fix limit and highlighting and empty docs

README.md

@@ -1,7 +1,7 @@
 ---
 title: CitationQA - Answers from science with citations
-emoji: 🐠
-colorFrom: pink
+emoji: 👩🏽‍🔬
+colorFrom: blue
 colorTo: indigo
 sdk: streamlit
 sdk_version: 1.10.0

app.py

@@ -10,7 +10,7 @@ def remove_html(x):
     text = soup.get_text()
     return text
 
-def search(term, limit=10):
+def search(term, limit=25):
     search = f"https://api.scite.ai/search?mode=citations&term={term}&limit={limit}&offset=0&user_slug=domenic-rosati-keW5&compute_aggregations=false"
     req = requests.get(
         search,
@@ -43,7 +43,9 @@ def find_source(text, docs):
             'citation_statement': '',
             'text': text,
             'from': '',
-            'supporting': ''
+            'supporting': '',
+            'source_title': '',
+            'source_link': ''
         }
 
 @st.experimental_singleton
@@ -89,11 +91,11 @@ def run_query(query):
         <div class="container-fluid">
         <div class="row align-items-start">
              <div  class="col-md-12 col-sm-12">
-            Sorry... no results for that question! Try another.
+            Sorry... no results for that question! Try another...
          </div>
         </div>
      </div>
-        """)
+        """, unsafe_allow_html=True)
 
     results = []
     model_results = qa_model(question=query, context=context, top_k=10)
@@ -113,11 +115,12 @@ def run_query(query):
     sorted_result = list({
         result['context']: result for result in sorted_result
     }.values())
+    sorted_result = sorted(sorted_result, key=lambda x: x['score'], reverse=True)
 
 
     for r in sorted_result:
         answer = r["answer"]
-        ctx = r["context"].replace(answer, f"<mark>{answer}</mark>").replace('<cite', '<a').replace('</cite', '</a').replace('data-doi="', 'href="https://scite.ai/reports/')
+        ctx = remove_html(r["context"]).replace(answer, f"<mark>{answer}</mark>").replace('<cite', '<a').replace('</cite', '</a').replace('data-doi="', 'href="https://scite.ai/reports/')
         title = r["title"].replace("_", " ")
         score = round(r["score"], 4)
         card(title, ctx, score, r['link'])