app.py

import os
import gradio as gr
import pandas as pd
import matplotlib.pyplot as plt
from groq import Groq
from datetime import datetime

# -------------------------------
# Groq Client
# -------------------------------
client = Groq(
    api_key=os.environ.get("GROQ_API_KEY")
)

# -------------------------------
# Session State
# -------------------------------
def init_session():
    return pd.DataFrame(columns=["Date", "Category", "Description", "Amount"])

# -------------------------------
# Add Expense
# -------------------------------
def add_expense(date, category, description, amount, df):
    if df is None:
        df = init_session()

    new_row = {
        "Date": date,
        "Category": category,
        "Description": description,
        "Amount": float(amount)
    }
    df = pd.concat([df, pd.DataFrame([new_row])], ignore_index=True)
    return df, df

# -------------------------------
# Visualization
# -------------------------------
def visualize(df, chart_type):
    if df is None or df.empty:
        return None

    plt.figure()

    if chart_type == "Line Chart":
        df.groupby("Date")["Amount"].sum().plot(kind="line")
    elif chart_type == "Bar Chart":
        df.groupby("Category")["Amount"].sum().plot(kind="bar")
    elif chart_type == "Column Chart":
        df.groupby("Category")["Amount"].sum().plot(kind="bar")
    elif chart_type == "Pie Chart":
        df.groupby("Category")["Amount"].sum().plot(kind="pie", autopct="%1.1f%%")

    plt.title("Expense Analysis")
    plt.ylabel("Amount")
    plt.tight_layout()
    return plt.gcf()

# -------------------------------
# AI Analysis using Groq
# -------------------------------
def ai_analysis(df):
    if df is None or df.empty:
        return "No expenses available for analysis."

    summary = df.groupby("Category")["Amount"].sum().to_dict()
    total = df["Amount"].sum()

    prompt = f"""
    You are an AI financial assistant.
    Analyze the following expense data:
    
    Category-wise Expenses: {summary}
    Total Expenses: {total}

    Provide:
    1. Spending insights
    2. Cost-saving suggestions
    3. Any unusual spending patterns
    """

    chat_completion = client.chat.completions.create(
        messages=[
            {"role": "user", "content": prompt}
        ],
        model="llama-3.3-70b-versatile",
    )

    return chat_completion.choices[0].message.content

# -------------------------------
# Gradio UI
# -------------------------------
with gr.Blocks(title="AI Expense Analyzer") as demo:
    gr.Markdown("# 💰 AI Expense Analyzer")
    gr.Markdown("Track expenses, visualize data, and get AI-powered insights.")

    df_state = gr.State(init_session())

    with gr.Row():
        date = gr.Textbox(label="Date", value=str(datetime.today().date()))
        category = gr.Textbox(label="Category (Food, Rent, Travel, etc.)")
        description = gr.Textbox(label="Description")
        amount = gr.Number(label="Amount")

    add_btn = gr.Button("➕ Add Expense")

    expense_table = gr.Dataframe(
        headers=["Date", "Category", "Description", "Amount"],
        interactive=False
    )

    add_btn.click(
        add_expense,
        inputs=[date, category, description, amount, df_state],
        outputs=[df_state, expense_table]
    )

    gr.Markdown("## 📊 Visualization")
    chart_type = gr.Dropdown(
        ["Line Chart", "Bar Chart", "Column Chart", "Pie Chart"],
        value="Bar Chart"
    )
    chart_btn = gr.Button("Generate Chart")
    chart_output = gr.Plot()

    chart_btn.click(
        visualize,
        inputs=[df_state, chart_type],
        outputs=chart_output
    )

    gr.Markdown("## 🤖 AI Expense Analysis")
    ai_btn = gr.Button("Analyze with AI")
    ai_output = gr.Textbox(lines=12)

    ai_btn.click(
        ai_analysis,
        inputs=df_state,
        outputs=ai_output
    )

demo.launch()