add search

app.py

@@ -1,18 +1,34 @@
 import streamlit as st
+from datasets import load_dataset
+from sentence_transformers import SentenceTransformer
 
-st.markdown("# Buscador de Noticias Salvadoreñas")
+def search(text, model, ds, n):
+  encoded_text = model.encode(text)
+  scores, retrieved_examples = ds.get_nearest_examples('embedding', encoded_text, k=n)
+  matching_titles = retrieved_examples["title"]
+  urls = retrieved_examples["link"]
+  contents = retrieved_examples["content"]
+  return list(zip(matching_titles, [c[:150] for c in contents], urls, scores))
 
-search_text = st.text_input(label="Búsqueda")
+@st.cache_data
+def get_dataset():
+    ds = load_dataset("justinian336/salvadoran-news-embedded")
+    embedded_ds.add_faiss_index(column="embedding")
+    return ds
 
-# ds = load_dataset("justinian336/salvadoran-news-embedded")
+@st.cache__resource
+def get_model():
+    model = SentenceTransformer("justinian336/chupeto")
+    return model
 
+ds = get_dataset()
+model = get_model()
 
-# def search(text, model, ds, n):
-#   encoded_text = model.encode(text)
-#   scores, retrieved_examples = ds.get_nearest_examples('embedding', encoded_text, k=n)
-#   matching_titles = retrieved_examples["title"]
-#   urls = retrieved_examples["link"]
-#   contents = retrieved_examples["content"]
-#   return list(zip(matching_titles, [c[:150] for c in contents], urls, scores))
+st.markdown("# Buscador de Noticias Salvadoreñas")
+search_text = st.text_input(label="Búsqueda")
 
-# search_results = search(search_text, model, embedded_ds, 10)
+search_results = search(search_text, model, embedded_ds, 10)
+for title, content, url in search_results:
+    st.markdown(f"""<div><a href="{url}">{title}</a></div>""", unsafe_allow_html=True)
+    st.markdown(f"""<div>{content}...</div>""")
+    st.markdown("---")