Changeset 489 for library

Timestamp:

08/08/09 13:43:16 (15 years ago)

Author:

smidl

Message:

Mpdf redesing is complet - all tests pass (except mlnorm<chmat> which is irrelevant)

Location:

library

Files:

• bdm/base/bdmbase.cpp (modified) (1 diff)
• bdm/base/bdmbase.h (modified) (5 diffs)
• tests/epdf_harness.cpp (modified) (1 diff)
• tests/mepdf.cfg (modified) (1 diff)
• tests/mgamma.cfg (modified) (1 diff)
• tests/mlnorm.cfg (modified) (1 diff)

library/bdm/base/bdmbase.cpp

@@ -273 +273 @@
 }
 
-void mepdf::condition ( const vec &cond ) {
-}
-
 void mepdf::from_setting ( const Setting &set ) {
         shared_ptr<epdf> e ( UI::build<epdf> ( set, "epdf", UI::compulsory ) );
-        ipdf = e;
+        iepdf = e;
+        set_ep(iepdf.get());
 }
 

library/bdm/base/bdmbase.h

@@ -419 +419 @@
                         ep = &iepdf;
                 }
+                void set_ep (epdf *iepdfp) {
+                        ep = iepdfp;
+                }
 
         public:
@@ -436 +439 @@
 
                 //! Returns \param N samples from the density conditioned on \c cond, \f$x \sim epdf(rv|cond)\f$. \param cond is numeric value of \c rv
-                virtual mat samplecond_m (const vec &cond, int N) {it_error ("Not implemented");return mat();}
+                virtual mat samplecond_m (const vec &cond, int N) {
+                        mat M(dimension(), N);
+                        for(int i=0;i<N;i++){M.set_col(i, samplecond(cond));}
+                        return M;
+                }
 
 
@@ -443 +450 @@
 
                 //! Matrix version of evallogcond
-                virtual vec evallogcond_m (const mat &Dt, const vec &cond) {it_error ("Not implemented");return vec (0);}
+                virtual vec evallogcond_m (const mat &Dt, const vec &cond) {
+                        vec v(Dt.cols());
+                        for(int i=0;i<Dt.cols();i++){v(i)= evallogcond(Dt.get_col(i),cond);}
+                        return v;
+                }
 
                 //! Array<vec> version of evallogcond
@@ -718 +729 @@
 {
 
-                shared_ptr<epdf> ipdf;
+                shared_ptr<epdf> iepdf;
         public:
                 //!Default constructor
@@ -724 +735 @@
 
                 mepdf (shared_ptr<epdf> em) {
-                        ipdf = em;
-                        set_ep (*ipdf.get());
+                        iepdf = em;
+                        set_ep (*iepdf.get());
                         dimc = 0;
                 }
 
                 //! empty
-                void condition (const vec &cond);
+                vec samplecond(const vec &cond){return iepdf->sample();}
+                double evallogcond(const vec &val, const vec &cond){return iepdf->evallog(val);}
 
                 //! Load from structure with elements:

library/tests/epdf_harness.cpp

@@ -175 +175 @@
         } while ( ( tc < CurrentContext::max_trial_count ) &&
                   !UnitTest::AreClose ( R, actual( tc - 1 ), tolerance ) );
+        
         if ( ( tc == CurrentContext::max_trial_count ) &&
              ( !UnitTest::AreClose ( R, actual( CurrentContext::max_trial_count - 1 ), tolerance ) ) ) {

library/tests/mepdf.cfg

@@ -18 +18 @@
   mean = [ 1.1, -1.0 ];
   R = ( "matrix", 2, 2, [ 1.0, -0.5, -0.5, 2.0 ] );
+  variance = [ 1.0, 2.0 ];
   tolerance = 0.1;
-  variance = [ 0.3, 0.3 ];
 } );
 

library/tests/mgamma.cfg

@@ -20 +20 @@
   mean = [ 1.5, 1.7 ];
   R = ( "matrix", 2, 2, [ 2.25, 0.0, 0.0, 2.89 ] );
-  variance = [ 0.4, 0.4 ];
+  variance = [ 2.25, 2.89];
   tolerance = 0.4;
 } );

library/tests/mlnorm.cfg

@@ -10 +10 @@
   cond = [ 1.5, 1.7 ];
   mean = [ 1.5, 1.7 ];
-  R = ( "matrix", 2, 2, [ 1.26402, 0.378231, 0.378231, 5.15925 ] );
+  R = ( "matrix", 2, 2, [ 1.2, 0.3, 0.3, 5.0 ] );
+  variance = [ 1.2, 5.0 ];
   tolerance = 0.6;
-  variance = [ 0.2, 0.2 ];
-} );
+},
+{
+  class = "mpdf_harness";
+  mpdf = {
+    class = "mlnorm<chmat>";
+    A = ( "matrix", 2, 2, [ 1.0, 0.0, 0.0, 1.0 ] );
+    const = [ 0, 0 ];
+    R = ( "matrix", 2, 2, [ 1.2, 0.3, 0.3, 5.0 ] );
+  };
+  cond = [ 1.5, 1.7 ];
+  mean = [ 1.5, 1.7 ];
+  R = ( "matrix", 2, 2, [ 1.2, 0.3, 0.3, 5.0 ] );
+  variance = [ 1.2, 5.0 ];
+  tolerance = 0.6;
+});