Changeset 3987

Timestamp:

Aug 26, 2020, 6:42:10 AM (3 years ago)

Author:

Peter

Message:

merge release 0.17.2 into trunk

Location:

trunk

Files:

• . (modified) (1 prop)
• Makefile.am (modified) (2 diffs)
• NEWS (modified) (1 diff)
• build_support/Makefile.am (modified) (1 diff)
• doc/Makefile.am (modified) (2 diffs)
• doc/doxygen.config.in (modified) (1 diff)
• m4/version.m4 (modified) (1 diff)
• test/Makefile.am (modified) (3 diffs)
• test/Suite.h (modified) (2 diffs)
• test/data/kendall.txt (copied) (copied from branches/0.17-stable/test/data/kendall.txt)
• test/kendall.cc (modified) (6 diffs)
• test/vector.cc (modified) (1 diff)
• yat/statistics/Kendall.cc (modified) (2 diffs)
• yat/utility/VectorMutable.cc (modified) (2 diffs)

trunk/Makefile.am

@@ -8 +8 @@
 # Copyright (C) 2006 Jari Häkkinen
 # Copyright (C) 2007, 2008 Jari Häkkinen, Peter Johansson
-# Copyright (C) 2009, 2010, 2011, 2012, 2013, 2014, 2015 Peter Johansson
+# Copyright (C) 2009, 2010, 2011, 2012, 2013, 2014, 2015, 2020 Peter Johansson
 #
 # This file is part of the yat library, http://dev.thep.lu.se/yat
@@ -65 +65 @@
 
 clean-local:
-  rm -rf doc/$(DX_HTML_OUTPUT) test/yathello test/testSubDir test/lib
+  rm -rf doc/html test/yathello test/testSubDir test/lib
 
 installcheck-local: $(INSTALL_CHECK_TARGETS)

trunk/NEWS

@@ -25 +25 @@
 yat 0.17.x series from http://dev.thep.lu.se/yat/svn/branches/0.17-stable
 
-version 0.17.2 (released NOT YET)
+version 0.17.2 (released 26 August 2020)
+  - Fixed bug in Kendall::p_value() that caused overstimation when
+    both variables contained triplets (see bug #955)
+  - VectorMutable::all(double) now works on zero-sized vector, which
+    means that Vector::resize(0) now works on a zero-sized vector (bug
+    #956)
   - Generating documentation now works with Doxygen 1.8.19 (bug #954)
   - GSL, HTSLIB, Boost.Thread and Boost.System are not declated private

trunk/build_support/Makefile.am

@@ -1 +1 @@
 ## $Id$
 
-# Copyright (C) 2008, 2009, 2010, 2011, 2012, 2013, 2014, 2016, 2017, 2019 Peter Johansson
+# Copyright (C) 2008, 2009, 2010, 2011, 2012, 2013, 2014, 2016, 2017, 2019, 2020 Peter Johansson
 #
 # This file is part of the yat library, http://dev.thep.lu.se/yat

trunk/doc/Makefile.am

@@ -4 +4 @@
 # Copyright (C) 2005 Peter Johansson
 # Copyright (C) 2006, 2007, 2008 Jari Häkkinen, Peter Johansson
-# Copyright (C) 2009, 2010, 2011, 2012 Peter Johansson
+# Copyright (C) 2009, 2010, 2011, 2012, 2020 Peter Johansson
 #
 # This file is part of the yat library, http://dev.thep.lu.se/yat
@@ -88 +88 @@
   || { echo "no documentation installed" 1>&2; exit 1; }
 
+MOSTLYCLEANFILES += doc/doxygen.config
 MOSTLYCLEANFILES += doc/doxygen.error

trunk/doc/doxygen.config.in

@@ -5 +5 @@
 # Copyright (C) 2006, 2007 Jari Häkkinen, Peter Johansson
 # Copyright (C) 2008 Jari Häkkinen, Peter Johansson, Markus Ringnér
-# Copyright (C) 2009, 2010, 2012, 2014, 2017, 2018 Peter Johansson
+# Copyright (C) 2009, 2010, 2012, 2014, 2017, 2018, 2020 Peter Johansson
 #
 # This file is part of the yat library, http://dev.thep.lu.se/yat

