app.py

app.py

@@ -0,0 +1,34 @@
+import streamlit as st
+import pandas as pd
+
+from transformers import pipeline
+
+@st.cache(allow_output_mutation=True)
+def get_model(model):
+	return pipeline("fill-mask", model=model, top_k=100)#seto maximum of tokens to be retrieved after each inference to model
+
+HISTORY_WEIGHT = 100 # set history weight (if found any keyword from history, it will priorities based on its weight)
+
+
+history_keyword_text = st.text_input("Enter users's history keywords (optional, i.e., 'Gates')", value="Gates")
+
+text = st.text_input("Enter a text for auto completion...", value='Where is Bill')
+
+model = st.selectbox("choose a model", ["roberta-base", "bert-base-uncased", "gpt2", "t5"])
+
+data_load_state = st.text('Loading model...')
+nlp = get_model(model)
+
+if text:
+    data_load_state = st.text('Inference to model...')
+    result = nlp(text+' '+nlp.tokenizer.mask_token)
+    data_load_state.text('')
+    for index, r in enumerate(result):
+        if r['token_str'].lower().strip() in history_keyword_text.lower().strip():
+            #found from history, then increase the score of tokens
+            result[index]['score']*=HISTORY_WEIGHT
+            
+    #sort the results        
+    df=pd.DataFrame(result).sort_values(by='score', ascending=False)
+    #show the results as a table
+    st.table(df)