from gradio_client import Client
from sklearn.datasets import load_linnerud
import pandas as pd
import numpy as np
from time import time

X, y = load_linnerud(return_X_y=True, as_frame=True)

# create a dataframe with 1000 randomly generated values for predicting
rng = np.random.default_rng(42)
num_pred = 10
X_pred = pd.DataFrame(
    {
        "Chins": 50 * rng.random(num_pred),
        "Situps": 80 * rng.random(num_pred),
        "Jumps": 20 * rng.random(num_pred),
    }
)

client = Client("AccelerationConsortium/sklearn-train-basic")

t0 = time()
result = client.predict(
    {
        "headers": X_pred.columns.tolist(),
        "data": X_pred.values.tolist(),
    },  # Dict(headers: List[str], data: List[List[Any]], metadata: Dict(str, List[Any] | None) | None) in 'X' Dataframe component
    api_name="/predict",
)

print(f"Time taken: {time() - t0:.2f}s")

result_df = pd.DataFrame(result["data"], columns=result["headers"])

print(result_df)