updates

app.py

@@ -4,20 +4,22 @@ from query import VectaraQuery
 import streamlit as st
 import os
 
+topics =  { 
+    'Standardized testing in education': '266',
+    'Ethical implications of genetic editing': '267'
+}
+
 def launch_bot():
-    def generate_response(question, role, topic):
-        response = vq.submit_query(question, role, topic)
+    def generate_response(question, cfg):
+        response = vq.submit_query(question, cfg.bot_role, cfg.topic, cfg.style)
         return response
 
     if 'cfg' not in st.session_state:
         cfg = OmegaConf.create({
             'customer_id': str(os.environ['VECTARA_CUSTOMER_ID']),
-            'corpus_id': str(os.environ['VECTARA_CORPUS_ID']),
+            'corpus_id': list(topics.values())[0],
             'api_key': str(os.environ['VECTARA_API_KEY']),
             'prompt_name': 'vectara-experimental-summary-ext-2023-12-11-large',
-            'topic': 'Standardized testing in education',
-            'human_role': 'in opposition to',
-            'bot_role': 'in support of'
         })
         st.session_state.cfg = cfg
         st.session_state.vq = VectaraQuery(cfg.api_key, cfg.customer_id, cfg.corpus_id, cfg.prompt_name)
@@ -35,34 +37,42 @@ def launch_bot():
         cfg.human_role = st.selectbox('Your are:', role_options)
         cfg.bot_role = role_options[1] if cfg.human_role == role_options[0] else role_options[0]
 
-        st.markdown(f"{cfg.topic}.\n\n")
+        topic_options = list(topics.keys())
+        cfg.topic = st.selectbox('The topic:', topic_options)
+        vq.corpus_id = topics[cfg.topic]
+
+        st.markdown("\n")
+        debate_styles = ['Lincoln-Douglas', 'Spontaneous Argumentation', 'Parliamentary debates']
+        cfg.style = st.selectbox('Debate Style:', debate_styles)
+
         st.markdown("---")
         st.markdown(
             "## How this works?\n"
-            "This app was built with [Vectara](https://vectara.com).\n"
+            "This app was built with [Vectara](https://vectara.com).\n\n"
+            "It demonstrates the use of the Chat functionality along with custom prompts and GPT4-Turbo (as part of our Scale plan)"
         )
         st.markdown("---")
 
     if "messages" not in st.session_state.keys():
-        st.session_state.messages = [{"role": "assistant", "content": f"Please make your opening statment."}]
+        st.session_state.messages = [{"role": "assistant", "content": f"Please make your opening statement.", "avatar": '🦖'}]
 
     # Display chat messages
     for message in st.session_state.messages:
-        with st.chat_message(message["role"]):
+        with st.chat_message(message["role"], avatar=message["avatar"]):
             st.write(message["content"])
 
     # User-provided prompt
     if prompt := st.chat_input():
-        st.session_state.messages.append({"role": "user", "content": prompt})
-        with st.chat_message("user"):
+        st.session_state.messages.append({"role": "user", "content": prompt, "avatar": '🧑‍💻'})
+        with st.chat_message("user", avatar='🧑‍💻'):
             st.write(prompt)
     
     # Generate a new response if last message is not from assistant
     if st.session_state.messages[-1]["role"] != "assistant":
-        with st.chat_message("assistant"):
-            stream = generate_response(prompt, cfg.bot_role, cfg.topic) 
+        with st.chat_message("assistant", avatar='🤖'):
+            stream = generate_response(prompt, cfg) 
             response = st.write_stream(stream) 
-            message = {"role": "assistant", "content": response}
+            message = {"role": "assistant", "content": response, "avatar": '🤖'}
             st.session_state.messages.append(message)
     
 if __name__ == "__main__":

query.py

@@ -16,7 +16,7 @@ class VectaraQuery():
         self.prompt_name = prompt_name if prompt_name else "vectara-experimental-summary-ext-2023-12-11-large"
         self.conv_id = None
 
-    def get_body(self, user_response: str, role: str, topic: str):
+    def get_body(self, user_response: str, role: str, topic: str, style: str):
         corpora_key_list = [{
             'customer_id': self.customer_id, 'corpus_id': self.corpus_id, 'lexical_interpolation_config': {'lambda': 0.025}
         }]
@@ -25,7 +25,9 @@ class VectaraQuery():
         [
             {{
                 "role": "system",
-                "content": "You are a professional debate bot. You are provided with search results related to {topic} 
+                "content": "You are a professional debate bot. 
+                You specialize in the {style} debate style.
+                You are provided with search results related to {topic} 
                 and respond to the previous arugments made so far. Be sure to provide a thoughtful and convincing reply.
                 Never mention search results explicitly in your response.
                 Do not base your response on information or knowledge that is not in the search results.
@@ -41,9 +43,10 @@ class VectaraQuery():
             }},
             {{
                 "role": "user",
-                "content": "provide a convincing reply {role} {topic}.
+                "content": "provide a convincing reply {role} {topic} in response to the last argument '{user_response}.
                 Consider the search results as relevant information with which to form your response.
-                Do not repeat earlier arguments and make sure your new response is coherent with the previous arguments, and responsive to the last argument: {user_response}."
+                Use the {style} debate style to make your argument.
+                Do not repeat earlier arguments and make sure your new response is coherent with the previous arguments."
             }}
         ]
         '''
@@ -93,10 +96,10 @@ class VectaraQuery():
             "grpc-timeout": "60S"
         }
 
-    def submit_query(self, query_str: str, role: str, topic: str):
+    def submit_query(self, query_str: str, bot_role: str, topic: str, style: str):
 
         endpoint = f"https://api.vectara.io/v1/stream-query"
-        body = self.get_body(query_str, role, topic)
+        body = self.get_body(query_str, bot_role, topic, style)
 
         response = requests.post(endpoint, data=json.dumps(body), verify=True, headers=self.get_headers(), stream=True) 
         if response.status_code != 200: