CLI preview

SXS runs in a terminal. These previews render output recorded from the actual CLI, with a styled frame for readability. They are not screenshots of a graphical application, and no new research run or discovery is implied.

Select a preview image to open it at full size. The expandable transcripts below each image remain readable and copyable on smaller screens.

Explore the commands

python -m src.cli --help

The four commands cover baseline recovery, scaled model qualification, candidate screening, and independent validation. Add --help after a command to inspect its options, for example python -m src.cli validate --help.

Copyable help output

SXS | SCIX Exoplanet Search

Usage: sxs <command> [options]
       python -m src.cli <command> [options]

Commands:
  baseline  Run or inspect the baseline recovery workflow
  scaleup   Build the scaled benchmark and qualify the production model
  search    Screen the deterministic unknown-target sample
  validate  Run independent statistical and astrophysical validation
  demo      Run an offline synthetic transit example
  analyze   Analyze a CSV/FITS light curve or a Kepler KIC
  report    Rebuild a self-contained analysis HTML report
  inject    Measure conditional transit injection recovery
  evaluate  Run nested target-grouped RF evaluation
  doctor    Check the installation and optional service connectivity
  config-check  Validate workflow YAML without running scientific stages
  init      Create an isolated workspace with validated configurations
  status    Inspect workspace configuration and recorded workflow state
  support-bundle  Create a privacy-conscious diagnostic ZIP
  verify    Verify checkpoint identity and recorded file hashes
  citation  Print the preferred archived-software citation

Pass --help after a command to see workflow-specific options.
Legacy workflows accept --workspace DIR for isolated configs and outputs.

Preview a run without executing it

python -m src.cli baseline --config configs/base.yaml --dry-run

dry_run means the pipeline was inspected, not executed. Each would_run entry is a planned stage, not a completed stage. This command does not download mission products, train models, or overwrite research artifacts.

The image is a compact JSON excerpt: it retains the exact status, config_path, and all six stages values. The expandable transcript includes the remaining JSON fields, except the two run timestamps, which are omitted to keep the example reproducible. Dependency warnings on stderr are not shown in these stdout previews and may differ between environments.

Copyable dry-run JSON (timestamps omitted)

{
  "schema_version": 2,
  "status": "dry_run",
  "config_path": "configs/base.yaml",
  "options": {
    "from_stage": 0,
    "to_stage": 5,
    "resume": false,
    "refresh_catalog": false,
    "dry_run": true
  },
  "stages": [
    {
      "stage": 0,
      "name": "environment",
      "status": "would_run"
    },
    {
      "stage": 1,
      "name": "acquisition",
      "status": "would_run"
    },
    {
      "stage": 2,
      "name": "preprocessing",
      "status": "would_run"
    },
    {
      "stage": 3,
      "name": "bls_detection",
      "status": "would_run"
    },
    {
      "stage": 4,
      "name": "machine_learning",
      "status": "would_run"
    },
    {
      "stage": 5,
      "name": "catalog_validation",
      "status": "would_run"
    }
  ]
}

Try it yourself

Follow Installation or the Container guide, then work through Quickstart. With the container, replace python -m src.cli with:

docker run --rm ghcr.io/science-experimental-technologies/exoplanet-search:main

Append --help or baseline --config configs/base.yaml --dry-run as above. Public pulls require the package's visibility to be Public; see the container guide if registry authorization is denied.

For archived scientific outputs rather than CLI operation, see Results and Interpreting results.

Refresh the previews

From a checkout with SXS dependencies installed:

python scripts/build_cli_previews.py
python scripts/build_cli_previews.py --check

The generator invokes only help and dry-run commands. It creates local SVGs and text transcripts under docs/assets/cli/; no external image service is used. CI checks that the committed previews still match the current CLI output.