Depictio ools Catalog
The catalog turns a bioinformatics tool's outputs into ready-made dashboard
components. Point Depictio at a pipeline run and every output it recognises
becomes a render — no manual wiring. Since v1.16.0, depictio ingest builds a
whole dashboard that way for a results directory no bundled template fits: see
Composed dashboards.
Tool
The upstream software — ivar, qiime2, mosdepth, nextclade, metaphlan, pangolin… and any tool not yet integrated (freebayes, bakta, gtdb-tk, abricate). One Tool = one catalog entry.
Data Collection
A file the tool emits — recognised and reshaped into a tidy, bindable table.
Render
A dashboard component bound to that data — figure, card, table, interactive filter, advanced viz, or MultiQC section.
Browse the catalog below: filter by component type or viz kind, preview each
render on real fixture data, and copy its use: snippet.
How a render is defined¶
Every render follows the same path — from a raw file on disk to a live dashboard component:
Detect — find
Recognise which file in a run is this tool's output, by filename or path pattern.
Reshape — recipe optional
Transform the raw file into tidy, bindable columns → the Data Collection. Already-tidy files skip this.
Render — renders
Bind those columns to a dashboard component — figure, card, table, interactive filter, advanced viz, or MultiQC section.
Validated bindings
Each Data Collection's column schema lives in exactly one place (the recipe), and CI validates every render binding against the real reshaped columns — so a tool that ships green is wired correctly, with no manual review.
For the configuration reference of each component type, see the Dashboard Components guide.
Contribute a tool¶
Adding a tool is a single-folder pull request under depictio/catalog/<tool>/.
One question decides who writes it: is the tool's output already a table with a
header row?
Already a table no recipe
Use Tool Studio. Drop the file in the browser, bind its columns in Depictio's own component builder, and it opens the pull request for you.
Needs reshaping recipe
The entry needs a recipe to tidy the file first, which the Studio cannot write. Follow the contributing guide. Half the committed outputs are in this case.
Either route produces the same three co-located files:
module.yaml
The tool's identity: id, display name, the description and homepage its catalog card shows, its biotools_url (null when bio.tools has no entry), and a pointer to an upstream source (an nf-core module today; other catalogs later). The identity is declared here, copied from the upstream metadata, never derived from it at runtime.
<output>.yaml one per file
find the raw file, optionally recipe it into tidy columns, and list the renders_as it offers — figure, card, table, interactive filter, advanced viz, or MultiQC section.
<output>.tsv fixture
A small sample of that file, right beside its YAML, so depictio catalog validate previews and checks every render in CI before it ships.
Use the catalog in a dashboard Compose a dashboard from a run Read the contributing guide Open Tool Studio
The catalog always shows the latest version
The gallery above is served from a single shared copy, so it shows the current catalog no matter which documentation version you are reading. Browsing the docs for an older release will still show today's tools and renders, which may not all have existed in that release.
This is a deliberate trade-off: the gallery is ~32 MB, and publishing a copy inside every released version is what pushed the documentation site past the size limit its hosting allows. Serving per-version catalogs again is planned.