Create components/decision_engine.py

components/decision_engine.py

@@ -1,31 +1,34 @@
+# components/decision_engine.py
+
 import streamlit as st
+from langchain_core.prompts import ChatPromptTemplate
+from langchain_core.runnables import Runnable
 from langchain_groq import ChatGroq
-from langchain.prompts import ChatPromptTemplate
-from langchain_core.output_parsers import StrOutputParser
-from utils.session import get_state_value
+from utils.session import get_canvas_data
+from utils.prompts import DECISION_ENGINE_PROMPT
+
 
 def run_decision_engine():
-    st.title("🎯 Strategy Suggestions")
-    st.markdown("Let the AI suggest smart next steps based on your Business Model Canvas.")
+    st.header("🎯 Strategy Suggestions")
 
-    canvas_data = get_state_value("canvas_data", None)
-    if not canvas_data:
+    canvas = get_canvas_data()
+    if not canvas:
         st.warning("Please complete the Canvas Assistant first.")
         return
 
-    sections_text = "\n\n".join(
-        [f"**{section.replace('_', ' ').title()}**:\n{value}" for section, value in canvas_data.items()]
-    )
-
-    prompt = ChatPromptTemplate.from_messages([
-        ("system", "You're a strategic business advisor. Analyze the provided Business Model Canvas data and suggest specific next steps to improve and grow the business."),
-        ("user", "{canvas_summary}")
-    ])
+    # Show canvas summary
+    st.subheader("📋 Canvas Overview")
+    for section, content in canvas.items():
+        st.markdown(f"**{section}**")
+        st.info(content)
 
-    chain = prompt | ChatGroq(model_name="llama3-8b-8192", temperature=0.4) | StrOutputParser()
+    with st.spinner("Analyzing canvas for strategic insights..."):
+        prompt = ChatPromptTemplate.from_template(
+            DECISION_ENGINE_PROMPT + "\n\nCanvas Data:\n{input}"
+        )
+        chain: Runnable = prompt | ChatGroq(model="llama3-8b-8192", temperature=0.3)
+        full_canvas_text = "\n".join([f"{k}: {v}" for k, v in canvas.items()])
+        result = chain.invoke({"input": full_canvas_text})
 
-    if st.button("🚀 Suggest Strategies"):
-        with st.spinner("Analyzing your business model..."):
-            result = chain.invoke({"canvas_summary": sections_text})
-            st.markdown("### 🔍 AI-Generated Strategy Suggestions")
-            st.markdown(result)
+        st.subheader("🧠 AI Strategy Suggestions")
+        st.success(result.content)