Upload 2 files

app.py

@@ -68,31 +68,23 @@ for msg in st.session_state.messages:
 st.markdown('</div>', unsafe_allow_html=True)  # Close chat-box
 st.markdown('</div>', unsafe_allow_html=True)  # Close chat-container
 
-if "input" not in st.session_state:
-    st.session_state.input = ""
-
-with st.form(key="chat_form"):
-    user_input = st.text_input("💬 ጥያቄዎን ያስገቡ:", value=st.session_state.input, key="input_form")
+# Form with clear_on_submit and direct input capture
+with st.form(key="chat_form", clear_on_submit=True):
+    user_input = st.text_input("💬 ጥያቄዎን ያስገቡ:")
     submit = st.form_submit_button("መልስ አውጣ")
 
-if submit:
-    user_input = st.session_state.input_form.strip()
-
-    if user_input == "":
-        st.warning("እባክዎ ጥያቄ ያስገቡ።")
-    else:
-        # Append user input
-        st.session_state.messages.append({"sender": "user", "text": user_input})
-
-        # Generate response
-        response = bot.get_answer(user_input)
-        if response == "__OUT_OF_SCOPE__":
-            response = "ይቅርታ፣ ይህንን ጥያቄ ማስተዋል አልቻልኩም። እባክዎ በሌላ መንገድ ይሞክሩ።"
-
-        st.session_state.messages.append({"sender": "bot", "text": response})
-
-        # Reset input manually
-        st.session_state.input_form = ""
+    if submit:
+        if user_input.strip() == "":
+            st.warning("እባክዎ ጥያቄ ያስገቡ።")
+        else:
+            # Append user input
+            st.session_state.messages.append({"sender": "user", "text": user_input})
 
+            # Generate response
+            response = bot.get_answer(user_input)
+            if response == "__OUT_OF_SCOPE__":
+                response = "ይቅርታ፣ ይህንን ጥያቄ ማስተዋል አልቻልኩም ከእኔ መረጃ ውጪ ነው። እባክዎ በሌላ መንገድ ይሞክሩ።"
 
+            # Append bot response
+            st.session_state.messages.append({"sender": "bot", "text": response})
 

chatbot_utils.py

@@ -6,7 +6,7 @@ import numpy as np
 from sklearn.metrics.pairwise import cosine_similarity
 
 class AmharicChatbot:
-    def __init__(self, csv_path, threshold=0.70):
+    def __init__(self, csv_path, threshold=0.80):
         self.df = pd.read_csv(csv_path)
         self.model = SentenceTransformer("intfloat/multilingual-e5-small")
         self.threshold = threshold
@@ -23,10 +23,21 @@ class AmharicChatbot:
     def get_answer(self, user_question, k=1):
         user_embedding = self.model.encode([f"query: {user_question}"])[0].astype("float32")
         D, I = self.index.search(np.array([user_embedding]), k)
+
+        if len(I[0]) == 0:
+            return "__OUT_OF_SCOPE__"
+
         top_idx = I[0][0]
-        top_question = self.df.iloc[top_idx]["question"]
-        top_embedding = self.model.encode([f"passage: {top_question}"])[0]
+        top_embedding = self.embeddings[top_idx]
+
+        # Normalize embeddings before cosine similarity
+        user_embedding = user_embedding / np.linalg.norm(user_embedding)
+        top_embedding = top_embedding / np.linalg.norm(top_embedding)
+
         score = cosine_similarity([user_embedding], [top_embedding])[0][0]
+
         if score < self.threshold:
             return "__OUT_OF_SCOPE__"
+
         return self.df.iloc[top_idx]["answer"]
+