Fix bigartm#758 ARTM.get_phi_dense() to extract phi without pandas.Da…

…taFrame

docs/release_notes/python.txt

@@ -13,6 +13,9 @@ v0.8.3
 
 * Enable copying of ARTM, LDA, hARTM and ARTM_Level objects with ``clone()`` method and ``copy.deepcopy(obj)``.
 * Experimental support for import/export/editing of theta matrices; for more details see python reference of ``ARTM.__init__(ptd_name='ptd')``.
+* ``ARTM.get_phi_dense()`` method to extract phi matrix without pandas.DataFrame, see `#758 <https://github.com/bigartm/bigartm/issues/758>`_.
+* Bug fix in ``ARTM.get_phi_sparse()`` to return tokens as rows, and topic names as columns (previously it was the opposite way)
+
 
 v0.8.2
 ------

python/artm/artm_model.py

@@ -644,10 +644,9 @@ def load(self, filename, model_name='p_wt'):
         self._num_online_processed_batches = 0
         self._phi_cached = None
 
-    def get_phi(self, topic_names=None, class_ids=None, model_name=None):
+    def get_phi_dense(self, topic_names=None, class_ids=None, model_name=None):
         """
-        :Description: get custom Phi matrix of model. The extraction of the\
-                      whole Phi matrix expects ARTM.phi_ call.
+        :Description: get phi matrix in dense format
 
         :param topic_names: list with topics or single topic to extract, None value means all topics
         :type topic_names: list of str or str or None
@@ -657,10 +656,10 @@ def get_phi(self, topic_names=None, class_ids=None, model_name=None):
                       reasonable to extract unnormalized counters
 
         :return:
-          * pandas.DataFrame: (data, columns, rows), where:
-          * columns --- the names of topics in topic model;
+          * a 3-tuple of (data, rows, columns), where
+          * data --- numpy.ndarray with Phi data (i.e., p(w|t) values)
           * rows --- the tokens of topic model;
-          * data --- content of Phi matrix.
+          * columns --- the names of topics in topic model;
         """
         if not self._initialized:
             raise RuntimeError('Model does not exist yet. Use ARTM.initialize()/ARTM.fit_*()')
@@ -682,6 +681,29 @@ def get_phi(self, topic_names=None, class_ids=None, model_name=None):
                   if class_ids is None or class_id in class_ids]
         topic_names = [topic_name for topic_name in topics_and_tokens_info.topic_name
                        if topic_names is None or topic_name in topic_names]
+        return nd_array, tokens, topic_names
+
+    def get_phi(self, topic_names=None, class_ids=None, model_name=None):
+        """
+        :Description: get custom Phi matrix of model. The extraction of the\
+                      whole Phi matrix expects ARTM.phi_ call.
+
+        :param topic_names: list with topics or single topic to extract, None value means all topics
+        :type topic_names: list of str or str or None
+        :param class_ids: list with class_ids or single class_id to extract, None means all class ids
+        :type class_ids: list of str or str or None
+        :param str model_name: self.model_pwt by default, self.model_nwt is also\
+                      reasonable to extract unnormalized counters
+
+        :return:
+          * pandas.DataFrame: (data, columns, rows), where:
+          * columns --- the names of topics in topic model;
+          * rows --- the tokens of topic model;
+          * data --- content of Phi matrix.
+        """
+        (nd_array, tokens, topic_names) =  self.get_phi_dense(topic_names=topic_names,
+                                                              class_ids=class_ids,
+                                                              model_name=model_name)
         phi_data_frame = DataFrame(data=nd_array,
                                    columns=topic_names,
                                    index=tokens)
@@ -744,8 +766,8 @@ def get_phi_sparse(self, topic_names=None, class_ids=None, model_name=None, eps=
         # Columns correspond to topics; get topic names from topic_model.topic_name
         data = sparse.csr_matrix((data, (row_ind, col_ind)),
                                  shape=(len(topic_model.token), len(topic_model.topic_name)))
-        rows = list(topic_model.topic_name)
-        columns = list(topic_model.token)
+        columns = list(topic_model.topic_name)
+        rows = list(topic_model.token)
         return data, rows, columns
 
     def get_theta(self, topic_names=None):