Allow specifying a run instead of just scanning the runfolder for it.

Also rework some of the command line options to group the commands
together, and to print helpful error messages when runfolder is being used
incorrectly.

diff --git a/htsworkflow/pipelines/runfolder.py b/htsworkflow/pipelines/runfolder.py
index b8cbc56b1902aaf7b9d7f790b4dba71929c6d9c9..6b953ce2f53c65760458559fae9c8e2e491d8794 100644 (file)
--- a/htsworkflow/pipelines/runfolder.py
+++ b/htsworkflow/pipelines/runfolder.py
@@ -214,6 +214,58 @@ def get_runs(runfolder):
 
     return runs
 
+def get_specific_run(gerald_dir):
+    """
+    Given a gerald directory, construct a PipelineRun out of its parents
+
+    Basically this allows specifying a particular run instead of the previous
+    get_runs which scans a runfolder for various combinations of
+    firecrest/ipar/bustard/gerald runs.
+    """
+    from htsworkflow.pipelines import firecrest
+    from htsworkflow.pipelines import ipar
+    from htsworkflow.pipelines import bustard
+    from htsworkflow.pipelines import gerald
+
+    bustard_dir = os.path.abspath(os.path.join(gerald_dir, '..'))
+    image_dir = os.path.abspath(os.path.join(gerald_dir, '..', '..'))
+
+    runfolder_dir = os.path.abspath(os.path.join(image_dir, '..','..'))
+   
+    logging.debug('--- use-run detected options ---')
+    logging.debug('runfolder: %s' % (runfolder_dir,))
+    logging.debug('image_dir: %s' % (image_dir,))
+    logging.debug('bustard_dir: %s' % (bustard_dir,))
+    logging.debug('gerald_dir: %s' % (gerald_dir,))
+
+    # find our processed image dir
+    image_run = firecrest.firecrest(image_dir)
+    if image_run is None:
+        image_run = ipar.ipar(image_dir)
+    if image_run is None:
+        msg = '%s does not contain an image processing step' % (image_dir,)
+        logging.error(msg)
+        return None
+
+    # find our base calling
+    base_calling_run = bustard.bustard(bustard_dir)
+    if base_calling_run is None:
+        logging.error('%s does not contain a bustard run' % (bustard_dir,))
+        return None
+
+    # find alignments
+    gerald_run = gerald.gerald(gerald_dir)
+    if gerald_run is None:
+        logging.error('%s does not contain a gerald run' % (gerald_dir,))
+        return None
+
+    p = PipelineRun(runfolder_dir)
+    p.image_analysis = image_run
+    p.bustard = base_calling_run
+    p.gerald = gerald_run
+    
+    logging.info('Constructed PipelineRun from %s' % (gerald_dir,))
+    return p
 
 def extract_run_parameters(runs):
     """

diff --git a/scripts/runfolder b/scripts/runfolder
index bf5c5d884d2c931a6b96040a477b6bc0577ab30f..dcdaebefa1b64bd3cb174191bdfa11240cc22a52 100644 (file)
--- a/scripts/runfolder
+++ b/scripts/runfolder
@@ -39,21 +39,34 @@ from htsworkflow.pipelines.runfolder import ElementTree
 def make_parser():
     usage = 'usage: %prog [options] runfolder_root_dir'
     parser = optparse.OptionParser(usage)
+
     parser.add_option('-v', '--verbose', dest='verbose', action='store_true',
                       default=False,
                       help='turn on verbose mode')
-    parser.add_option('-s', '--summary', dest='summary', action='store_true',
-                      default=False,
-                      help='produce summary report')
-    parser.add_option('-a', '--archive', dest='archive', action='store_true',
-                      default=False,
-                      help='generate run configuration archive')
-    parser.add_option('--extract-results', action='store_true',
+    commands = optparse.OptionGroup(parser, 'Commands')
+
+    commands.add_option('-s', '--summary', dest='summary', action='store_true',
+                        default=False,
+                        help='produce summary report')
+    commands.add_option('-a', '--archive', dest='archive', action='store_true',
+                        default=False,
+                        help='generate run configuration archive')
+    commands.add_option('--extract-results', action='store_true',
            default=False,
-           help='extract result files out of runfolder into a simpler archive')
+           help='create run-xml summary, compress the eland result files, and '
+                'copy them and the Summary.htm file into archival directory.')
+    parser.add_option_group(commands)
+
     parser.add_option('-o', '--output-dir', default=None,
            help="specify the default output directory for extract results")
 
+    parser.add_option('-u', '--use-run', dest='use_run', default=None,
+                      help='Specify which run to use instead of autoscanning '
+                           'the runfolder. You do this by providing the final '
+                           ' GERALD directory, and it assumes the parent '
+                           'directories are the bustard and image processing '
+                           'directories.')
+                    
     parser.add_option('--run-xml', dest='run_xml',
            default=None,
            help='specify a run_<FlowCell>.xml file for summary reports')
@@ -77,18 +90,34 @@ def main(cmdlist=None):
         opt.run_xml = os.path.expanduser(opt.run_xml)
         tree = ElementTree.parse(opt.run_xml).getroot()
         runs.append(runfolder.PipelineRun(xml=tree))
+
+    # look for manually specified run
+    if opt.use_run is not None:
+        runs.append(runfolder.get_specific_run(opt.use_run))
+
+    # scan runfolders for runs
     for run_pattern in args:
         # expand args on our own if needed
         for run_dir in glob(run_pattern):
             runs.extend(runfolder.get_runs(run_dir))
 
     if len(runs) > 0:
+        command_run = False
         if opt.summary:
             print runfolder.summary_report(runs)
+            command_run = True
         if opt.archive:
             runfolder.extract_run_parameters(runs)
+            command_run = True
         if opt.extract_results:
             runfolder.extract_results(runs, opt.output_dir)
+            command_run = True
+        if command_run == False:
+            print "You need to specify a command."+os.linesep
+            parser.print_help()
+    else:
+        print "You need to specify some run folders to process..."+os.linesep
+        parser.print_help()
 
     return 0