Spaces:
Sleeping
Sleeping
| from typing import List | |
| import numpy as np | |
| def calculate_engagement_score(likes: int, comments: int, views: int) -> float: | |
| """ | |
| Calculates the engagement score based on the formula: | |
| (Likes + Comments * 2) / Views | |
| """ | |
| if views <= 0: | |
| return 0.0 | |
| return (likes + (comments * 2)) / views | |
| def preprocess_text_features(title: str, tags: List[str]) -> str: | |
| """ | |
| Combines title and tags into a single string for TF-IDF vectorization. | |
| """ | |
| tag_str = " ".join(tags) | |
| return f"{title} {tag_str}".lower().strip() | |
| def calculate_rank_velocity(rank_history: List[int]) -> float: | |
| """ | |
| Calculates the derivative (rate of change) of the rank. | |
| Positive value = Dropping in rank (bad) | |
| Negative value = Climbing in rank (good, as 1 is top) | |
| """ | |
| if len(rank_history) < 2: | |
| return 0.0 | |
| # Simple slope calculation between first and last point | |
| return (rank_history[-1] - rank_history[0]) / len(rank_history) | |
| def encode_time_features(hour: int, day_of_week: int) -> np.ndarray: | |
| """ | |
| Example of transforming time into cyclical features (sin/cos) | |
| if you want to upgrade your models later. | |
| """ | |
| # This is a placeholder for future advanced feature engineering | |
| return np.array([hour, day_of_week]) | |