Quick Start

This guide walks through the shortest end-to-end Adagio workflow:

add actions to the canvas
connect them into a valid graph
expose the right run-time inputs and parameters
run the pipeline
export the pipeline for CLI use

1. Open the pipeline editor

2. Create a new pipeline

The editor has three main areas:

the action tray on the left
the canvas in the middle
the inspector on the right

3. Add actions

Each action is one plugin method or pipeline. To add one:

Browse or search in the action tray.
Add an action such as demux, dada2, or another plugin action relevant to your workflow.
Repeat until the core steps of your analysis are on the canvas.

4. Connect nodes

Connect an output port to a downstream input port.

Adagio only accepts type-compatible connections. If a port does not highlight as a valid target, the semantic types do not match. See Connections and Semantic Types.

5. Promote the run-time interface

For each required input that should come from the runner rather than another node:

Select the unconnected input port.
Click Promote Input.

For each parameter that should stay configurable at run time:

Select the node.
In the parameter controls, enable Promote.
Optionally assign a default later from the pipeline summary.

Anything you do not promote stays fixed in the pipeline.

6. Save and verify

Use the To-do tab to clear validation issues.

Adagio blocks pipeline download until the graph has no remaining required-input, required-parameter, or naming problems.

7. Run from the UI

Click Run in the toolbar. Adagio will prompt you to provide values for any required inputs and parameters that have no defaults. Once submitted, you can watch each step complete in the run view.

The run form has two parts:

Run arguments for files and promoted parameters
Run configuration for the analysis name, compute settings, and account

8. Export the pipeline

From the editor you can export:

the pipeline itself as a .adg file
an arguments JSON template for CLI runs

The arguments export captures the current parameter values and input file names. Before using it with the CLI, replace those file names with real local paths.

9. Run the same pipeline with the CLI

Typical flow:

adagio pipeline show my-pipeline.adg

adagio run \
  --pipeline my-pipeline.adg \
  --arguments my-arguments.json \
  --cache-dir ./adagio-cache