Update app.py

app.py

@@ -1,23 +1,30 @@
 import streamlit as st
 import pandas as pd
 import openai
+from dateutil import parser
 
 # Function to calculate late interest
 def calculate_late_interest(data, late_interest_rate):
-    # Check if 'due_date' and 'payment_date' columns exist
-    if 'due_date' not in data.columns or 'payment_date' not in data.columns:
-        st.error("Error: 'due_date' or 'payment_date' columns not found in the uploaded file.")
-        return None
-
-    # Convert 'due_date' and 'payment_date' to datetime
-    data['due_date'] = pd.to_datetime(data['due_date'], errors='coerce')
-    data['payment_date'] = pd.to_datetime(data['payment_date'], errors='coerce')
-
     # Calculate late days and late interest
     data['late_days'] = (data['payment_date'] - data['due_date']).dt.days.clip(lower=0)
     data['late_interest'] = data['late_days'] * data['amount'] * (late_interest_rate / 100)
     return data
 
+# Function to analyze Excel sheet and extract relevant information
+def analyze_excel(df):
+    # Extract due dates and payment dates
+    due_dates = df.iloc[:, 0].dropna().tolist()
+    payment_dates = df.iloc[:, 1].dropna().tolist()
+    amounts = []
+
+    # Extract and clean amounts from third column
+    for amount in df.iloc[:, 2]:
+        if isinstance(amount, str):
+            amount = amount.replace('"', '').replace(',', '')
+        amounts.append(float(amount))
+
+    return due_dates, payment_dates, amounts
+
 # Streamlit App
 def main():
     st.title("Invoice Interest Calculator and Conversation")
@@ -32,29 +39,56 @@ def main():
         st.write("Uploaded Data:")
         st.write(df)
 
+        # Analyze Excel sheet
+        due_dates, payment_dates, amounts = analyze_excel(df)
+
         # Allow user to specify late interest rate
         late_interest_rate = st.number_input("Enter Late Interest Rate (%):", min_value=0.0, max_value=100.0, step=0.1)
 
-        # Analyze data and calculate late interest
-        df_with_interest = calculate_late_interest(df, late_interest_rate)
-
-        if df_with_interest is not None:
-            # Display analyzed data
-            st.write("Data with Late Interest Calculated:")
-            st.write(df_with_interest)
-
-            # Allow user to engage in conversation
-            user_input = st.text_input("Start a conversation:")
-            if st.button("Send"):
-                openai.api_key = api_key  # Set OpenAI API key
-                completion = openai.Completion.create(
-                    engine="davinci-codex",
-                    prompt=user_input,
-                    max_tokens=100
-                )
-                response = completion.choices[0].text.strip()
-                st.write("AI's Response:")
-                st.write(response)
+        # Calculate late interest if due dates and payment dates are available
+        if due_dates and payment_dates:
+            # Create DataFrame with extracted due dates, payment dates, and placeholder amount
+            df_calculate = pd.DataFrame({
+                'due_date': due_dates,
+                'payment_date': payment_dates,
+                'amount': [0] * len(due_dates)  # Placeholder amount for calculation
+            })
+
+            # Calculate late interest
+            df_with_interest = calculate_late_interest(df_calculate, late_interest_rate)
+
+            # Display calculated late interest
+            st.write("Calculated Late Interest:")
+            st.write(df_with_interest['late_interest'].sum())
+
+        # Generate conversation prompt
+        prompt = "I have analyzed the provided Excel sheet. "
+        if due_dates:
+            prompt += f"The due dates in the sheet are: {', '.join(str(date) for date in due_dates)}. "
+        if payment_dates:
+            prompt += f"The payment dates in the sheet are: {', '.join(str(date) for date in payment_dates)}. "
+        if amounts:
+            prompt += f"The amounts in the sheet are: {', '.join(str(amount) for amount in amounts)}. "
+        prompt += "Based on this information, what would you like to discuss?"
+
+        # Allow user to engage in conversation
+        user_input = st.text_input("Start a conversation:")
+        if st.button("Send"):
+            if 'api_key' not in st.session_state:
+                st.session_state.api_key = st.text_input("Enter your OpenAI API key:")
+            openai.api_key = st.session_state.api_key  # Set OpenAI API key
+            
+            completion = openai.ChatCompletion.create(
+                model="gpt-3.5-turbo",
+                messages=[
+                    {"role": "system", "content": prompt},
+                    {"role": "user", "content": user_input}
+                ],
+                max_tokens=800  # Adjust this value to allow longer responses
+            )
+            response = completion.choices[0].message['content']
+            st.write("AI's Response:")
+            st.write(response)
 
 if __name__ == "__main__":
     main()