app.py

import os
import gradio as gr
import pandas as pd
import plotly.express as px
import re
import io
import subprocess

# Ensure the script is executable
os.system("chmod +x gpu_info_collector.sh")

# ==========================================
# 1. Define function to run the script
# ==========================================
def run_shell_script(secret_key):
    # Security check: Verify the secret key to prevent unauthorized execution.
    # Note: Set "RUN_KEY" in Space Settings -> Variables and secrets.
    expected_key = os.environ.get("RUN_KEY")
    
    if not expected_key:
        return "❌ Auth failed: RUN_KEY environment variable is not configured on the server!"
        
    if secret_key != expected_key:
        return "❌ Auth failed: Incorrect secret key!"

    print("Command received, starting script execution...")
    
    # Execute the .sh file
    try:
        result = subprocess.run(
            ["./gpu_info_collector.sh"], 
            shell=True, 
            capture_output=True, 
            text=True
        )
        log_output = f"Standard Output:\n{result.stdout}\n\nError Output:\n{result.stderr}"
        print(log_output) 
        return f"✅ Script execution completed!\n{log_output}"
    except Exception as e:
        return f"⚠️ Execution error: {str(e)}"

# ==========================================
# 2. Data Reading Engine
# ==========================================

def clean_and_read_file(file_path):
    if not file_path or not os.path.exists(file_path):
        return pd.DataFrame()

    # --- Strategy A: Try reading as Excel ---
    try:
        df = pd.read_excel(file_path)
        return df
    except Exception:
        pass

    # --- Strategy B: Read as Text ---
    raw_data = b""
    try:
        with open(file_path, 'rb') as f:
            raw_data = f.read()
    except Exception as e:
        print(f"File read error: {e}")
        return pd.DataFrame()

    # Decode
    content = ""
    for enc in ['utf-8', 'gb18030', 'gbk']:
        try:
            content = raw_data.decode(enc)
            break
        except UnicodeDecodeError:
            continue
    if not content:
        content = raw_data.decode('utf-8', errors='replace')

    # --- Cleaning ---
    content = re.sub(r"\\", "", content)

    lines = content.splitlines()
    cleaned_lines = []
    buffer = ""
    date_pattern = re.compile(r'^\s*202\d-\d{2}-\d{2}')

    for line in lines:
        line = line.strip()
        if not line:
            continue

        is_header = "Date" in line and ("," in line)
        is_date_row = date_pattern.match(line) is not None

        if is_header or is_date_row:
            if buffer:
                cleaned_lines.append(buffer)
            buffer = line
        else:
            buffer += " " + line

    if buffer:
        cleaned_lines.append(buffer)

    csv_content = "\n".join(cleaned_lines)
    try:
        df = pd.read_csv(io.StringIO(csv_content))
    except Exception:
        try:
            df = pd.read_csv(io.StringIO(csv_content),
                             sep=None,
                             engine='python')
        except Exception:
            return pd.DataFrame()

    return df

# ==========================================
# 3. Data Processing
# ==========================================

def process_gpu_data(df):
    if df.empty:
        return df
    df.columns = [str(c).strip() for c in df.columns]

    if 'Date' in df.columns:
        df['Date'] = pd.to_datetime(df['Date'], errors='coerce')

    def clean_currency(x):
        if isinstance(x, (int, float)):
            return float(x)
        if isinstance(x, str):
            match = re.search(r'(\d+\.?\d*)', x)
            return float(match.group(1)) if match else 0.0
        return 0.0

    target_col = None
    if 'Cloud Rent (/hr)' in df.columns:
        target_col = 'Cloud Rent (/hr)'
    else:
        for c in df.columns:
            if 'Rent' in c or '/hr' in c:
                target_col = c
                break

    if target_col:
        df['Rent_Price_Num'] = df[target_col].apply(clean_currency)

    return df

def process_llm_data(df):
    if df.empty:
        return df
    df.columns = [str(c).strip() for c in df.columns]

    if 'Date' in df.columns:
        df['Date'] = pd.to_datetime(df['Date'], errors='coerce')

    return df

# ==========================================
# 4. Plotting Logic
# ==========================================

def plot_gpu_trends(df):
    if df is None or df.empty or 'Rent_Price_Num' not in df.columns:
        return None

    plot_df = df.dropna(subset=['Date', 'Rent_Price_Num'])
    if plot_df.empty:
        return None

    # Defensive fix: Prevent Index out of bounds if df columns are insufficient
    chip_col = 'Chip' if 'Chip' in df.columns else (df.columns[1] if len(df.columns) > 1 else None)

    fig = px.line(plot_df,
                  x='Date',
                  y='Rent_Price_Num',
                  color=chip_col if chip_col in df.columns else None,
                  title='GPU Cloud Rental Price Trends ($/hr)',
                  labels={
                      'Rent_Price_Num': 'Price ($/hr)',
                      'Date': 'Date'
                  },
                  markers=True)
    return fig

def plot_llm_trends(df):
    if df is None or df.empty:
        return None

    value_vars = [c for c in df.columns if c != 'Date']
    if not value_vars:
        return None

    plot_df = df[['Date'] + value_vars].copy().dropna(subset=['Date'])

    df_long = plot_df.melt(id_vars=['Date'], var_name='Model', value_name='Price')

    fig = px.line(
        df_long,
        x='Date',
        y='Price',
        color='Model',
        title='LLM API Price Trends',
        labels={'Price': 'Price', 'Date': 'Date', 'Model': 'Model Type'},
        markers=True
    )
    return fig

# ==========================================
# 5. Gradio Interface
# ==========================================

DEFAULT_GPU_FILE = "gpu_price_history.csv"
DEFAULT_LLM_FILE = "llm_price_trends.csv"

def load_gpu_pipeline():
    df = clean_and_read_file(DEFAULT_GPU_FILE)
    df = process_gpu_data(df)
    return df, plot_gpu_trends(df)

def load_llm_pipeline():
    df = clean_and_read_file(DEFAULT_LLM_FILE)
    df = process_llm_data(df)
    return df, plot_llm_trends(df)

# --- UI Definition ---
with gr.Blocks(title="AI Price Tracker") as demo:
    gr.Markdown("## 📊 AI Compute & Model Price Trends")

    with gr.Tabs():
        # GPU Tab
        with gr.TabItem("GPU Prices"):
            with gr.Row():
                with gr.Column(scale=1):
                    gpu_plot = gr.Plot(label="Price Trend")
            with gr.Row():
                with gr.Accordion("Data Preview", open=False):
                    gpu_table = gr.DataFrame()

        # LLM Tab
        with gr.TabItem("LLM Prices"):
            with gr.Row():
                with gr.Column(scale=1):
                    llm_plot = gr.Plot(label="Price Trend")

            with gr.Row():
                with gr.Accordion("Data Preview", open=False):
                    llm_table = gr.DataFrame()

    # Hidden components to expose the API safely without breaking UI
    api_input = gr.Textbox(visible=False)
    api_output = gr.Textbox(visible=False)
    api_trigger = gr.Button(visible=False)
    
    api_trigger.click(
        fn=run_shell_script,
        inputs=[api_input],
        outputs=[api_output],
        api_name="run_collector"
    )

    # --- Initialization Logic ---
    def init_on_load():
        g_df, g_fig = load_gpu_pipeline()
        l_df, l_fig = load_llm_pipeline()
        return g_fig, g_df, l_fig, l_df

    demo.load(
        init_on_load,
        inputs=None,
        outputs=[gpu_plot, gpu_table, llm_plot, llm_table]
    )

if __name__ == "__main__":
    demo.launch(share=True)