add corpus

app.py

@@ -1,6 +1,6 @@
 import time
 import json
-from pyserini.search.lucene import LuceneImpactSearcher
+from pyserini.search.lucene import LuceneImpactSearcher, LuceneSearcher
 import streamlit as st
 from pathlib import Path
 import sys
@@ -53,6 +53,8 @@ encoder, index = encoder_index_map[encoder][runtime_index], encoder_index_map[en
 searcher = LuceneImpactSearcher(
     f'indexes/{index}', f'{encoder}', encoder_type=f'{runtime}')
 
+corpus = LuceneSearcher(f'indexes/index-unicoil')
+
 col1, col2 = st.columns([9, 1])
 with col1:
     search_query = st.text_input(label="search query", placeholder="Search")
@@ -72,11 +74,10 @@ if search_query or button_clicked:
     for i, result in enumerate(search_results[:10]):
         result_score = result.score
         result_id = result.docid
-        if result.raw:
-            contents = json.loads(result.raw)
-            contents = contents['contents'] if 'contents' in contents else contents['content']
-        else:
-            contents = ''
+        contents = json.loads(result.raw)
+        contents = contents['contents'] if 'contents' in contents else contents['content']
+        if contents == "":
+            contents = json.loads(corpus.doc(result_id).raw())['contents']
 
         output = f'<div class="row"> <b>Rank</b>: {i+1} | <b>Document ID</b>: {result_id} | <b>Score</b>:{result_score:.2f}</div>'