app.py

import gradio as gr
import pandas as pd
import numpy as np
from xgboost import XGBClassifier

# =============== Your feature engineering functions ===============
def calculate_angle(x1, y1, x2, y2, x3, y3):
    v1 = [x2 - x1, y2 - y1]
    v2 = [x3 - x2, y3 - y2]
    dot = v1[0]*v2[0] + v1[1]*v2[1]
    mag1 = np.sqrt(v1[0]**2 + v1[1]**2)
    mag2 = np.sqrt(v2[0]**2 + v2[1]**2)
    
    if mag1 == 0 or mag2 == 0:
        return 0.0
    cos_theta = np.clip(dot / (mag1 * mag2), -1.0, 1.0)
    angle = np.degrees(np.arccos(cos_theta))
    return angle

def label_angle(angle):
    if np.isnan(angle):    
        return 0
    if angle < 30:
        return 0
    elif angle < 60:
        return 1
    elif angle < 90:
        return 2
    elif angle < 120:
        return 3
    elif angle < 150:
        return 4
    else:
        return 5

def extract_features(extracted, time_diff=0.05):
    extracted['distance_covered'] = 0.0
    extracted['idle_time'] = 0.0

    for i in range(1, len(extracted)):
        dx = extracted.loc[i, 'x'] - extracted.loc[i-1, 'x']
        dy = extracted.loc[i, 'y'] - extracted.loc[i-1, 'y']
        distance = np.sqrt(dx**2 + dy**2)
        extracted.loc[i, 'distance_covered'] = distance

        if extracted.loc[i, 'x'] == extracted.loc[i-1, 'x'] and extracted.loc[i, 'y'] == extracted.loc[i-1, 'y']:
            extracted.loc[i, 'idle_time'] = extracted.loc[i-1, 'idle_time'] + time_diff
        else:
            extracted.loc[i, 'idle_time'] = 0.0

    extracted['cursor_speed'] = extracted['distance_covered'] / time_diff
    extracted['acceleration'] = extracted['cursor_speed'] / time_diff

    angles = []
    for i in range(1, len(extracted) - 1):
        angle = calculate_angle(
            extracted.loc[i-1, 'x'], extracted.loc[i-1, 'y'],
            extracted.loc[i, 'x'], extracted.loc[i, 'y'],
            extracted.loc[i+1, 'x'], extracted.loc[i+1, 'y']
        )
        angles.append(angle)

    extracted = extracted.iloc[1:-1].copy()
    extracted['movement_angle'] = angles
    extracted['prev_movement_angle'] = [0] + angles[1:]

    extracted['angle_label'] = extracted['movement_angle'].apply(label_angle)
    extracted['prev_angle_label'] = extracted['prev_movement_angle'].apply(label_angle)
    return extracted

# =============== Load model ===============
model = XGBClassifier()
model.load_model("xgb_confusion_detector.model")

# =============== Gradio prediction function ===============
def predict_fn(file):
    # file is a pandas-readable file (csv, etc.)
    raw_df = pd.read_csv(file)

    extracted = extract_features(raw_df, time_diff=0.05)

    try:
        features_to_use = extracted.drop(['isConfused'], axis=1)
    except:
        features_to_use = extracted

    prediction = model.predict(features_to_use)
    total_predictions = len(prediction)
    confused_predictions = np.sum(prediction == 1)

    confusion_ratio = confused_predictions / total_predictions
    is_user_confused = confusion_ratio > 0.3

    return {
        "Prediction": prediction.tolist(),
        "Confidence": model.predict_proba(features_to_use).tolist(),
        "Confusion Ratio": confusion_ratio,
        "User Confused?": is_user_confused
    }

# =============== Gradio Interface ===============
interface = gr.Interface(
    fn=predict_fn,
    inputs=gr.File(label="Upload cursor movement CSV"),
    outputs=[
        gr.JSON(label="Prediction Details")
    ],
    title="Confusion Detector",
    description="Upload a CSV file with cursor movements (timestamp,x,y,isClick) to predict if the user is confused."
)

interface.launch()