diff kmersvm/scripts/nullseq_generate.py @ 7:fd740d515502 draft default tip

Uploaded revised kmer-SVM to include modules from kmer-visual.

--- a/kmersvm/scripts/nullseq_generate.py	Mon Aug 20 21:42:29 2012 -0400
+++ b/kmersvm/scripts/nullseq_generate.py	Sun Jun 16 18:06:14 2013 -0400
@@ -71,8 +71,7 @@
 def sample_sequences(positions, buildname, basedir, options):
 	"""
 	"""
-	rpt_err = options.rpt_err
-	gc_err = options.gc_err
+	max_fails = 20
 	max_trys = options.max_trys
 	norpt = options.norpt
 	nogc = options.nogc
@@ -121,6 +120,12 @@
 		else:
 			count = options.count
 
+		#initialize paramter
+		#added by dlee 2/17/13	
+		ncfails = 0
+		rpt_err = options.rpt_err
+		gc_err = options.gc_err
+
 		sampled_positions = []
 		while len(sampled_positions) < count:
 			sampled_prof = random.choice(profiles)
@@ -128,6 +133,15 @@
 			sampled_gc = sampled_prof[2]
 			sampled_rpt = sampled_prof[3]
 
+			#relax rpt_err and gc_err if it keep fail to sample a region
+			#added by dlee 2/17/13	
+			if ncfails >= max_fails:
+				if options.quiet == False:
+					sys.stderr.write("reached max_fail. relax gc and rpt err criteria\n")
+				ncfails = 0
+				rpt_err += 0.01
+				gc_err += 0.01
+
 			rpt_err_allowed = int(rpt_err*sampled_len)
 			gc_err_allowed = int(gc_err*sampled_len)
 			trys = 0
@@ -156,9 +170,17 @@
 
 				sampled_positions.append((chrom, pos, pos_e))
 
+				#reset the counter of consecutive fails
+				#added by dlee 2/17/13	
+				ncfails = 0
+
 				#print trys, chrom, pos, pos_e, sampled_len, pos_rpt, sampled_rpt, pos_gc, sampled_gc
 				break
 			else:
+				#increase the counter of consecutive fails
+				#added by dlee 2/17/13	
+				ncfails += 1
+
 				if options.quiet == False:
 					sys.stderr.write(' '.join(["fail to sample from", \
 							"len=", str(sampled_len), \