broomfilter

A (pretty fast) bloom filter implementation in Rust.

🤔 What is this?

A bloom filter is a probabilistic data structure that can tell you:

❌ "Definitely not in the set" — 100% accurate
🤷 "Probably in the set" — maybe, who knows

Think of it as that friend who says "I'm pretty sure I locked the door." You can trust them when they say they didn't, but when they say they did... you might want to go check.

🧹 Theory of broom sweeping (simplified)

    item ──→ [ hash hash hash ] ──→ flip bits in the array
                  🧹💨

    query ──→ [ hash hash hash ] ──→ check bits
                                      ├── all set?   → "probably yes" 🤷
                                      └── any unset? → "definitely no" ❌

💡 Why does this exist?

Because I wanted to learn how a bloom filter actually worked instead of just reciting the definition in tech interviews. Turns out its super fun to optimize!

🚀 Usage

use broomfilter::Filter;

// Create a filter with 2^14 bits, optimized for 10,000 items
let mut filter = Filter::new(14, 10_000).unwrap();

// Insert some things 🧹
filter.insert("hello");
filter.insert("world");

// Check for things
assert!(filter.contains("hello"));    // true  — probably  🤷
assert!(!filter.contains("goodbye")); // false — definitely ❌

🏗️ Status

⚠️ Work in progress — handle with care (or don't, I'm a dev, not a cop)

It works, it has benchmarks, it probably won't break. But then again, the whole point of a bloom filter is that you can never be 100% sure about anything.

Feature	Status
Insert items	✅ Swept in
Check membership	✅ Probably
False positives	✅ It's a feature
False negatives	🚫 Impossible (finally, some certainty)
Delete items	🧹 Nope, the broom only sweeps one way

📜 License

MIT — do whatever you want with it, just don't blame me when it breaks.