tmap-silicon 1.0.19

A Python package for visualizing large, high-dimensional data sets.

tmap (Python 3.9+ & Apple Silicon Compatible)

This is a fork of tmap that is compatible with Python 3.9-3.12 and Apple Silicon (arm64).

tmap is a very fast visualization library for large, high-dimensional data sets. The graph layouts are based on the OGDF library.

Installation

Option 1: Install from GitHub (Recommended)

pip install git+https://github.com/afloresep/silicon-tmap-compatible.git

Option 2: Install from pre-built wheels

Check the Releases page for pre-built wheels for your platform.

Option 3: Build from source

For Apple Silicon Macs or if you need to build from source:

# Clone the repository
git clone https://github.com/afloresep/silicon-tmap-compatible.git
cd silicon-tmap-compatible

# Create conda environment with build dependencies
conda create -n tmap python=3.12 -y
conda activate tmap
conda install -c conda-forge cmake ninja llvm-openmp numpy scipy matplotlib annoy pybind11 -y

# Build OGDF
cd ogdf-conda/src
mkdir -p build installed
cd build
cmake .. \
  -DCMAKE_INSTALL_PREFIX="$PWD/../installed" \
  -DCMAKE_BUILD_TYPE=Release \
  -DCMAKE_OSX_ARCHITECTURES=arm64  # Only for Apple Silicon
cmake --build . --parallel
cmake --install .
cd ../../..

# Install tmap
export LIBOGDF_INSTALL_PATH="$PWD/ogdf-conda/src/installed"
pip install . --no-build-isolation

Visualization

We recommend using faerun to plot the data laid out by tmap:

pip install faerun

Quick Example

import tmap as tm
import numpy as np

# Generate random data
data = np.random.rand(1000, 100)

# Create MinHash fingerprints
mh = tm.Minhash(128)
fps = [tm.VectorFloat(row) for row in data]
lf = tm.LSHForest(128)
lf.batch_add(mh.batch_from_weight_array(fps))
lf.index()

# Layout
cfg = tm.LayoutConfiguration()
cfg.k = 20
x, y, s, t, _ = tm.layout_from_lsh_forest(lf, cfg)

# Plot with faerun or matplotlib
import matplotlib.pyplot as plt
plt.scatter(x, y, s=1)
plt.show()

Tutorial and Documentation

See the original documentation at http://tmap.gdb.tools

Examples

Name	Description
MNIST	Visualization of the MNIST data set
Drugbank	All drugs registered in Drugbank
Project Gutenberg	Linguistic relationships between books

Compatibility

Python	Linux	macOS Intel	macOS Apple Silicon	Windows
3.9	✅	✅	✅	✅
3.10	✅	✅	✅	✅
3.11	✅	✅	✅	✅
3.12	✅	✅	✅	✅

Changes from Original

• Fixed compatibility with Python 3.9-3.12
• Fixed Apple Silicon (arm64) builds
• Updated deprecated distutils imports
• Added proper pyproject.toml build configuration

License

BSD 3-Clause License - see LICENSE.md

Credits

Original tmap by Daniel Probst and the Reymond Group.