trunk/m4/version.m4

@@ -94 +94 @@
 # yat-0.17   14:0:0
 # yat-0.17.1 14:1:0
+# yat-0.17.2 14:2:0
 #
 # *Accidently, the libtool number was not updated for yat 0.5

trunk/test/Makefile.am

@@ -5 +5 @@
 # Copyright (C) 2005 Jari Häkkinen, Peter Johansson
 # Copyright (C) 2006, 2007, 2008 Jari Häkkinen, Peter Johansson, Markus Ringnér
-# Copyright (C) 2009, 2010, 2011, 2012, 2013, 2014, 2015, 2016, 2017, 2018, 2019 Peter Johansson
+# Copyright (C) 2009, 2010, 2011, 2012, 2013, 2014, 2015, 2016, 2017, 2018, 2019, 2020 Peter Johansson
 #
 # This file is part of the yat library, http://dev.thep.lu.se/yat
@@ -188 +188 @@
 test/cmd_test.log: test/help.test $(shell_test_deps)
 test/help_test.log: test/help.test $(shell_test_deps)
+test/kendall.log: $(top_srcdir)/test/data/kendall.txt
 test/pileup.log: test/data/foo.sorted.bam
 test/static_test.log: $(srcdir)/m4/yat.m4 $(shell_test_deps)
@@ -312 +313 @@
 test/data/foo.vcf \
 test/data/isoform.peaks \
+test/data/kendall.txt \
 test/data/knni_matrix.data \
 test/data/knni_result.data \

trunk/test/Suite.h

@@ -6 +6 @@
 /*
   Copyright (C) 2008 Jari Häkkinen, Peter Johansson
-  Copyright (C) 2009, 2010, 2011, 2013, 2014, 2015, 2016, 2018, 2019 Peter Johansson
+  Copyright (C) 2009, 2010, 2011, 2013, 2014, 2015, 2016, 2018, 2019, 2020 Peter Johansson
 
   This file is part of the yat library, http://dev.thep.lu.se/yat
@@ -191 +191 @@
 
   /**
-     \return absolute path to test src dir
-  */
-  std::string abs_srcdir(void);
-
-  /**
+     If top source directory is located in /home/strangelove/yat,
+     function returns "/home/strangelove/yat" + "test/" + local_path
+
      \return absolute path to file
      \param local_path path relative to srcdir

trunk/test/kendall.cc

@@ -2 +2 @@
 
 /*
-  Copyright (C) 2011, 2012 Peter Johansson
+  Copyright (C) 2011, 2012, 2020 Peter Johansson
 
   This file is part of the yat library, http://dev.thep.lu.se/yat
@@ -24 +24 @@
 #include "Suite.h"
 
+#include "yat/random/random.h"
 #include "yat/statistics/Kendall.h"
+#include "yat/utility/Matrix.h"
 
 #include <gsl/gsl_cdf.h>
 
 #include <cmath>
+#include <fstream>
 
 using namespace theplu::yat;
@@ -37 +40 @@
 void test_p(test::Suite&);
 void test_p_exact(test::Suite&);
+void test_p_with_ties(test::Suite&);
 void test_score(test::Suite&);
 void test_score_with_ties(test::Suite&);
@@ -48 +52 @@
   test_p(suite);
   test_p_exact(suite);
+  test_p_with_ties(suite);
   test_score(suite);
   test_score_with_ties(suite);
@@ -110 +115 @@
   kendall.add(8,1.5);
   suite.add(suite.equal_fix(0.031, kendall.p_right(true), 0.0005));
+}
+
+
+void test_p_with_ties(test::Suite& suite)
+{
+  suite.out() << YAT_TEST_PROLOGUE;
+  std::string fn = test::filename("data/kendall.txt");
+  std::ifstream is(fn.c_str());
+  if (!is.good()) {
+    suite.add(false);
+    suite.err() << "failed open '" << fn << "'\n";
+    return;
+  }
+  utility::Matrix x(is);
+  is.close();
+  Kendall kendall;
+  for (size_t i=0; i<x.rows(); ++i)
+    kendall.add(x(i,0), x(i, 1));
+  double score = kendall.score();
+  suite.out() << "score: " << score << "\n";
+  double p = kendall.p_value();
+  suite.out() << "p: " << p << "\n";
+  if (p > 0.0001) {
+    suite.err() << "error: too large p\n";
+    suite.add(false);
+  }
+  if (p < 0.00001) {
+    suite.err() << "error: too small p\n";
+    suite.add(false);
+  }
 }
 
