- Created a new eveai_chat plugin to support the new dynamic possibilities of the Specialists. Currently only supports standard Rag retrievers (i.e. no extra arguments).

eveai_app/views/interaction_views.py

@@ -72,11 +72,21 @@ def handle_chat_session_selection():
 @interaction_bp.route('/view_chat_session/<int:chat_session_id>', methods=['GET'])
 @roles_accepted('Super User', 'Tenant Admin')
 def view_chat_session(chat_session_id):
+    # Get chat session with user info
     chat_session = ChatSession.query.get_or_404(chat_session_id)
+
+    # Get interactions with specialist info
     interactions = (Interaction.query
                     .filter_by(chat_session_id=chat_session.id)
-                    .order_by(Interaction.question_at)
-                    .all())
+                    .join(Specialist, Interaction.specialist_id == Specialist.id, isouter=True)
+                    .add_columns(
+                        Interaction.id,
+                        Interaction.question_at,
+                        Interaction.specialist_arguments,
+                        Interaction.specialist_results,
+                        Specialist.name.label('specialist_name'),
+                        Specialist.type.label('specialist_type')
+                    ).order_by(Interaction.question_at).all())
 
     # Fetch all related embeddings for the interactions in this session
     embedding_query = (db.session.query(InteractionEmbedding.interaction_id,
@@ -84,7 +94,7 @@ def view_chat_session(chat_session_id):
                                         DocumentVersion.object_name)
                        .join(Embedding, InteractionEmbedding.embedding_id == Embedding.id)
                        .join(DocumentVersion, Embedding.doc_vers_id == DocumentVersion.id)
-                       .filter(InteractionEmbedding.interaction_id.in_([i.id for i in interactions])))
+                       .filter(InteractionEmbedding.interaction_id.in_([i.id for i, *_ in interactions])))
 
     # Create a dictionary to store embeddings for each interaction
     embeddings_dict = {}