FuncCraft

FuncCraft is a Python package backed by a C++17 benchmark-function generator for black-box optimization research. It is designed for scalable benchmark suite generation across dimensions: one editable suite YAML file can generate hundreds, thousands, or practically unlimited numbers of distinct benchmark functions while controlling the constructed optimum location and optimum value.

Install

python -m pip install --upgrade funccraft
python -m pip install numpy scipy minionpy

YAML-First Workflow

Suite YAML is the easiest way to configure FuncCraft from files. The same structure can also be passed as a Python dictionary to BenchmarkSuite.

supported_dimensions: any
base_functions: [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34]
composition_base_functions: [4, 8, 10, 11, 12, 13, 15, 16, 17, 18, 19, 20, 21, 23, 28, 30, 34, 33, 2, 3, 5, 9, 26, 27]
coordinate_transforms:
  - kind: none
    probability: 0.0
    parameters: []
  - kind: rotation
    probability: 0.34
    parameters: []
  - kind: affine
    probability: 0.33
    parameters: []
  - kind: subspace-rotation
    probability: 0.33
    parameters: []
value_transforms:
  - kind: none
    probability: 0.5
    parameters: []
  - kind: power
    probability: 0.25
    parameters: []
  - kind: oscillatory
    probability: 0.25
    parameters: []
  - kind: cosine-zero
    probability: 0.0
    parameters: []
compositions:
  - kind: cpm-wsum
    probability: 0.1
    parameters: []
  - kind: cpm-power-mean
    probability: 0.1
    parameters: [3.0]
  - kind: cpm-power-mean
    probability: 0.1
    parameters: [0.1]
  - kind: cpm-level-well
    probability: 0.2
    parameters: []
  - kind: dpm-softmax
    probability: 0.25
    parameters: [0.005]
  - kind: dpm-bgsoftmax
    probability: 0.25
    parameters: [0.005, 1.0, 0.01]
min_components: 2
max_components: 5
max_nested_composition_depth: 1
nested_probability: 0.1
requested_number_of_functions: 500
master_seed: 1
lower_bound: -100
upper_bound: 100
assigned_fopt: 100.0
xopt_domain_shrink_factor: 0.8
suite_label: my-suite

Load the YAML and evaluate a function. Function indices are one-based:

import funccraft as fc

dimension = 10
function_index = 1

suite = fc.BenchmarkSuite("my_suite.yaml", dimension=dimension)
f = suite.function(function_index)

points = [[0.0] * dimension, [1.0] * dimension]
values = f.evaluate(points)
print(values)

The packaged 2026 suite is also YAML-backed and exposed through a shortcut:

import funccraft as fc

dimension = 10
function_index = 1
year = 2026
version = 1
suite = fc.SuiteCollection(year=year, version=version).benchmark_suite(dimension)
f = suite.function(function_index)
values = f.evaluate([[0.0] * dimension])

All evaluations are batched: pass a list of points, not one flat point vector.

Optimization

SciPy:

import numpy as np
from scipy.optimize import differential_evolution
import funccraft as fc

dimension = 10
function_index = 1
year = 2026
version = 1
suite = fc.SuiteCollection(year=year, version=version).benchmark_suite(dimension)
f = suite.function(function_index)
domain = f.domain
bounds = list(zip(domain.lower_bound, domain.upper_bound))

def objective(x):
    return f.evaluate([np.asarray(x, dtype=float).tolist()])[0]

result = differential_evolution(objective, bounds, seed=1, polish=False)
print(result.x, result.fun)

MinionPy:

import funccraft as fc
import minionpy as mpy

dimension = 10
function_index = 1
year = 2026
version = 1
suite = fc.SuiteCollection(year=year, version=version).benchmark_suite(dimension)
f = suite.function(function_index)
domain = f.domain

optimizer = mpy.Minimizer(
    func=f.evaluate,
    x0=[
        [0.0] * dimension,
        [1.0] * dimension,
        [-0.5] * dimension,
    ],
    bounds=list(zip(domain.lower_bound, domain.upper_bound)),
    algo="ARRDE",
    maxevals=10000,
    callback=None,
    seed=None,
    options=None,
)
result = optimizer.optimize()
print(result.x, result.fun)

Mechanism Summary

FuncCraft builds functions from primitive benchmark landscapes, coordinate transforms, value transforms, and composition rules:

f(x) = assigned_fopt + scale_factor * psi(x, z_1(x), ..., z_m(x))
z_i(x) = component_scale_i * phi_i(g_i(T_i(x)) - f_i*)

Omitted component scales and the final scale_factor are estimated deterministically at construction time and materialized in exported specs.

