app.py

import requests
import pandas as pd
import gradio as gr
import plotly.express as px
from datetime import datetime

# Function to fetch and process the transaction data from the API
def fetch_transactions():
    url = "https://li.quest/v1/analytics/transfers?integrator=valory"
    headers = {"accept": "application/json"}
    response = requests.get(url, headers=headers)
    return response.json()

# Function to parse the transaction data and prepare it for visualization
def process_transactions(data):
    transactions = data["transfers"]

    # Convert the data into a pandas DataFrame for easy manipulation
    rows = []
    for tx in transactions:
        # Normalize amounts
        sending_amount = float(tx["sending"]["amount"]) / (10 ** tx["sending"]["token"]["decimals"])
        receiving_amount = float(tx["receiving"]["amount"]) / (10 ** tx["receiving"]["token"]["decimals"])

        # Convert timestamps to datetime objects
        sending_timestamp = datetime.utcfromtimestamp(tx["sending"]["timestamp"])
        receiving_timestamp = datetime.utcfromtimestamp(tx["receiving"]["timestamp"])

        # Prepare row data
        rows.append({
            "transactionId": tx["transactionId"],
            "from_address": tx["fromAddress"],
            "to_address": tx["toAddress"],
            "sending_chain": tx["sending"]["chainId"],
            "receiving_chain": tx["receiving"]["chainId"],
            "sending_token_symbol": tx["sending"]["token"]["symbol"],
            "receiving_token_symbol": tx["receiving"]["token"]["symbol"],
            "sending_amount": sending_amount,
            "receiving_amount": receiving_amount,
            "sending_amount_usd": float(tx["sending"]["amountUSD"]),
            "receiving_amount_usd": float(tx["receiving"]["amountUSD"]),
            "sending_gas_used": int(tx["sending"]["gasUsed"]),
            "receiving_gas_used": int(tx["receiving"]["gasUsed"]),
            "sending_timestamp": sending_timestamp,
            "receiving_timestamp": receiving_timestamp,
            "date": sending_timestamp.date(),  # Group by day
            "week": sending_timestamp.strftime('%Y-%W')  # Group by week
        })
    
    df = pd.DataFrame(rows)
    return df

# Function to create visualizations based on the metrics
def create_visualizations():
    transactions_data = fetch_transactions()
    df = process_transactions(transactions_data)

    # Ensure that chain IDs are strings for consistent grouping
    df["sending_chain"] = df["sending_chain"].astype(str)
    df["receiving_chain"] = df["receiving_chain"].astype(str)

    # 1. Number of Transactions per Chain per Day per Agent
    tx_per_chain_agent = df.groupby(["date", "from_address", "sending_chain"]).size().reset_index(name="transaction_count")
    fig_tx_chain_agent = px.bar(tx_per_chain_agent, x="date", y="transaction_count", color="sending_chain", barmode="group",
                                facet_col="from_address", title="Number of Transactions per Chain per Agent per Day")

    # 2. Number of Opportunities Taken per Agent per Day
    opportunities_per_agent = df.groupby(["date", "from_address"]).size().reset_index(name="opportunities_taken")
    fig_opportunities_agent = px.bar(opportunities_per_agent, x="date", y="opportunities_taken", color="from_address",
                                     title="Number of Opportunities Taken per Agent per Day")

    # 3. Amount of Investment in Pools Daily per Agent (Note: Assuming sending_amount_usd as investment)
    # Since we might not have explicit data about pool investments, we'll use sending_amount_usd
    investment_per_agent = df.groupby(["date", "from_address"])["sending_amount_usd"].sum().reset_index()
    fig_investment_agent = px.bar(investment_per_agent, x="date", y="sending_amount_usd", color="from_address",
                                  title="Amount of Investment (USD) per Agent per Day")

    # 4. Number of Swaps per Day
    # Assuming each transaction is a swap if sending and receiving tokens are different
    df["is_swap"] = df.apply(lambda x: x["sending_token_symbol"] != x["receiving_token_symbol"], axis=1)
    swaps_per_day = df[df["is_swap"]].groupby("date").size().reset_index(name="swap_count")
    fig_swaps_per_day = px.bar(swaps_per_day, x="date", y="swap_count", title="Number of Swaps per Day")

    # 5. Aggregated Metrics over All Traders
    amount_usd = df["sending_amount_usd"]
    stats = {
        "Total": amount_usd.sum(),
        "Average": amount_usd.mean(),
        "Min": amount_usd.min(),
        "Max": amount_usd.max(),
        "25th Percentile": amount_usd.quantile(0.25),
        "50th Percentile (Median)": amount_usd.median(),
        "75th Percentile": amount_usd.quantile(0.75),
    }
    stats_df = pd.DataFrame(list(stats.items()), columns=["Metric", "Value"])

    # Visualization for Aggregated Metrics
    fig_stats = px.bar(stats_df, x="Metric", y="Value", title="Aggregated Transaction Amount Metrics (USD)")

    return fig_tx_chain_agent, fig_opportunities_agent, fig_investment_agent, fig_swaps_per_day, fig_stats

# Gradio interface
def dashboard():
    with gr.Blocks() as demo:
        gr.Markdown("# Valory Transactions Dashboard")

        # Fetch and display visualizations
        with gr.Tab("Transactions per Chain per Agent"):
            fig_tx_chain_agent, _, _, _, _ = create_visualizations()
            gr.Plot(fig_tx_chain_agent)
        
        with gr.Tab("Opportunities per Agent"):
            _, fig_opportunities_agent, _, _, _ = create_visualizations()
            gr.Plot(fig_opportunities_agent)
        
        with gr.Tab("Investment per Agent"):
            _, _, fig_investment_agent, _, _ = create_visualizations()
            gr.Plot(fig_investment_agent)
        
        with gr.Tab("Swaps per Day"):
            _, _, _, fig_swaps_per_day, _ = create_visualizations()
            gr.Plot(fig_swaps_per_day)
        
        with gr.Tab("Aggregated Metrics"):
            _, _, _, _, fig_stats = create_visualizations()
            gr.Plot(fig_stats)
    
    return demo

# Launch the dashboard
if __name__ == "__main__":
    dashboard().launch()