buttons introduced

app.py

@@ -41,9 +41,8 @@ def build_model():
     st.markdown("Model weights loaded")
     return model
     
-def inference(txt, mode=None):
-    # infers classes for text topic based on the trained model from above
-    # has separate mode 'print' for just output
+def inference(txt):
+    # infers classes for text topic based on loaded trained model
     t2 = tokenizer(txt.lower().replace('\n', ''), 
        padding='max_length', max_length = 512, truncation=True,
        return_tensors="pt")
@@ -56,38 +55,51 @@ def inference(txt, mode=None):
     out = out/out.sum() * 100
     res = [(l, o) for l, o in zip (list(labels.keys()), out.tolist())]   
     return res   
+
+def infer_and_display_result(txt):
+    start_time = time()
+
+    res = inference(txt, mode=None)
+    res.sort(key = lambda x : - x[1])
     
-model = build_model()
+    st.markdown("###Inference results:")    
+    for lbl, score in res:
+        if score >=1:
+            st.write(f"[        {lbl:<7}] {labels_decoder[lbl]:<35}  {score:.1f}%")
     
-st.markdown("###Predict topic by abstract.")
+    res_plot = []  # storage for plot data
+    total=0
+    for r in res:
+        if total < 95:
+            res_plot.append(r)
+            total += r[1]
+        else:
+            break
+    res.sort(key = lambda x : x[1])         
+    
+    fig, ax = plt.subplots(figsize=(10, len(res_plot)))
+    for r in res_plot :
+        ax.barh(r[0], r[1])
+    st.pyplot(fig)
+    st.markdown(f"cycle time = {time() - start_time:.2f} s.")
+   
+# ======================================            
+model = build_model()
+
+st.title('Big-data cloud application for scientific article topic inference using in-memory computing and stuff.')    
 image = Image.open('dilbert_big_data.jpg')
 st.image(image)
+st.write('test application for ML-2 class, YSDA-2022' )
 # st.markdown("<img width=200px src='https://i.pinimg.com/736x/11/33/19/113319f0ffe91f4bb0f468914b9916da.jpg'>", unsafe_allow_html=True)
+# st.markdown("###Predict topic by abstract.")
 
-text = st.text_area("ENTER TEXT HERE")
-
-start_time = time()
-
-res = inference(text, mode=None)
-res.sort(key = lambda x : - x[1])
-
-st.markdown("INFERENCE RESULT")    
-for lbl, score in res:
-    if score >=1:
-        st.markdown(f"[        {lbl:<7}] {labels_decoder[lbl]:<35}  {score:.1f}%")
-
-res_plot = []
-total=0
-for r in res:
-    if total < 95:
-        res_plot.append(r)
-        total += r[1]
-    else:
-        break
+text = st.text_area("ENTER ARTICLE TITLE OR ABSTRACT HERE")
+action = st.button('click here to infer topic')
+if action: 
+    infer_and_display_result(text)
+    
+action2 = st.button('click here to infer topic')
+if action2: 
+    st.write(text.upper())   
 
-fig, ax = plt.subplots(figsize=(10, len(res_plot)+1))
-for r in res_plot :
-    ax.barh(r[0], r[1])
-st.pyplot(fig)
 
-st.markdown(f"cycle time = {time() - start_time:.2f} s.")