Implemented mechanism families include:

34 primitive base functions, including Sphere, Rosenbrock, Ackley, Rastrigin, Griewank, Schwefel, Katsuura, Levy, BentCigar, HappyCat, HGBat, and StyblinskiTang.
coordinate transforms: none, rotation, affine, subspace-rotation.
value transforms: none, power, oscillatory, cosine-zero, huber, log, softplus-threshold, dead-zone, saturating, piecewise-power, noisy-smooth.
compositions: none, cpm-wsum, cpm-power-mean, cpm-level-well, cpm-max, cpm-smoothmax, cpm-constraint-penalty, cpm-lexicographic, cpm-product, cpm-max-plus-mean, cpm-cvar, dpm-softmax, dpm-bgsoftmax.

Components can also be nested composed functions, up to the configured max_nested_composition_depth.

Names are parsed permissively: case, spaces, hyphens, and underscores are normalized before matching.

Exported Manifests

Input YAML is for configuration and editing. Exported YAML is a materialized record of what FuncCraft built:

f.export_yaml("function_materialized.yaml")
suite.export_manifest("suite_manifest.yaml")

Exported YAML records include generated matrices, selected subspaces, DPM centers and biases, assigned optima, scale factors, labels, and metadata.

Warning: the packaged 2026_v1 suite contains 1,000,000 functions. Exporting the full shipped suite manifest will write every generated function record and is usually unnecessary. To export a smaller reproducible subset, copy the collection configuration dictionary, reduce requested_number_of_functions, build the smaller suite, and export that suite:

import funccraft as fc

collection = fc.SuiteCollection(year=2026, version=1)
config = collection.config
config["requested_number_of_functions"] = 500

suite = fc.BenchmarkSuite(config, dimension=2)
suite.export_manifest("suite_2026_v1_first_500.yaml")

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distributions

No source distribution files available for this release.See tutorial on generating distribution archives.

Built Distributions

If you're not sure about the file name format, learn more about wheel file names.

The dropdown lists show the available interpreters, ABIs, and platforms. Enable javascript to be able to filter the list of wheel files.

funccraft-0.3.1-cp314-cp314-win_amd64.whl (661.8 kB view details)

Uploaded Aug 1, 2026 CPython 3.14Windows x86-64

funccraft-0.3.1-cp314-cp314-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (813.5 kB view details)

Uploaded Aug 1, 2026 CPython 3.14manylinux: glibc 2.27+ x86-64manylinux: glibc 2.28+ x86-64

funccraft-0.3.1-cp314-cp314-macosx_11_0_arm64.whl (487.4 kB view details)

Uploaded Aug 1, 2026 CPython 3.14macOS 11.0+ ARM64

funccraft-0.3.1-cp314-cp314-macosx_10_15_x86_64.whl (520.0 kB view details)

Uploaded Aug 1, 2026 CPython 3.14macOS 10.15+ x86-64

funccraft-0.3.1-cp313-cp313-win_amd64.whl (640.7 kB view details)

Uploaded Aug 1, 2026 CPython 3.13Windows x86-64

funccraft-0.3.1-cp313-cp313-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (813.5 kB view details)

Uploaded Aug 1, 2026 CPython 3.13manylinux: glibc 2.27+ x86-64manylinux: glibc 2.28+ x86-64

funccraft-0.3.1-cp313-cp313-macosx_11_0_arm64.whl (486.8 kB view details)

Uploaded Aug 1, 2026 CPython 3.13macOS 11.0+ ARM64

funccraft-0.3.1-cp313-cp313-macosx_10_15_x86_64.whl (519.5 kB view details)

Uploaded Aug 1, 2026 CPython 3.13macOS 10.15+ x86-64

funccraft-0.3.1-cp312-cp312-win_amd64.whl (640.6 kB view details)

