Internal format of molecular structure in the InChI software

[stuchalk]

In a recent IUPAC meeting there was a brief (but important) discussion of how best to standardize unique identifier generation across community tools and services. There was a comment by Ian Bruno that InChI has an internal representation of chemical structure that could be a (canonical?) format that might help with this. I would interested to know more about the format. Stuart Chalk

[ijbruno]

This has also come up as a topic of discussion at the July 2025 Aachen InChI Technical meeting and there is interest in this idea from others as well. There is also a lot of discussion about the practicality of exposing this! One interesting suggestion was to have a Python API that provides access to the internal representation that can be output in whatever formats people prefer.

[stuchalk]

A Python package to allow use of the internal representation format API with data output as JSON would be awesome. There could also be a JSON schema provided to define the elements in the JSON, adding documentation and datatypes which would make nature of the data clear. It could then be reworked as a JSON-LD version ... so many possibilities! Sign me up :)