Installation

v0.1.7

Get started with SynaptIX in minutes

Quick Install

Install SynaptIX from PyPI using pip — the core (with scikit-learn) has no heavy requirements:

bash

python

pip install synaptix

Verify the installation:

python

import synaptix as sx
print(sx.__version__)
print(sx.list_datasets())

Requirements

Core dependencies ship with every install; the rest are extras:

Package	Minimum Version
Python	>= 3.9
numpy	>= 1.21.0
pandas	>= 1.3.0
scipy	>= 1.7.0
matplotlib	>= 3.4.0
scikit-learn	>= 1.2.0
requests	>= 2.25.0
tensorflow	>= 2.12.0 (optional extra: dl)
pymc	>= 5.0.0 (optional extra: bayes)
viewx	>= 0.3.0 (optional extra: viewx)
statslibx	>= 0.4.0 (optional extra: stats)

Optional Extras

TensorFlow for neural networks and DQN.

bayes

PyMC for full Bayesian inference models.

viewx

Export results to HTML dashboards, PDF and presentations.

stats

StatsLibX for statistical model diagnostics.

nlp

NLTK for the legacy NLP API.

all

Everything above, plus polars and plotly.

bash

python

pip install 'synaptix[dl]'

bash

python

pip install 'synaptix[viewx,stats]'

bash

python

pip install 'synaptix[all]'

Development Install

Clone the repository and install in editable mode:

bash

python

git clone https://github.com/GhostAnalyst30/SynaptIX.git
cd SynaptIX
pip install -e .

Quick Start

Train and evaluate your first model in a few lines:

python

import synaptix as sx

df = sx.load_dataset("iris")
X, y = df.drop(columns="species"), df["species"]
X_train, X_test, y_train, y_test = sx.preprocessing.train_test_split(
    X, y, test_size=0.2, random_state=42
)

model = sx.supervised.RandomForestClassifier(n_estimators=100)
model.fit(X_train, y_train)
model.evaluate(X_test, y_test, plot=True)

First Steps

Load Data

Use sx.load_dataset() or any pandas/polars DataFrame.

Preprocess

DataCleaner, Scaler, Encoder and Imputer keep column names.

Train & Evaluate

fit() then evaluate() — or let AutoPipeline do everything.

Improve & Export

Tune with GridSearch, diagnose with diagnostics() (statslibx) and export with to_html() (viewx).

Troubleshooting

Command not found: pip

Ensure Python and pip are installed and added to your PATH. Try using python -m pip install synaptix instead.

ImportError: No module named 'synaptix'

The package may not be installed correctly. Re-run pip install synaptix and verify with pip list.

ImportError about tensorflow when importing synaptix.neural

Neural networks (MLP, CNN, LSTMNet) require Keras/TensorFlow. Install the extra: pip install 'synaptix[dl]'. Every other module works without it.

Model diagnostics() raises a statslibx error

diagnostics() uses StatsLibX for inferential tests on residuals. Install the extra: pip install 'synaptix[stats]'.

to_html()/to_report() raise 'viewx not installed'

Result export to dashboards/PDF/slides needs ViewX. Install the extra: pip install 'synaptix[viewx]'.