Uploaded Aug 1, 2026 CPython 3.12Windows x86-64

funccraft-0.3.1-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (813.5 kB view details)

Uploaded Aug 1, 2026 CPython 3.12manylinux: glibc 2.27+ x86-64manylinux: glibc 2.28+ x86-64

funccraft-0.3.1-cp312-cp312-macosx_11_0_arm64.whl (486.7 kB view details)

Uploaded Aug 1, 2026 CPython 3.12macOS 11.0+ ARM64

funccraft-0.3.1-cp312-cp312-macosx_10_15_x86_64.whl (519.4 kB view details)

Uploaded Aug 1, 2026 CPython 3.12macOS 10.15+ x86-64

funccraft-0.3.1-cp311-cp311-win_amd64.whl (639.9 kB view details)

Uploaded Aug 1, 2026 CPython 3.11Windows x86-64

funccraft-0.3.1-cp311-cp311-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (811.5 kB view details)

Uploaded Aug 1, 2026 CPython 3.11manylinux: glibc 2.27+ x86-64manylinux: glibc 2.28+ x86-64

funccraft-0.3.1-cp311-cp311-macosx_11_0_arm64.whl (486.6 kB view details)

Uploaded Aug 1, 2026 CPython 3.11macOS 11.0+ ARM64

funccraft-0.3.1-cp311-cp311-macosx_10_15_x86_64.whl (516.1 kB view details)

Uploaded Aug 1, 2026 CPython 3.11macOS 10.15+ x86-64

funccraft-0.3.1-cp310-cp310-win_amd64.whl (638.8 kB view details)

Uploaded Aug 1, 2026 CPython 3.10Windows x86-64

funccraft-0.3.1-cp310-cp310-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (809.7 kB view details)

Uploaded Aug 1, 2026 CPython 3.10manylinux: glibc 2.27+ x86-64manylinux: glibc 2.28+ x86-64

funccraft-0.3.1-cp310-cp310-macosx_11_0_arm64.whl (485.6 kB view details)

Uploaded Aug 1, 2026 CPython 3.10macOS 11.0+ ARM64

funccraft-0.3.1-cp310-cp310-macosx_10_15_x86_64.whl (514.7 kB view details)

Uploaded Aug 1, 2026 CPython 3.10macOS 10.15+ x86-64

funccraft-0.3.1-cp39-cp39-win_amd64.whl (655.4 kB view details)

Uploaded Aug 1, 2026 CPython 3.9Windows x86-64

funccraft-0.3.1-cp39-cp39-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl (809.9 kB view details)

Uploaded Aug 1, 2026 CPython 3.9manylinux: glibc 2.27+ x86-64manylinux: glibc 2.28+ x86-64

funccraft-0.3.1-cp39-cp39-macosx_11_0_arm64.whl (485.7 kB view details)

Uploaded Aug 1, 2026 CPython 3.9macOS 11.0+ ARM64

funccraft-0.3.1-cp39-cp39-macosx_10_15_x86_64.whl (514.8 kB view details)

Uploaded Aug 1, 2026 CPython 3.9macOS 10.15+ x86-64

File details

Details for the file funccraft-0.3.1-cp314-cp314-win_amd64.whl.

File metadata

Download URL: funccraft-0.3.1-cp314-cp314-win_amd64.whl
Upload date: Aug 1, 2026
Size: 661.8 kB
Tags: CPython 3.14, Windows x86-64
Uploaded using Trusted Publishing? No
Uploaded via: twine/6.2.0 CPython/3.12.10

File hashes

Hashes for funccraft-0.3.1-cp314-cp314-win_amd64.whl
Algorithm	Hash digest
SHA256	`537d6316526ea12afbf240507e0f3d25e6e7093ce2ac3eda47942d19d94660ae`
MD5	`627b8c7a74b6b910c1642968cbd5c0bd`
BLAKE2b-256	`bb4e1aa0a4e04457e7a605510ba02a16cfb5e6cc12f0f41e8fc76d667ca5a4d1`

Algorithm	Hash digest
SHA256	`46c870b146e4a6e463695fc9574513898d14c1c324118643bcaa7215468df940`
MD5	`93f4afed4989c2842d2083ba52a58ebe`
BLAKE2b-256	`d7fb353693b49bc90580135bdccdaf66a2e7c549c65bdf4688f7ed0bae2dd616`

