add_argument(
+            '--exclude_scores',
+            dest='exclude_scores',
+            required=False,
+            action='store_true',
+            help=('Do not include bit-score support scores in the lineage of '
+                'a classification output file.'))
+    optional.add_argument(
+            '--force',
+            dest='force',
+            required=False,
+            action='store_true',
+            help='Force overwrite existing files.')
+    optional.add_argument(
+            '-q',
+            '--quiet',
+            dest='quiet',
+            required=False,
+            action='store_true',
+            help='Suppress verbosity.')
+    optional.add_argument(
+            '-h',
+            '--help',
+            action='help',
+            help='Show this help message and exit.')
 
     (args, extra_args) = parser.parse_known_args()
 
     extra_args = [arg for (i, arg) in enumerate(extra_args) if
                   (i, arg) != (0, 'add_names')]
     if len(extra_args) > 0:
-        sys.exit('error: too much arguments supplied:\n{0}'
-                 ''.format('\n'.join(extra_args)))
+        sys.exit('error: too much arguments supplied:\n{0}'.format(
+            '\n'.join(extra_args)))
+
+    # Add extra arguments.
+    shared.expand_arguments(args)
 
     return args
 
 
-def add_names(args):
-    (input_file,
-     output_file,
-     taxonomy_folder,
-     only_official,
-     exclude_scores,
-     force,
-     quiet) = check.convert_arguments(args)
+def run():
+    args = parse_arguments()
 
-    # Currently add_names does not allow for a log file.
-    log_file = None
-
     message = '# CAT v{0}.'.format(about.__version__)
-    shared.give_user_feedback(message, log_file, quiet, show_time=False)
+    shared.give_user_feedback(message, args.log_file, args.quiet,
+            show_time=False)
 
     errors = []
 
-    errors.append(check.check_input_file(input_file, log_file, quiet))
+    errors.append(
+            check.check_input_file(args.input_file, args.log_file, args.quiet))
 
-    if not force:
-        errors.append(check.check_output_file(output_file, log_file, quiet))
+    if not args.force:
+        errors.append(
+                check.check_output_file(
+                    args.output_file, args.log_file, args.quiet))
 
     if True in errors:
         sys.exit(1)
-
-    (nodes_dmp,
-     names_dmp,
-     prot_accession2taxid_file) = check.inspect_taxonomy_folder(taxonomy_folder)
 
-    (taxid2parent, taxid2rank) = tax.import_nodes(nodes_dmp, log_file, quiet)
-    taxid2name = tax.import_names(names_dmp, log_file, quiet)
+    (taxid2parent,
+            taxid2rank) = tax.import_nodes(
+                    args.nodes_dmp, args.log_file, args.quiet)
+    taxid2name = tax.import_names(args.names_dmp, args.log_file, args.quiet)
 
     message = 'Appending names...'
-    shared.give_user_feedback(message, log_file, quiet)
+    shared.give_user_feedback(message, args.log_file, args.quiet)
 
-    with open(input_file, 'r') as f1:
+    with open(args.input_file, 'r') as f1:
         for line in f1:
             if line.startswith('#'):
                 line = line.rstrip().split('\t')
@@ -128,11 +133,9 @@ def add_names(args):
                     lineage_index = line.index('lineage')
                 except:
                     message = ('{0} is not a supported classification file.'
-                               ''.format(input_file))
-                    shared.give_user_feedback(message,
-                                              log_file,
-                                              quiet,
-                                              error=True)
+                            ''.format(input_file))
+                    shared.give_user_feedback(
+                            message, args.log_file, args.quiet, error=True)
 
                     sys.exit(1)
 
@@ -145,20 +148,20 @@ def add_names(args):
 
                 break
         else:
-            message = ('{0} is not a supported classification file.'
-                       ''.format(input_file))
+            message = ('{0} is not a supported classification file.'.format(
+                args.input_file))
             shared.give_user_feedback(message, log_file, quiet, error=True)
 
             sys.exit(1)
 
-    with open(input_file, 'r') as f1, open(output_file, 'w') as outf1:
+    with open(args.input_file, 'r') as f1, open(args.output_file, 'w') as outf1:
         for line in f1:
             line = line.rstrip()
 
             if line.startswith('#'):
-                if only_official:
+                if args.only_official:
                     outf1.write('{0}\tsuperkingdom\tphylum\tclass\torder\t'
-                                'family\tgenus\tspecies\n'.format(line))
+                            'family\tgenus\tspecies\n'.format(line))
                 else:
                     outf1.write('{0}\tfull lineage names\n'.format(line))
 
@@ -173,7 +176,7 @@ def add_names(args):
                 continue
 
             if (line[1].startswith('no taxid found') or
-                line[2].startswith('no taxid found')):
+                    line[2].startswith('no taxid found')):
                 # ORF has database hits but the accession number is not found
                 # in the taxonomy files.
                 outf1.write('{0}\n'.format('\t'.join(line)))
@@ -182,34 +185,26 @@ def add_names(args):
 
             lineage = line[lineage_index].split(';')
 
-            if scores_index and not exclude_scores:
+            if scores_index and not args.exclude_scores:
                 scores = line[scores_index].split(';')
             else:
                 scores = None
 
-            if only_official:
-                names = tax.convert_to_official_names(lineage,
-                                                      taxid2rank,
-                                                      taxid2name,
-                                                      scores)
+            if args.only_official:
+                names = tax.convert_to_official_names(
+                        lineage, taxid2rank, taxid2name, scores)
             else:
-                names = tax.convert_to_names(lineage,
-                                             taxid2rank,
-                                             taxid2name,
-                                             scores)
+                names = tax.convert_to_names(
+                        lineage, taxid2rank, taxid2name, scores)
 
             outf1.write('{0}\t{1}\n'.format('\t'.join(line), '\t'.join(names)))
 
-    message = 'Names written to {0}!'.format(output_file)
-    shared.give_user_feedback(message, log_file, quiet)
-
-
-def run():
-    args = parse_arguments()
+    message = 'Names written to {0}!'.format(args.output_file)
+    shared.give_user_feedback(message, args.log_file, args.quiet)
+
+    return
+
 
-    add_names(args)
-
-
 if __name__ == '__main__':
-    sys.exit('Please run \'CAT add_names\' to add taxonomic names to CAT or '
-             'BAT output files.')
+    sys.exit('Run \'CAT add_names\' to add taxonomic names to CAT or BAT '
+            'output files.')