gaworkflow/automation/runner.py

   1 #!/usr/bin/env python
   2 from glob import glob
   3 import logging
   4 import os
   5 import re
   6 import sys
   7 import time
   8 import threading
   9
  10 from benderjab import rpc
  11
  12 from gaworkflow.pipeline.configure_run import *
  13
  14 #s_fc = re.compile('FC[0-9]+')
  15 s_fc = re.compile('_[0-9a-zA-Z]*$')
  16
  17
  18 def _get_flowcell_from_rundir(run_dir):
  19     """
  20     Returns flowcell string based on run_dir.
  21     Returns None and logs error if flowcell can't be found.
  22     """
  23     junk, dirname = os.path.split(run_dir)
  24     mo = s_fc.search(dirname)
  25     if not mo:
  26         logging.error('RunDir 2 FlowCell error: %s' % (run_dir))
  27         return None
  28
  29     return dirname[mo.start()+1:]
  30
  31
  32
  33 class Runner(rpc.XmlRpcBot):
  34     """
  35     Manage running pipeline jobs.
  36     """
  37     def __init__(self, section=None, configfile=None):
  38         #if configfile is None:
  39         #    self.configfile = "~/.gaworkflow"
  40         super(Runner, self).__init__(section, configfile)
  41
  42         self.cfg['notify_users'] = None
  43         self.cfg['genome_dir'] = None
  44         self.cfg['base_analysis_dir'] = None
  45
  46         self.cfg['notify_users'] = None
  47         self.cfg['notify_postanalysis'] = None
  48
  49         self.conf_info_dict = {}
  50
  51         self.register_function(self.sequencingFinished)
  52         #self.eventTasks.append(self.update)
  53
  54
  55     def read_config(self, section=None, configfile=None):
  56         super(Runner, self).read_config(section, configfile)
  57
  58         self.genome_dir = self._check_required_option('genome_dir')
  59         self.base_analysis_dir = self._check_required_option('base_analysis_dir')
  60
  61         self.notify_users = self._parse_user_list(self.cfg['notify_users'])
  62         #FIXME: process notify_postpipeline cfg
  63
  64
  65     def _parser(self, msg, who):
  66         """
  67         Parse xmpp chat messages
  68         """
  69         help = u"I can send [start] a run, or report [status]"
  70         if re.match(u"help", msg):
  71             reply = help
  72         elif re.match("status", msg):
  73             words = msg.split()
  74             if len(words) == 2:
  75                 reply = self.getStatusReport(words[1])
  76             else:
  77                 reply = u"Status available for: %s" \
  78                         % (', '.join([k for k in self.conf_info_dict.keys()]))
  79         elif re.match(u"start", msg):
  80             words = msg.split()
  81             if len(words) == 2:
  82                 self.sequencingFinished(words[1])
  83                 reply = u"starting run for %s" % (words[1])
  84             else:
  85                 reply = u"need runfolder name"
  86         else:
  87             reply = u"I didn't understand '%s'" %(msg)
  88
  89         logging.debug("reply: " + str(reply))
  90         return reply
  91
  92
  93     def getStatusReport(self, fc_num):
  94         """
  95         Returns text status report for flow cell number
  96         """
  97         if fc_num not in self.conf_info_dict:
  98             return "No record of a %s run." % (fc_num)
  99
 100         status = self.conf_info_dict[fc_num].status
 101
 102         if status is None:
 103             return "No status information for %s yet." \
 104                    " Probably still in configure step. Try again later." % (fc_num)
 105
 106         output = status.statusReport()
 107
 108         return '\n'.join(output)
 109
 110
 111     def sequencingFinished(self, run_dir):
 112         """
 113         Sequenceing (and copying) is finished, time to start pipeline
 114         """
 115         logging.debug("received sequencing finished message")
 116
 117         # Setup config info object
 118         ci = ConfigInfo()
 119         ci.base_analysis_dir = self.base_analysis_dir
 120         ci.analysis_dir = os.path.join(self.base_analysis_dir, run_dir)
 121
 122         # get flowcell from run_dir name
 123         flowcell = _get_flowcell_from_rundir(run_dir)
 124
 125         # Store ci object in dictionary
 126         self.conf_info_dict[flowcell] = ci
 127
 128
 129         # Launch the job in it's own thread and turn.