Algorithm	Hash digest
SHA256	`09f41b43c6a3386fd7b121cd0937453938bd8a666ede4b0cbfc080136fd02e47`
MD5	`8c23b3d4174b0d55f69a75253006f261`
BLAKE2b-256	`2e565bed97472095e6195a9e68a4a8d84899cdbd072bea056406e8b75e6d4dad`

Algorithm	Hash digest
SHA256	`76edf84b584337cbdcfd7c0542f72ef028afe40175d496dd43933f93cf313a22`
MD5	`7b3ae4784788717f67887c75e243a042`
BLAKE2b-256	`4f77a8d339955c710cc8627c819e104dae6282b0a9a5daa9e2b1971b8fb04bcf`

Algorithm	Hash digest
SHA256	`a2adb7869eaeb3bbeb4e38d464ff244ba8f5eb5659ca44f315c6074047b52a75`
MD5	`b6d3f68a004633148e1673e075249e25`
BLAKE2b-256	`2e9ae53dbbe27ccc7676c144bfe4cb1676547735e0291ec30331a9637af7e97b`

Algorithm	Hash digest
SHA256	`bdcbe237e222647c2878f16c89136569f49459d4c13494b2f023cf8925c16c3b`
MD5	`ef09cddbfd19882acd0031505f614ba8`
BLAKE2b-256	`4f837d653855367c2af0276e3c666c6d86a2250a228eecbb9366dcf0e98af042`

Algorithm	Hash digest
SHA256	`8f747e3aed9afe848e0adb88cfbe55e83f7e02d426c19e236cd4cbc93a1a03f6`
MD5	`8368c16289354c3ebe33babb561d059b`
BLAKE2b-256	`c480cea1d871a7efb0e7e41848bfa42d8c00029948d359f3cfad3fea30123327`

Algorithm	Hash digest
SHA256	`c897ae14fd97d6053059135891f729917b179c52e2ccc23e741480e18fcd6f3f`
MD5	`432d9c3bf63389764523c4c0137d00e2`
BLAKE2b-256	`9ccc9e2b44021a4a90b78828e246b7bc00d24029cca2a7900982fb30cadb4f3a`

Algorithm	Hash digest
SHA256	`c402e2d0dfb99439777c884a464c078e91dff03b86510787ef130413f4cded91`
MD5	`89356672e1d15df59aa2051e180561e0`
BLAKE2b-256	`1db675cf423af8a8b72450a690d7437fd22028b9737343625af2ae833a062315`

Algorithm	Hash digest
SHA256	`263234d713e134a875168264aa07ee0229b7b4189d3679b2378a26521a76502d`
MD5	`b3d36d3f5782b15bdbc049bdc9b3d790`
BLAKE2b-256	`f150aa2ab284cdaad7cebee5eb2414cb35ec0f8f30dfb478d7aa3e27238098c7`

Algorithm	Hash digest
SHA256	`a8a46bdcf47f23d644e2e6fe3109977a94a56b1e796c3cc58e4c810fae85c39d`
MD5	`198173d570fdbdd5263c3a37a52e324f`
BLAKE2b-256	`c2dd2040657faf9591c1fd5d1076faca3f4596ab6febc88ac9d2d22b0392aac9`

Algorithm	Hash digest
SHA256	`adb5ff439344968602af81c2e0208a338616c42e0dce5a9259a8725ebea823aa`
MD5	`8b74b1297b4cf0c08f60f25840dad75e`
BLAKE2b-256	`746500b6903e2b4998dfea8d6c1817712eaa8d091347f46f3d4b4f2815164878`

Algorithm	Hash digest
SHA256	`e2d72060d4a45369060c2fcbea71487b7769ed35afde6698335afa589d35bc84`
MD5	`d76a1f439e2a2d38221b02e2feaf85a3`
BLAKE2b-256	`69600e5618ab3f72da0e616e3f186485b557efff99e1c4cf3ec22d3d6c1c235d`

