Find Median from a Data Stream

Question

## Problem

Design a data structure that supports:
- `addNum(int num)` — add a number from the stream
- `findMedian()` — return the median of all numbers added so far

## Example
```
addNum(1), addNum(2) → findMedian() = 1.5
addNum(3)            → findMedian() = 2
```

## Constraints
- -10^5 ≤ num ≤ 10^5
- Up to 5 × 10^4 calls

## Expected approach
Two heaps: a max-heap for the lower half and a min-heap for the upper half. Rebalance after every insert.

## Follow-up
If 99% of numbers are in [0, 100], how would you optimise memory?