DrawMolToMPL: RDKit mol drawing with matplotlib

A quickfix for rdkit.Chem.Draw.MolToMPL to save drawings as vector graphics.

What is this for?

DrawMolToMPL generates a drawing of rdkit.Chem.rdchem.Mol objects by matplotlib line graphics, and thus it can be saved as a vector image.

I had a problem when I tried to output RDKit molecular graphics into a PDF or EPS file. If I convert it into a PIL image by rdkit.Chem.Draw.MolToImage for matplotlib subplots, then the output quality was in a low resolution because it's a pixel image. But, apparently SVG-based drawings (like rdkit.Chem.Draw.rdMolDraw2D.MolDraw2DSVG) cannot simply be saved to a PDF file. Furthermore, it cannot work with matplotlib subplots, and it's frustrating when we want to draw many pairs or sets of molecules in a single panel.

DrawMolToMPL fixes this issue to work with matplotlib subplots. The following example successfully generates a vector-graphcis output like out.pdf.

matplotlib.pyplot.subplots provides flexible and powerful layouts for multiple graphics, and we can enjoy its significant benefit by using DrawMolToMPL. Even for a simple grid layout, it can be useful. A RDKit's practice would be something like this

and the output is saved as PNG (a raster image) or SVG (a vector image). To get the output in PDF, it further requires a converter such as rsvg-convert and cairosvg (or Illustrator, Inkscape, etc), which is annoying. Check out out_examples.

$ cairosvg rdkit_out.svg -o rdkit_cairosvg.pdf
$ rsvg-convert -f pdf -o rdkit_librsvg.pdf rdkit_out.svg

How to use this?

Just download mpl_rdkit_utils.py by

$ curl -O https://raw.githubusercontent.com/itakigawa/mpl_rdkit_utils/main/mpl_rdkit_utils.py

and import it like

from mpl_rdkit_utils import DrawMolToMPL

For complete examples, see example.ipynb.

To work with matplotlib subplots

Internally DrawMolToMPL draw a molecule in the [0,1]x[0,1] area, and returns (xlim, ylim) representing its bounding box. So you can work using this information as you like.