Create app.py

app.py

@@ -0,0 +1,54 @@
+import streamlit as st
+import bm25s
+from operator import itemgetter
+import os
+import re
+
+
+@st.cache_data
+def load_data():
+    df = pd.read_csv("/kaggle/input/cleanlist/cleaned_list.csv",header = None)
+    df.columns = ['document']
+    corpus = [doc for doc in df['document'].to_list()]
+
+    retriever = bm25s.BM25(corpus=corpus)
+    retriever.index(bm25s.tokenize(corpus)
+
+    return retriever
+
+def extract_hscode(text):
+    match = re.search(r'hs_code:\s*(\d+)', text)
+    if match:
+        return match.group(1)
+    return None
+
+df2 = pd.read_csv("/kaggle/input/hscode-miss/hscode_main.csv")
+new_col = [len(str(code))for code in df2['hs_code'].to_list()]
+df2['len'] = new_col
+
+new_hscode = [str(code) for code in df2['hs_code']]
+
+for i in range(len(new_col)):
+    if new_col[i]==5:
+        new_hscode[i] = '0'+ new_hscode[i]
+df2['hs_code'] = new_hscode
+df2=df2.drop(columns='len')
+
+if 'retriever' not in st.session_state:
+    st.session_state.retriever = None
+
+if st.session_state.retriever is None:
+    st.session_state.retriever = load data()
+
+
+sentence = st.text_input("please enter description:")
+
+if sentence !='':
+    results,_ = st.session_state.retriever.retrieve(bm25s.tokenize(query), k=5)
+    doc = [d for d in results]
+    hscodes = [extract_hscode(item) for item in doc[0]]
+    for code in hscodes:
+	filter_df = df2[df2['hs_code']==code]
+	answer = filter_df['full_description'].iloc[0]
+	st.write("Hscode:",code)
+    	st.write("answer:",answer)