expressive 1.5.0

A library for quickly applying symbolic expressions to NumPy arrays

expressive

A library for quickly applying symbolic expressions to NumPy arrays

By enabling callers to front-load sample data, developers can move the runtime cost of Numba's JIT to the application's initial loading (or an earlier build) and also avoid exec during runtime, which is otherwise needed when lambdifying symbolic expressions

Inspired in part by this Stack Overflow Question Using numba.autojit on a lambdify'd sympy expression

installation

via pip https://pypi.org/project/expressive/

pip install expressive

usage

refer to tests for examples for now

generally follow a workflow like

• create instance expr = Expressive("a + log(b)")
• build instance expr.build(sample_data)
• instance is now callable expr(full_data)

The data should be provided as dict of NumPy arrays

sample_data = {  # types are used to compile a fast version for full data
    "a": numpy.array([1,2,3,4], dtype="int64"),
    "b": numpy.array([4,3,2,1], dtype="int64"),
}
full_data = {
    "a": numpy.array(range(1_000_000), dtype="int64"),
    "b": numpy.array(range(1_000_000), dtype="int64"),
}

testing

Only docker and docker compose (v2) are require, which are used to host/generate the test environment

Then just directly run the test script from the root of the repository, it will build the docker test environment and run itself inside it automatically

sudo apt install docker.io docker-compose-v2
./test/runtests.sh

contributing

The development process is currently private (though most fruits are available here!), largely due to this being my first public project with the potential for other users than myself, and so the potential for more public gaffes is far greater

Please refer to CONTRIBUTING.md and LICENSE.txt and feel free to provide feedback, bug reports, etc. via Issues, subject to the former

additional future intentions for contributing

• improve internal development history as time, popularity, and practicality allows
• move to parallel/multi-version/grid CI over all-in-1, single-version dev+test container
• greatly relax dependency version requirements to improve compatibility

version history

v1.5.0

• added ._repr_html_() method for improved display in Jupyter/IPython notebooks

v1.4.2

• greatly improved verify
  • numpy.allclose() takes exactly 2 arrays to compare (further args are passed to rtol, atol)
  • SymPy namespace special values oo, zoo, nan are coerced to NumPy equivalents (inf, -inf, nan)
  • raise when result is False
  • groundwork to maintain an internal collection of results
• internal symbols collection maintains IndexedBase instances (e.atoms(Symbol) returns Symbol instances)
• improve Exceptions from data that can't be used
• new custom warning helper for testing as assertWarnsRegex annoyingly eats every warning it can

v1.4.1

• more sensibly fill the result array for non-floats when not provided (only float supports NaN)

v1.4.0

• add build-time verify step to help identify math and typing issues
• some improved logic flow and improved warn()

v1.3.2 (unreleased)

• improved publishing workflow
• improved README

v1.3.1

• fix bad math related to indexing range
• add an integration test

v1.3.0

• add support for parsing equality to result
• add support for (optionally) passing result array
• hugely improve docstrings

v1.2.1

• add more detail to contributing block
• switch array dimensions checking from .shape to .ndim
• switch tests from numpy.array(range()) to numpy.arange()

v1.2.0

• enable autobuilding (skip explicit .build() call)
• basic display support for Expressive instances

v1.1.1

• add version history block

v1.1.0

• fixed bug: signature ordering could be unaligned with symbols, resulting in bad types
• added support for non-vector data arguments

v1.0.0

• completely new code tree under Apache 2 license
• basic support for indexed offsets

v0.2.0 (unreleased)

v0.1.0

• very early version with support for python 3.5