update NewChromSweep signature to support fisher. May have conflicts with...

update NewChromSweep signature to support fisher. May have conflicts with @nkindlon's upcoming changes. Will resolve then.

.gitignore

@@ -11,3 +11,4 @@ src/utils/version/version_git.h
 .project
 .cproject
 nbproject
+sandbox

src/utils/NewChromsweep/NewChromsweep.h

@@ -12,8 +12,6 @@
 #ifndef NEW_CHROMSWEEP_H
 #define NEW_CHROMSWEEP_H
 
-using namespace std;
-
 #include <string>
 #include "BTlist.h"
 #include "RecordKeyList.h"
@@ -31,7 +29,7 @@ class ContextIntersect;
 class NewChromSweep {
 public:
 
-    NewChromSweep(ContextIntersect *context);
+    NewChromSweep(ContextIntersect *context, bool useMergedIntervals = false);
     
     
     ~NewChromSweep(void);
@@ -60,7 +58,9 @@ protected:
     int _numDBs; //don't really need this stored, but here for code brevity.
     vector<FileRecordMgr *> _dbFRMs;
 
-     unsigned long _queryRecordsTotalLength;
+    bool _useMergedIntervals;
+
+    unsigned long _queryRecordsTotalLength;
     vector<unsigned long> _dbFileRecordsLength; //each value in this vector have the
     //length of all records in the corresponding db file.
 

test/fisher/a.bed

+chr1	10	20
+chr1	30	40
+chr1	51	52

test/fisher/b.bed

+chr1	15	25
+chr1	51	52

test/fisher/t.500.genome

+chr1	500

test/fisher/t.60.genome

+chr1	60

test/fisher/test-fisher.sh

+BT=${BT-../../bin/bedtools}
+
+check()
+{
+	if diff $1 $2; then
+    	echo ok
+		return 1
+	else
+    	echo fail
+		return 0
+	fi
+}
+
+echo "    fisher.t1...\c"
+echo \
+"# Number of query intervals: 3
+# Number of db intervals: 2
+# Number of overlaps: 2
+# Number of possible intervals (estimated): 34
+# phyper(2 - 1, 3, 34 - 3, 2, lower.tail=F)
+# Contingency Table Of Counts
+#_________________________________________
+#           |  in -b       | not in -b    |
+#     in -a | 2            | 1            |
+# not in -a | 0            | 31           |
+#_________________________________________
+# p-values for fisher's exact test
+left	right	two-tail	ratio
+1	0.0053476	0.0053476	inf" > exp
+$BT fisher -a a.bed -b b.bed -g t.500.genome > obs
+check obs exp
+rm obs exp
+
+
+echo "    fisher.t2...\c"
+echo \
+"# Number of query intervals: 3
+# Number of db intervals: 2
+# Number of overlaps: 2
+# Number of possible intervals (estimated): 4
+# phyper(2 - 1, 3, 4 - 3, 2, lower.tail=F)
+# Contingency Table Of Counts
+#_________________________________________
+#           |  in -b       | not in -b    |
+#     in -a | 2            | 1            |
+# not in -a | 0            | 1            |
+#_________________________________________
+# p-values for fisher's exact test
+left	right	two-tail	ratio
+1	0.5	1	inf" > exp
+$BT fisher -a a.bed -b b.bed -g t.60.genome > obs
+check obs exp
+rm obs exp

test/test.sh

@@ -22,6 +22,9 @@ cd expand; bash test-expand.sh; cd ..
 echo " Testing bedtools flank:"
 cd flank; bash test-flank.sh; cd ..
 
+echo " Testing bedtools fisher:"
+cd fisher; bash test-fisher.sh; cd ..
+
 echo " Testing bedtools genomecov:"
 cd genomecov; bash test-genomecov.sh; cd ..