 130         self.launchJob(run_dir, flowcell, ci)
 131         return "started"
 132
 133
 134     def pipelineFinished(self, run_dir):
 135         # need to strip off self.watch_dir from rundir I suspect.
 136         logging.info("pipeline finished in" + str(run_dir))
 137         #pattern = self.watch_dir
 138         #if pattern[-1] != os.path.sep:
 139         #    pattern += os.path.sep
 140         #stripped_run_dir = re.sub(pattern, "", run_dir)
 141         #logging.debug("stripped to " + stripped_run_dir)
 142
 143         # Notify each user that the run has finished.
 144         if self.notify_users is not None:
 145             for u in self.notify_users:
 146                 self.send(u, 'Pipeline run %s finished' % (run_dir))
 147
 148         #if self.notify_runner is not None:
 149         #    for r in self.notify_runner:
 150         #        self.rpc_send(r, (stripped_run_dir,), 'sequencingFinished')
 151
 152     def reportMsg(self, msg):
 153
 154         if self.notify_users is not None:
 155             for u in self.notify_users:
 156                 self.send(u, msg)
 157
 158
 159     def _runner(self, run_dir, flowcell, conf_info):
 160
 161         # retrieve config step
 162         cfg_filepath = os.path.join(conf_info.analysis_dir,
 163                                     'config32auto.txt')
 164         status_retrieve_cfg = retrieve_config(conf_info,
 165                                           flowcell,
 166                                           cfg_filepath,
 167                                           self.genome_dir)
 168         if status_retrieve_cfg:
 169             logging.info("Runner: Retrieve config: success")
 170             self.reportMsg("Retrieve config (%s): success" % (run_dir))
 171         else:
 172             logging.error("Runner: Retrieve config: failed")
 173             self.reportMsg("Retrieve config (%s): FAILED" % (run_dir))
 174
 175
 176         # configure step
 177         if status_retrieve_cfg:
 178             status = configure(conf_info)
 179             if status:
 180                 logging.info("Runner: Configure: success")
 181                 self.reportMsg("Configure (%s): success" % (run_dir))
 182                 self.reportMsg(
 183                     os.linesep.join(glob(os.path.join(run_dir,'Data','C*')))
 184                 )
 185             else:
 186                 logging.error("Runner: Configure: failed")
 187                 self.reportMsg("Configure (%s): FAILED" % (run_dir))
 188
 189             #if successful, continue
 190             if status:
 191                 # Setup status cmdline status monitor
 192                 #startCmdLineStatusMonitor(ci)
 193
 194                 # running step
 195                 print 'Running pipeline now!'
 196                 run_status = run_pipeline(conf_info)
 197                 if run_status is True:
 198                     logging.info('Runner: Pipeline: success')
 199                     self.reportMsg("Pipeline run (%s): Finished" % (run_dir,))
 200                 else:
 201                     logging.info('Runner: Pipeline: failed')
 202                     self.reportMsg("Pipeline run (%s): FAILED" % (run_dir))
 203
 204
 205     def launchJob(self, run_dir, flowcell, conf_info):
 206         """
 207         Starts up a thread for running the pipeline
 208         """
 209         t = threading.Thread(target=self._runner,
 210                         args=[run_dir, flowcell, conf_info])
 211         t.setDaemon(True)
 212         t.start()
 213
 214
 215
 216 def main(args=None):
 217     bot = Runner()
 218     return bot.main(args)
 219
 220 if __name__ == "__main__":
 221     sys.exit(main(sys.argv[1:]))
 222