How RxCubed organises work

How RxCubed organises work

The levels

Every saved or generated result is organised as Category → Tool → Run → Evidence.

• A Category is a broad area: Literature, File Collection, Model, or Documents.
• A Tool is a connector or computational operation.
• A Run is one search, upload, or tool execution, with its settings and timestamp.
• Evidence is what that run produced: one row per file.

One run often produces several evidence rows — saving a structure as both mmCIF and legacy PDB, or a tool that writes both a TSV and a JSON. Each stays its own row so a later tool can select the exact file format it needs, while related rows stay visibly together under the run that made them.

Evidence

Evidence includes saved papers, database records, files, computational outputs, and reports. Each row records its source, type, creator, metadata, and (when present) a checksum of the file.

States

Search results may be candidates until you select and save them. Saved evidence can be marked selected, generated, rejected, or superseded. These states describe your judgement; they do not delete the underlying record.

Provenance

A derived result points to its inputs. Browse shows the saved evidence; Graph shows the selected run's upstream inputs and downstream results. Clicking a graph node opens its details without changing the focused run.

Notes and chat

Chat is private to you. Notes are shared project-level working documents. Evidence and reports are persistent, checksummed records with provenance.