A memory-efficient spell checker built using a Bloom filter — a probabilistic data structure for fast set membership checks with minimal memory usage. This command-line tool allows you to test whether a word might exist in a dictionary.
Based on the Bloom filter algorithm.
- Fast constant-time insertions and lookups
- Probabilistic membership checking (no false negatives, possible false positives)
- Bloom filter persistence (save/load from JSON)
- Automatic parameter recommendation based on expected item count and desired false positive rate
- Set operations: union, intersection
- CLI interface and optional metadata output
- Logging and unit tests for development and debugging
Written in Python, using xxhash for fast, consistent hashing and numpy for efficiency.
Create a virtual environment and install dependencies:
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txtRun the spell checker from the command line:
python main.py --dict dict.txt
You can also save or load filters:
python main.py --dict dict.txt --save bloom.json
python main.py --load bloom.json
Interactive CLI session:
Type words to check. Type 'exit' to quit.
Check word: pinapple
Might exist!
Check word: bubblebee
Not found!A Bloom filter is a space-efficient probabilistic data structure used to test whether an element is likely to be a member of a set.
- It may return false positives, but never false negatives.
- Commonly used in caching, databases, search engines, and distributed systems.
from bloom import BloomFilter
from recommender import BloomParamRecommender
# Create a bloom filter with a size of 1000 bits and 4 hash functions
bf = BloomFilter(size=1000, hash_num=4)
# Add a word
bf.add("apple")
# Query membership
print(bf.query("apple")) # likely true
print(bf.query("banana")) # likely false
# Get an estimate number of inserted items
print(bf.estimate_num_of_items())
# Combine two filters (union)
bf2 = BloomFilter(size=1000, hash_num=4)
bf2.add("banana")
combined = bf.union(bf2)