@@ -158 +193 @@
   double vu = 3*2*11 + 2*1*9;
   double v1 = (6+6)*(6+2)/(2.0*n*(n-1));
-  double v2 = (3*2*1+3*2*1)*(3*2*1);
+  double v2 = (3*2*1+3*2*1)*(3*2*1) / (9.0*n*(n-1)*(n-2));
   double v = (v0-vt-vu)/18 + v1 + v2;
   double z = (nc - nd)/std::sqrt(v);

trunk/test/vector.cc

@@ -354 +354 @@
     theplu::yat::utility::Vector vec(0, 1.2);
     theplu::yat::utility::Vector vec2(vec);
+    // test ticket #956
+    vec.resize(0);
   }
 

trunk/yat/statistics/Kendall.cc

@@ -2 +2 @@
 
 /*
-  Copyright (C) 2011, 2012 Peter Johansson
+  Copyright (C) 2011, 2012, 2020 Peter Johansson
 
   This file is part of the yat library, http://dev.thep.lu.se/yat
@@ -102 +102 @@
       // avoid overflow
       double n = this->n();
-      double var = n*(n-1)*(2*n+5)/18;
-      if (have_ties()) {
-        var -= (v_correction(x_count_)+v_correction(y_count_))/18.0;
-        double v1 = count_pair(x_count_)*count_pair(y_count_)/(2*n*(n-1));
-        if (v1) {
-          var += v1;
-          // v_corr3 is zero if v1=0
-          var += count_triple(x_count_)*count_triple(y_count_);
+      /*
+        According to wikipedia,
+        z = k / sqrt(v)
+        is approximately standard normal
+        v = (v0 - vt - vu)/18 + v1 + v2
+        v0 = n(n-1)(2n+5)
+        vt = \sum t(t-1)(2t+5)
+        vu = \sum u(u-1)(2u+5)
+        v1 = \sum t(t-1)) * \sum u(u-1) / (2n(n-1))
+        v2 = sum t(t-1)(t-2) \sum u(u-1)(u-2) / (9n(n-1)(n-2))
+
+        where t is number of equal values in group i
+        and similarly u for y.
+       */
+      double v0 = n*(n-1)*(2*n+5);
+      double vt = 0;
+      double vu = 0;
+      double v1 = 0;
+      double v2 = 0;
+      // all correction terms above are zero in absence of ties
+      bool x_have_ties = x_.size() != x_count_.size();
+      bool y_have_ties = y_.size() != y_count_.size();
+      if (x_have_ties || y_have_ties) {
+        if (x_have_ties)
+          vt = v_correction(x_count_);
+        if (y_have_ties)
+          vu = v_correction(y_count_);
+        if (x_have_ties && y_have_ties) {
+          v1 = count_pair(x_count_) * (count_pair(y_count_) / (2*n*(n-1)));
+          v2 = count_triple(x_count_);
+          if (v2)
+            v2 *= count_triple(y_count_) / (9*n*(n-1)*(n-2));
         }
       }
-      return gsl_cdf_gaussian_Q(k, std::sqrt(var));
+      double v = (v0 - vt - vu)/18 + v1 + v2;
+      return gsl_cdf_gaussian_Q(k, std::sqrt(v));
     }
 

trunk/yat/utility/VectorMutable.cc

@@ -6 +6 @@
   Copyright (C) 2005, 2006, 2007 Jari Häkkinen, Peter Johansson, Markus Ringnér
   Copyright (C) 2008, 2009 Jari Häkkinen, Peter Johansson
-  Copyright (C) 2010, 2012, 2016, 2017, 2018 Peter Johansson
+  Copyright (C) 2010, 2012, 2016, 2017, 2018, 2020 Peter Johansson
 
   This file is part of the yat library, http://dev.thep.lu.se/trac/yat
@@ -69 +69 @@
   void VectorMutable::all(double value)
   {
-    assert(vec_);
-    gsl_vector_set_all(vec_,value);
+    if (vec_)
+      gsl_vector_set_all(vec_,value);
   }