Algorithm	Hash digest
SHA256	`3167e4fe3d33cd04a71f3b32a15db9b408b521c08cfc2460fce2b8f9d57dccb8`
MD5	`d03e4f74930319b6c774945652f827a4`
BLAKE2b-256	`f5c17012ffce5390bea4757d056e2585785ae7789f6775c1156985eff82029d5`

Algorithm	Hash digest
SHA256	`59cacd146528707a708508917b0cab088a4c1b4484334978e91d33e5fb76ca34`
MD5	`7fd8ed27022ee68407c784145bc52ef0`
BLAKE2b-256	`238cd3f052f2aab2f2c49d6084b97343188464f2cb4805a39b3d029685a8b307`

Algorithm	Hash digest
SHA256	`d8ceac55be1c70ebcae9201e62353434a00209c33e7d7d63c66bcb5ac0763c63`
MD5	`e75108943cc9855826d1524d2d240d74`
BLAKE2b-256	`2ae7be4e53c90717e4dc9de61e63d51b97df0620bf26614f59c301b97d9320ec`

Algorithm	Hash digest
SHA256	`40e3c44a8a928f88fbffd2cc52a1c7d9dcb3df3f0e10cdc35d3c99d2bd195659`
MD5	`6ab20557a361afd8fa244f77bfd9816d`
BLAKE2b-256	`d2d6f39a1dcb83e0dbd3d07544a3dc34b4c4776c613c48f8d3f171a607cf0289`

Algorithm	Hash digest
SHA256	`f87e9d970e555794901161287731d72c6588334aa705f0ad9bc9e26796469072`
MD5	`db6d17e43012b70c6351afd4218a795c`
BLAKE2b-256	`f166115ce99b24c7450e9408f8262c9f4ce7d80e8c3e8d00f0e981bd38b67b6a`

Algorithm	Hash digest
SHA256	`c1543d17a0a7b58c27750faf4073254acdac5d002913c428bf225337c19efbdb`
MD5	`380ec5e72195ce1440d78f92a0d11683`
BLAKE2b-256	`ab92fa044502b3cf6466709a5cddba46c2da5990159d7e945918114c231fcdca`

Algorithm	Hash digest
SHA256	`db3827a3ea7bc20fc1f01617a8fc0a82b60d3a1aadd5bbd967ecb43adb24b377`
MD5	`68ebcca393a7bd7e972a5ce37350baa9`
BLAKE2b-256	`10be3fb1f5ba06b20fc87fab28e3b3c72c18678d8e0ed897a7848c90c9be801a`

Algorithm	Hash digest
SHA256	`1df8e2afc6dfdc794be6a6e023c8dac241d688d64b1227b343a2ab693042773b`
MD5	`6cbdabcdc290727ecef5a367d464e95c`
BLAKE2b-256	`2f9d89f37ac6fa019b989205bc32558112f5415a6959bab3838cf42fb93119fb`

funccraft 0.3.1

FuncCraft

Install

YAML-First Workflow

Optimization

Mechanism Summary

Exported Manifests

Links

Project links

Key dates

1 maintainer

Credits

Requires

Classifiers

Download files

Source Distributions

Built Distributions

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

File details

File metadata

File hashes

Algorithm	Hash digest
SHA256	`966209ae8e5082de805faa7c0eeda58712ad30b98a7b86a844fae566112944c4`
MD5	`70c8c22ceba9f79922d039aaa163d7ff`
BLAKE2b-256	`a82d22bf454fcc22e311e001ac653978c6def9837cf9551d9cd7e8ab2ede4a7c`

Algorithm	Hash digest
SHA256	`36af4020db79e63ffbfc956e243363c8c8501e33d1d3cf3cac083598ddf307f1`
MD5	`33112bc0b01101dee8207b68bae1487b`
BLAKE2b-256	`b25901150834ab1474a4879e705ff113806bcb10fc9e9f3770ee8fa3c6f2167d`

Algorithm	Hash digest
SHA256	`bde4a7d77e7e59e65423c96c3d342b98ba0703d3897e9766176fdd5d467d2a29`
MD5	`96e4cbe48863b1a4a616ded1e512a60d`
BLAKE2b-256	`e35ea3c23ffc23b1d1836d43a13a23f2df0b6e2614e500d0e932120785bd7209`