Clean scores and return top 5 students

Implement a Python function to clean and rank student scores. You are given a table (or DataFrame) `students` with schema: | column | type | notes | |---|---|---| | student_id | int | unique identifier | | math_score | float | may be null | | english_score | float | may be null | | physics_score | float | may be null | ### Task Write a function (e.g., `select_top_students(students_df) -> pd.DataFrame`) that: 1. **Removes students with at least 2 missing scores** across the three subjects. 2. For remaining students, **fill missing scores with the subject median** (median computed per subject using the remaining students’ non-missing values). 3. **Sort** the remaining students by: - `math_score` descending, then - `physics_score` descending - (optional tie-breaker) `student_id` ascending. 4. **Return the top 5** students as a table with the same columns. ### Notes / edge cases to handle - Missing values may be represented as `None`/`NaN`. - If fewer than 5 students remain, return all remaining. - If a subject median is undefined (e.g., all remaining values are null for that subject), specify and implement a reasonable behavior (e.g., leave as null or raise an error)—state your assumption in comments.