+ first version of a script that uses pythongrid

git-svn-id: http://svn.tuebingen.mpg.de/ag-raetsch/projects/QPalma@9001 e1793c9e-67f9-0310-80fc-b846ff1f7b36

diff --git a/scripts/grid_predict.py b/scripts/grid_predict.py
new file mode 100644 (file)
index 0000000..7dffb16
--- /dev/null
+++ b/scripts/grid_predict.py
@@ -0,0 +1,105 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+
+import cPickle
+import sys
+import pdb
+import os
+import os.path
+import math
+
+import pythongrid
+
+from qpalma_main import *
+
+
+def get_slices(dataset_size,num_nodes):
+   all_instances = []
+
+   part = dataset_size / num_nodes
+   begin = 0
+   end = 0
+   for idx in range(1,num_nodes+1):
+      
+      if idx == num_nodes:
+         begin = end
+         end   = dataset_size
+      else:
+         begin = end
+         end = begin+part
+
+      params = (begin,end)
+
+      all_instances.append(params)
+
+   return all_instances
+
+
+def makeJobs(run,dataset_fn,chunks,param):
+   """
+   """
+
+   jobs=[]
+
+   for current_chunk in chunks:
+      current_job = KybJob(predict,[run,prediction_set,param])
+      current_job.h_vmem = '5.0G'
+      current_job.express = 'True'
+
+      print "job #1: ", j1.nativeSpecification
+
+      jobs.append(j1)
+
+  return jobs
+
+
+def create_and_submit():
+   """
+
+   """
+
+   jp = os.path.join
+
+   run_dir = '/fml/ag-raetsch/home/fabio/tmp/newest_run/alignment/run_enable_quality_scores_+_enable_splice_signals_+_enable_intron_length_+'
+
+   run   = cPickle.load(jp(run_dir,'run_obj.pickle'))
+   param = cPickle.load(jp(run_dir,'param_526.pickle'))
+
+   dataset_fn        = ''
+   prediction_keys   = ''
+
+   num_splits = 10
+   slices = get_slices(prediction_keys,num_splits)
+   chunks = []
+   for slice in slices:
+      chunks.append(prediction_keys[slice[0]:slice[1]])
+
+
+   functionJobs = makeJobs(run,dataset_fn,chunks,param)
+
+   print "output ret field in each job before sending it onto the cluster"
+   for (i, job) in enumerate(functionJobs):
+      print "Job with id: ", i, "- ret: ", job.ret
+
+   print ""
+   print "sending function jobs to cluster"
+   print ""
+
+   #processedFunctionJobs = processJobs(functionJobs)
+
+   print "ret fields AFTER execution on cluster"
+   for (i, job) in enumerate(processedFunctionJobs):
+      print "Job with id: ", i, "- ret: ", job.ret
+
+
+def predict(run,prediction_set,param):
+   """
+   
+   """
+
+   qp = QPalma()
+   qp.predict(run,dataset_fn,prediction_keys,param):
+
+
+if __name__ == '__main__':
+   create_and_submit():