Welcome to hirola!

∘ MIT license ∘ PyPI ∘ Documentation ∘ Source code ∘ Bug reports ∘ Support

A vectorized hash table written in C for fast set/dict like operations on NumPy arrays.

Hirola provides fast indexing and de-duplication of keys. It can be used as an extension of numpy.unique() and a very light (20-30KB download size) and much faster alternative to pandas.Categorical(). Hirola obtains its speed in the same way that NumPy does – vectorising, translating into C and imposing the following constraints:

• Keys must all be of the same predetermined type and size.

• The maximum size of a table must be chosen in advance and managed explicitly.

• To get any performance boost, operations should be done in bulk.

• Elements can not be removed.

Installation

Install Hirola with pip:

pip install hirola

Usage

• Quickstart
  • HashTable
    • Choosing a max size
    • Structured key data types
    • Handling of nearly full hash tables
  • Recipes
    • Using a HashTable as a dict
    • Using a HashTable as a set
    • Using a HashTable as a collections.Counter
      • A Security Note

API Reference

• HashTable
  • HashTable
    • HashTable.__init__()
    • HashTable.add()
    • HashTable.contains()
    • HashTable.copy()
    • HashTable.destroy()
    • HashTable.get()
    • HashTable.resize()
    • HashTable.almost_full
    • HashTable.dtype
    • HashTable.key_size
    • HashTable.keys
    • HashTable.length
    • HashTable.max
• hirola.exceptions
• AlmostFull
• HashTableDestroyed
• HashTableFullError
• HirolaException

About

• Benchmarks
  • random_16
  • random_128
  • id_like
  • permutative
  • floating_32
  • floating_64
  • textual
  • utf8
• Changelog
  • v0.3.0
  • v0.2.1
  • v0.2.0
  • v0.1.0
• Contributing Guide
  • Quick Links
  • Guide to Submitting Changes
    • Setting up
    • Codebase Roadmap
    • Commands Cheatsheet
      • Compile or recompile the C code
      • Test
      • Run coverage
      • Run automatic code formatter
      • Build documentation
      • Trigger continuous integration
    • Benchmarking
    • Before Submission