openpls-engine 0.7.0a4

PLS-SEM compute engine, a maintained fork of plspm-python with advanced analyses (IPMA, PLSpredict, moderation, FIMIX-PLS) and extended metrics (SRMR, d_ULS, HTMT, Q², MGA) powering OpenPLS.

OpenPLS Engine

The compute engine behind OpenPLS, a Python 3 library for Partial Least Squares Structural Equation Modeling (PLS-SEM).

openpls-engine is a fork of plspm-python by Jez Humble (Google). It keeps the original algorithm intact and adds the metrics and quality criteria that modern PLS-SEM reporting requires: HTMT, SRMR, d_ULS, adjusted R², BIC, Stone-Geisser Q², Cronbach α, Dijkstra-Henseler ρ, multi-group analysis, and a progress-streaming long bootstrap. It also ships advanced analyses that go beyond what plspm-python covers: PLSpredict out-of-sample validation, IPMA (Importance-Performance Map Analysis), two-stage moderation, FIMIX-PLS finite-mixture segmentation, and a quasi-Newton (BFGS) inner-weighting scheme and a Lohmöller PCA scheme, two new alternatives to the classical centroid, factorial, and path schemes (five inner-weighting schemes total).

The engine also powers OpenPLS (the hosted web application) and the CLI / Docker self-host distribution planned for the next phase of the roadmap.

Status: pre-release. All planned ports from the OpenPLS web app are in. Numerical alignment with reference implementations is tracked in TODO.md. See CHANGELOG.md for the version history.

Why fork

• Upstream has not seen a release since 2020.
• OpenPLS adds substantial extensions: SRMR, d_ULS, HTMT, Q², adjusted R², BIC, multi-group analysis (MGA), permutation tests, mean replacement, and a long-running bootstrap with BCa confidence intervals.
• The OpenPLS hosted product depends on a single source of truth for the algorithm; a maintained, versioned package makes that practical.

Installation

openpls-engine is published on PyPI. The current line is a pre-release, so opt in explicitly:

pip install --pre openpls-engine

Or pin a specific version:

pip install openpls-engine==0.7.0a3

To work from source instead:

git clone https://github.com/jojacobsen/openpls-engine.git
cd openpls-engine
python3 -m pip install -e .

Development

python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements-dev.txt
pip install -e .

pytest          # run the test suite
ruff check .    # lint

CI runs lint and tests on Python 3.10 through 3.13 against every push and pull request against main.

Versioning

openpls-engine follows Semantic Versioning. The current pre-1.0 line is 0.x. Tagged releases (v0.7.0, v1.0.0, ...) trigger a GitHub Actions workflow that builds the package and publishes it to PyPI via OIDC trusted publishing. The version is the single source of truth in pyproject.toml and is exposed at runtime as plspm.__version__.

See CHANGELOG.md for the per-version history.

Usage

The public API mirrors upstream plspm 0.5.7. The upstream documentation is still authoritative for the inherited interface; OpenPLS-specific docs will land alongside the metric extensions.

import pandas as pd
import plspm.config as c
from plspm.plspm import Plspm
from plspm.scheme import Scheme
from plspm.mode import Mode

satisfaction = pd.read_csv("tests/data/satisfaction.csv", index_col=0)

structure = c.Structure()
structure.add_path(["IMAG"], ["EXPE", "SAT", "LOY"])
structure.add_path(["EXPE"], ["QUAL", "VAL", "SAT"])
structure.add_path(["QUAL"], ["VAL", "SAT"])
structure.add_path(["VAL"], ["SAT"])
structure.add_path(["SAT"], ["LOY"])

config = c.Config(structure.path(), scaled=False)
for lv in ["IMAG", "EXPE", "QUAL", "VAL", "SAT", "LOY"]:
    config.add_lv_with_columns_named(lv, Mode.A, satisfaction, lv.lower())

result = Plspm(satisfaction, config, Scheme.CENTROID)
print(result.inner_summary())
print(result.path_coefficients())

License

GNU General Public License v3.0, see LICENSE. Inherited from upstream plspm-python (also GPL-3.0).

Attribution

This project is a fork of googlecloudplatform/plspm-python by Jez Humble. The upstream R package plspm by Gaston Sanchez and the seminr package by Soumya Ray and Nicholas Danks remain the conceptual references for the algorithm. See ATTRIBUTION.md for details.

OpenPLS is an independent project and not affiliated with Google.