Validate an RDAT file

The legacy RMDB site offered an in-browser RDAT validator. The static site delegates this to the open-source rdat_kit Python package (≥ v1.7.0).

Install

pip install 'rdat_kit>=1.7.0'

Validate a file

rdat_kit validate path/to/entry.rdat

Output is one line per file:

path/to/entry.rdat: OK   (1 construct(s), 272 data row(s), RDAT_VERSION=0.34)

Exit code:

Code	Meaning
`0`	File parsed cleanly and `RDATFile.validate()` returned no warnings
`1`	Parse failure — malformed file
`2`	Parsed OK but at least one schema warning (see stdout)

You can pass multiple files at once:

rdat_kit validate *.rdat

What is checked

RDAT_VERSION is one of the supported revisions (0.32 / 0.33 / 0.34).
SEQUENCE is a valid RNA string (ACGUTacguTNn); length matches SEQPOS extent.
STRUCTURE has balanced brackets and matches SEQUENCE length.
Every REACTIVITY:<i> has a matching ANNOTATION_DATA:<i>.
ANNOTATION keys are in the controlled vocabulary (warning only — unknown keys are allowed but flagged).
Numeric fields parse as floats.

See the RDAT format spec for full definitions.

Companion subcommands

rdat_kit exposes three subcommands in total:

rdat_kit validate   <file.rdat>            # this page
rdat_kit to_md      <file.rdat>            # emit Jekyll front-matter stub
rdat_kit thumbnail  <file.rdat> --out DIR  # render reactivity heatmap PNG

to_md auto-populates sequence, structure, offset, construct_count, data_points, annotation, comments, and the citation skeleton from the RDAT contents — saving the contributor from writing the YAML by hand.

thumbnail requires matplotlib (an optional dep): pip install 'rdat_kit[thumbnail]' or 'pip install matplotlib. It truncates huge libraries to the first 1000 rows by default (--max-rows to override) and renders a heatmap to drop into assets/thumbnails/.

See the contribute page for the full submission workflow.