The tools that get data in and ready.

Everything an agent needs before it can train: bringing a file into TuiML, looking at it, reshaping it, or synthesising one when there is no real data to hand.

Tools

  • tuiml_upload_data: Register a dataset from a file path or from inline
text, giving it a name later tools can refer to.
  • tuiml_read_data: Show rows (head, tail, sample, or specific indices)
so the agent can see what it is working with.
  • tuiml_profile_data: Shape, dtypes, missing values, class balance and
summary statistics.
  • tuiml_generate_data: Synthesise a dataset from a built-in generator
(Blobs, Agrawal, Friedman, ...).
  • tuiml_preprocess: Apply preprocessing steps and save the result.
  • tuiml_select_features: Reduce to the most informative columns.

Notes

Anywhere these take a data argument, it accepts a built-in dataset name ("iris"), a path to a file, or the name of a previous upload.

Modules