Merge pull request #110 from sashafrey/master

#109 and ThetaMatrix.item_title
bigartm · Jan 26, 2015 · 8416560 · 8416560
2 parents 59f141f + 5b9fda5
commit 8416560
Show file tree

Hide file tree

Showing 20 changed files with 529 additions and 231 deletions.
diff --git a/docs/ref/messages.txt b/docs/ref/messages.txt
@@ -1473,6 +1473,7 @@ Represents a theta matrix.
     repeated FloatArray item_weights = 3;
     repeated string topic_name = 4;
     optional int32 topics_count = 5;
+    repeated string item_title = 6;
   }
 
 .. attribute:: ThetaMatrix.model_name
@@ -1482,7 +1483,7 @@ Represents a theta matrix.
 
 .. attribute:: ThetaMatrix.item_id
 
-   A set of item IDs.
+   A set of item IDs corresponding to :attr:`Item.id` values.
 
 .. attribute:: ThetaMatrix.item_weights
 
@@ -1500,6 +1501,10 @@ Represents a theta matrix.
    A value that describes the number of topics in the topic model.
    This value will match :attr:`ModelConfig.topics_count` value, defined in the model config.
 
+.. attribute:: ThetaMatrix.item_id
+
+   A set of item titles, corresponding to :attr:`Item.title` values.
+
 .. _CollectionParserConfig:
 
 CollectionParserConfig

diff --git a/src/artm/core/helpers.cc b/src/artm/core/helpers.cc
@@ -274,6 +274,7 @@ bool BatchHelpers::PopulateThetaMatrixFromCacheEntry(
   }
 
   theta_matrix->set_topics_count(topic_indices.size());
+  bool has_title = (cache.item_title_size() == cache.item_id_size());
   for (int item_index = 0; item_index < cache.item_id_size(); ++item_index) {
     const artm::FloatArray& item_theta = cache.theta(item_index);
     if (all_topics) {
@@ -290,6 +291,7 @@ bool BatchHelpers::PopulateThetaMatrixFromCacheEntry(
     }
 
     theta_matrix->add_item_id(cache.item_id(item_index));
+    if (has_title) theta_matrix->add_item_title(cache.item_title(item_index));
   }
 
   return true;

diff --git a/src/artm/core/internals.pb.cc b/src/artm/core/internals.pb.cc