Find median given frequency of numbers

# Table: Numbers
#
# +-------------+------+
# | Column Name | Type |
# +-------------+------+
# | num         | int  |
# | frequency   | int  |
# +-------------+------+
# num is the primary key (column with unique values) for this table.
# Each row of this table shows the frequency of a number in the database.
#
#
#
# The median is the value separating the higher half from the lower half of a data sample.
#
# Write a solution to report the median of all the numbers in the database after decompressing the Numbers table. Round the median to one decimal point.
#
# The result format is in the following example.
#
# Example 1:
# Input:
# Numbers table:
# +-----+-----------+
# | num | frequency |
# +-----+-----------+
# | 0   | 7         |
# | 1   | 1         |
# | 2   | 3         |
# | 3   | 1         |
# +-----+-----------+
# Output:
# +--------+
# | median |
# +--------+
# | 0.0    |
# +--------+
# Explanation:
# If we decompress the Numbers table, we will get [0, 0, 0, 0, 0, 0, 0, 1, 2, 2, 2, 3], so the median is (0 + 0) / 2 = 0.

import pandas as pd


def median_frequency(numbers: pd.DataFrame) -> pd.DataFrame:
  numbers = numbers.sort_values("num")
  numbers["cum_freq"] = numbers["frequency"].cumsum()
  total = numbers["frequency"].sum()
  lower_pos = (total + 1) // 2
  upper_pos = (total + 2) // 2
  candidates = numbers[
    (numbers["cum_freq"] >= lower_pos)
    & (numbers["cum_freq"] - numbers["frequency"] < upper_pos)
  ]
  return pd.DataFrame({"median": [round(candidates["num"].mean(), 1)]})