app updated

app.py

@@ -12,10 +12,13 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-import streamlit as st
 from datetime import datetime
+
+import streamlit as st
+
 from utils.helper import *
 
+
 def run():
     st.set_page_config(page_title="Momentum-Strategy", page_icon="💹", layout="wide")
 
@@ -123,16 +126,22 @@ def run():
         )
 
     with st.expander("Please expand/collapse to chat with an AI advisor:"):
-        user_question = st.text_input('Enter a question:', 'What does the company NVIDIA do?')
-        ai_answer = run_langchain_agent_(question=user_question)
+        user_question = st.text_input(
+            "Enter a question:", "What does the company NVIDIA do?"
+        )
+        ai_answer = call_gpt(
+            prompt=user_question,
+            content="""
+            You are a Financial Advisor. You know all about FinTech, Corporate Finance, Modern Portfolio Theory, and knowledge like that.
+            """,
+        )
         st.markdown(ai_answer)
 
-
-    # Credit    
+    # Credit
     def current_year():
         now = datetime.now()
         return now.year
-    
+
     # Example usage:
     current_year = current_year()  # This will print the current year
     st.markdown(

utils/helper.py

@@ -9,6 +9,7 @@ import yfinance as yf
 from langchain import LLMChain, PromptTemplate
 from langchain.agents import initialize_agent, load_tools
 from langchain.llms import OpenAI
+from openai import OpenAI
 
 
 def download_stock_data(
@@ -293,9 +294,7 @@ def display_simulated_ef_with_random(
 OPENAI_API_KEY = os.environ["OPENAI_API_KEY"]
 
 
-def run_langchain_agent_(
-    question: str = "What is your question?"
-) -> str:
+def run_langchain_agent_(question: str = "What is your question?") -> str:
     """
     Executes a language chain agent to answer questions by using a series of tools.
 
@@ -320,9 +319,7 @@ def run_langchain_agent_(
     You are a financial advisor and user has a question above regarding related tickers provided.    
     Let's think step by step.
     Answer: """
-    prompt = PromptTemplate(
-        template=template, input_variables=["question"]
-    )
+    prompt = PromptTemplate(template=template, input_variables=["question"])
 
     # Building a chain of language model actions based on the prompt template
     llm_chain = LLMChain(prompt=prompt, llm=llm)
@@ -332,7 +329,10 @@ def run_langchain_agent_(
 
     # Initializing the agent with the loaded tools, the language model, default name, and verbosity
     agent = initialize_agent(
-        tools, llm, agent="zero-shot-react-description", verbose=True,
+        tools,
+        llm,
+        agent="zero-shot-react-description",
+        verbose=True,
         max_iterations=5,
     )
 
@@ -341,3 +341,41 @@ def run_langchain_agent_(
 
     # Return
     return output_
+
+
+openai_client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])
+
+
+def call_gpt(prompt: str, content: str) -> str:
+    """
+    Sends a structured conversation context including a system prompt, user prompt,
+    and additional background content to the GPT-3.5-turbo model for a response.
+    This function is responsible for generating an AI-powered response by interacting
+    with the OpenAI API. It puts together a preset system message, a formatted user query,
+    and additional background information before requesting the completion from the model.
+    Args:
+        prompt (str): The main question or topic that the user wants to address.
+        content (str): Additional background information or details relevant to the prompt.
+    Returns:
+        str: The generated response from the GPT model based on the given prompts and content.
+    Note: 'openai_client' is assumed to be an already created and authenticated instance of the OpenAI
+          openai_client, which should be set up prior to calling this function.
+    """
+
+    # Generates a response from the model based on the interactive messages provided
+    response = openai_client.chat.completions.create(
+        model="gpt-3.5-turbo",  # The AI model being queried for a response
+        messages=[
+            # System message defining the assistant's role
+            {"role": "system", "content": "You are a helpful assistant."},
+            # User message containing the prompt
+            {"role": "user", "content": f"I want to ask you a question: {prompt}"},
+            # Assistant message asking for background content
+            {"role": "assistant", "content": "What is the background content?"},
+            # User providing the background content
+            {"role": "user", "content": content},
+        ],
+    )
+
+    # Extracts and returns the response content from the model's completion
+    return response.choices[0].message.content