reductions now have predict functions, but tests break

1 files changed, 45 insertions, 32 deletions

diff --git a/vowpalwabbit/csoaa.cc b/vowpalwabbit/csoaa.cc
index 8b207f0b..d8781ba2 100644
--- a/vowpalwabbit/csoaa.cc
+++ b/vowpalwabbit/csoaa.cc
@@ -311,33 +311,39 @@ namespace CSOAA {
     print_update(all, is_test_label((label*)ec->ld), ec);
   }
 
-  void learn(void* d, learner& base, example* ec) {
+  template <bool is_learn>
+  void predict_or_learn(void* d, learner& base, example* ec) {
     csoaa* c = (csoaa*)d;
     vw* all = c->all;
     label* ld = (label*)ec->ld;
 
     size_t prediction = 1;
     float score = FLT_MAX;
+    label_data simple_temp;
+    simple_temp.initial = 0.;
+    ec->ld = &simple_temp;
     for (wclass *cl = ld->costs.begin; cl != ld->costs.end; cl ++)
       {
         uint32_t i = cl->weight_index;
-	label_data simple_temp;
-	simple_temp.initial = 0.;
-	
-	if (cl->x == FLT_MAX || !all->training)
+
+	if (is_learn)
 	  {
-	    simple_temp.label = FLT_MAX;
-	    simple_temp.weight = 0.;
+	    if (cl->x == FLT_MAX || !all->training)
+	      {
+		simple_temp.label = FLT_MAX;
+		simple_temp.weight = 0.;
+	      }
+	    else
+	      {
+		simple_temp.label = cl->x;
+		simple_temp.weight = 1.;
+	      }
+
+	    base.learn(ec);
 	  }
 	else
-	  {
-	    simple_temp.label = cl->x;
-	    simple_temp.weight = 1.;
-	  }
-
-	ec->ld = &simple_temp;
+	  base.predict(ec);
 
-	base.learn(ec, i - 1);
         cl->partial_prediction = ec->partial_prediction;
 	if (ec->partial_prediction < score || (ec->partial_prediction == score && i < prediction)) {
           score = ec->partial_prediction;
@@ -378,7 +384,7 @@ namespace CSOAA {
     *(all.p->lp) = cs_label_parser;
     all.sd->k = nb_actions;
 
-    learner* l = new learner(c, learn, all.l, nb_actions);
+    learner* l = new learner(c, predict_or_learn<true>, predict_or_learn<false>, all.l, nb_actions);
     l->set_finish_example(finish_example);
     return l;
   }
@@ -623,16 +629,15 @@ namespace LabelDict {
     v_array<CSOAA::wclass> costs = ld->costs;
     label_data simple_label;
 
+    simple_label.initial = 0.;
+    simple_label.label = FLT_MAX;
+    simple_label.weight = 0.;
     for (size_t j=0; j<costs.size(); j++) {
-      simple_label.initial = 0.;
-      simple_label.label = FLT_MAX;
-      simple_label.weight = 0.;
-      ec->partial_prediction = 0.;
 
       LabelDict::add_example_namespace_from_memory(l, ec, costs[j].weight_index);
-      
+
       ec->ld = &simple_label;
-      base.learn(ec); // make a prediction
+      base.predict(ec); // make a prediction
       costs[j].partial_prediction = ec->partial_prediction;
 
       if (ec->partial_prediction < *min_score) {
@@ -650,7 +655,7 @@ namespace LabelDict {
   }
 
 
-
+  template <bool is_learn>
   void do_actual_learning_wap(vw& all, ldf& l, learner& base, size_t start_K)
   {
     size_t K = l.ec_seq.size();
@@ -721,7 +726,10 @@ namespace LabelDict {
               simple_label.weight = value_diff;
               ec1->partial_prediction = 0.;
               subtract_example(all, ec1, ec2);
-              base.learn(ec1);
+	      if (is_learn)
+		base.learn(ec1);
+	      else
+		base.predict(ec1);
               unsubtract_example(all, ec1);
               
               LabelDict::del_example_namespace_from_memory(l, ec2, costs2[j2].weight_index);
@@ -738,6 +746,7 @@ namespace LabelDict {
     }
   }
 
+  template <bool is_learn>
   void do_actual_learning_oaa(vw& all, ldf& l, learner& base, size_t start_K)
   {
     size_t K = l.ec_seq.size();
@@ -796,7 +805,10 @@ namespace LabelDict {
           //cerr << "[" << ec->partial_prediction << "," << ec->done << "]";
           //ec->done = false;
           LabelDict::add_example_namespace_from_memory(l, ec, costs[j].weight_index);
-          base.learn(ec);
+	  if (is_learn)
+	    base.learn(ec);
+	  else
+	    base.predict(ec);
           LabelDict::del_example_namespace_from_memory(l, ec, costs[j].weight_index);
           ec->example_t = example_t;
         }
@@ -816,7 +828,7 @@ namespace LabelDict {
     }
   }
 
-
+  template <bool is_learn>
   void do_actual_learning(vw& all, ldf& l, learner& base)
   {
     if (l.ec_seq.size() <= 0) return;  // nothing to do
@@ -848,8 +860,8 @@ namespace LabelDict {
     }
 
     /////////////////////// learn
-    if (l.is_wap) do_actual_learning_wap(all, l, base, start_K);
-    else          do_actual_learning_oaa(all, l, base, start_K);
+    if (l.is_wap) do_actual_learning_wap<is_learn>(all, l, base, start_K);
+    else          do_actual_learning_oaa<is_learn>(all, l, base, start_K);
     
     /////////////////////// remove header
     if (start_K > 0)
@@ -934,7 +946,8 @@ namespace LabelDict {
     l->first_pass = false;
   }
 
-  void learn(void* data, learner& base, example *ec) 
+  template <bool is_learn>
+  void predict_or_learn(void* data, learner& base, example *ec) 
   {
     ldf* l=(ldf*)data;
     vw* all = l->all;
@@ -951,7 +964,7 @@ namespace LabelDict {
       if (l->ec_seq.size() >= all->p->ring_size - 2 && l->first_pass)
         cerr << "warning: length of sequence at " << ec->example_counter << " exceeds ring size; breaking apart" << endl;
 	
-      do_actual_learning(*all, *l, base);
+      do_actual_learning<is_learn>(*all, *l, base);
 
       if (!LabelDict::ec_seq_is_label_definition(*l, l->ec_seq) && l->ec_seq.size() > 0)
         global_print_newline(*all);
@@ -966,7 +979,7 @@ namespace LabelDict {
       if (!((!all->training) || CSOAA::example_is_test(ec))) {
         l->ec_seq.erase();
         l->ec_seq.push_back(ec);
-        do_actual_learning(*all, *l, base);
+        do_actual_learning<is_learn>(*all, *l, base);
         l->ec_seq.erase();
       }
 
@@ -1018,7 +1031,7 @@ namespace LabelDict {
   {
     ldf* l=(ldf*)data;
     vw* all = l->all;
-    do_actual_learning(*all, *l, *(l->base));
+    do_actual_learning<true>(*all, *l, *(l->base));
     output_example_seq(*all, *l);
     clear_seq(*all, *l);
     l->ec_seq.delete_v();
@@ -1096,7 +1109,7 @@ namespace LabelDict {
 
     ld->read_example_this_loop = 0;
     ld->need_to_clear = false;
-    learner* l = new learner(ld, learn, all.l);
+    learner* l = new learner(ld, predict_or_learn<true>, predict_or_learn<false>, all.l);
     if (ld->is_singleline)
       l->set_finish_example(finish_example);
     else