Exclude seed papers from results.

app.py

@@ -61,6 +61,9 @@ def read_user(seed_json):
         st.session_state['pid2sent_vectors_user'] = pid2sent_vectors_user
         st.session_state['user_kps'] = user_kps
         st.session_state['username'] = uname
+        for pd in seed_json['papers']:
+            norm_title = " ".join(pd['title'].lower().strip().split())
+            st.session_state.seed_titles.append(norm_title)
         return doc_vectors_user, pid2idx_user, pid2sent_vectors, user_kps
     else:
         return st.session_state.doc_vectors_user, st.session_state.pid2idx_user, \
@@ -240,6 +243,10 @@ def second_stage_ranked_docs(selected_query_kps, first_stage_pids, pid2abstract,
     # Get the docs
     retrieved_papers = collections.OrderedDict()
     for pid in top_pids:
+        # Exclude papers from the seed set in the result set.
+        norm_title = " ".join(pid2abstract[pid]['title'].lower().strip().split())
+        if norm_title in st.session_state.seed_titles:
+            continue
         retrieved_papers[pid2abstract[pid]['title']] = {
             'title': pid2abstract[pid]['title'],
             'kp_explanations': pid2kp_expls[pid],