Beta

chematic for Python

chematic ships as a Python package built on the same Rust core used by the Rust crate and the WebAssembly build. It targets the workflows Python users already have — notebooks, pandas DataFrames, batch jobs over a directory of structure files — with a subset of operations that behave like their RDKit equivalents. It is not a drop-in RDKit replacement: parts of RDKit's surface area (advanced reaction handling, the full 3D toolkit, most plugins) aren't implemented.

Install

pip install chematic

Published on PyPI as version 1.0.12, classified Beta, supporting Python 3.9–3.13. A runnable introduction is in the quickstart notebook in the repository.

Jupyter and notebooks

A typical session parses a structure, inspects it, and renders it inline:

import chematic

mol = chematic.parse_smiles("CC(=O)OC1=CC=CC=C1C(=O)O")
mol.canonical_smiles()
mol.molecular_weight()
mol.depict_svg()

This is illustrative of the shape of the API, not a guaranteed-stable code sample — check the API reference for exact current method names before depending on it.

pandas and batch analysis

For screening a column of SMILES in a DataFrame or walking a directory of SDF files, chematic's descriptor and fingerprint calls are plain Python functions with no server or database required — the same pattern the browser and Rust versions use, so a batch script and a notebook cell call the same underlying logic.

What "RDKit-compatible subset" means here

The physicochemical core (molecular weight, LogP, TPSA, H-bond donor/acceptor counts, aromatic ring count) matches RDKit at 100% agreement on a 4,999-molecule ChEMBL corpus. An extended descriptor family (Kappa shape indices, BertzCT, BalabanJ, BCUT2D, VSA descriptors, MQN, SA Score) diverges substantially from RDKit at corpus scale and should not be assumed validated just because it is one of the 194 values Mol.descriptors() returns.

Mol.descriptors() returns 194 values in total — that count includes both the validated core and the extended family above, so "194 descriptors" is not the same claim as "194 validated descriptors." See Validation for the full breakdown and methodology.