Skip to content

ALMA analysis workflow

Use this workflow to move ALMA data into a CANFAR Session, reduce it with CASA, inspect the products, and keep the results in persistent storage. You can use the browser for Session management and file transfers, with optional command-line steps for repeatable work.

Prerequisites

  • A CADC account and access to the CANFAR Science Platform. Contact support@canfar.net if your account or project access is not ready.
  • A persistent destination: use /arc/home/[username] for personal files or /arc/projects/[project] for shared work. Treat /scratch as temporary; it is suitable for fast intermediate files, not final results.
  • A Session Kind that matches the task: use Desktop for CASA and other GUI applications, Notebook for Python-based analysis, and CARTA for inspecting images and spectral cubes.
  • Optional command-line access: follow Install and set up before using the CLI examples below. Run canfar login cadc to authenticate and canfar auth show to check your selection.

See Getting Started for account and portal setup.

Workflow

1. Choose a Session and storage location

From the Science Portal, create a Desktop, Notebook, or CARTA Session and choose an image that supports the application you need. From a terminal, list the available images before creating a Session:

canfar image ls --kind desktop
canfar create desktop IMAGE_NAME --name alma-reduction
canfar ps --kind desktop
canfar open [session-id]

Replace bracketed values with the Session ID, username, and project name used by your account. Replace IMAGE_NAME with an image returned by canfar image ls --kind desktop that includes the CASA version required by the reduction. The image listing is the source of truth; do not assume that a CASA tag remains unchanged.

2. Stage the archive data

Download the requested data from the ALMA Science Archive to your local machine, then upload them in your browser. For a scripted transfer, create a destination and verify the copy with the CANFAR data commands:

canfar data mkdir -p arc:/home/[username]/alma/raw
canfar data cp local:/path/to/alma-data.tar.gz \
    arc:/home/[username]/alma/raw/alma-data.tar.gz
canfar data ls -lh arc:/home/[username]/alma/raw

The local: source is the machine where the command runs. To copy data that is already in Vault, use a vault: source instead:

canfar data cp vault:/ALMA/test-data/cutouts/test-4d-cube.fits \
    arc:/home/[username]/alma/raw/test-4d-cube.fits

Inside a Session, the same /arc paths are available from the file browser and terminal. For a large reduction, copy working data to /scratch and keep the source and final products under /arc.

3. Start CASA in a Desktop Session

In a Desktop Session, open Applications > Astro Software, choose the CASA version required by the project, and open its terminal. Start CASA in the mode required by the supplied reduction script:

casa
casa --pipeline

Follow the README delivered with your archive data and use its required CASA version. For deliveries that include these scripts:

  1. Run scriptForPI.py using the invocation in the archive README to calibrate the observations.
  2. Check that calibration completed and locate the calibrated measurement sets, usually in the delivery's calibrated directory. A measurement set contains radio observations ready for imaging; it is not an image.
  3. Follow the supplied scriptForImaging.py instructions to produce CASA images, or use the version-appropriate imaging procedure in the README.
  4. Check the imaging output before proceeding to visual inspection.

Keep the scripts, measurement sets, and intermediate products under a persistent project directory or copy them back to /arc when the reduction finishes. The CASA containers guide lists version-specific package notes and known issues.

4. Inspect and analyse the products

Open the image products from CASA imaging in a CARTA Session for cube navigation, spectra, regions, or moment maps. In CARTA, use File > Open Image and choose a CASA image or an exported FITS image. CARTA's file browser accepts image formats; use CASA to work with measurement sets. Use a Notebook Session for Python analysis. Both Session Kinds can read files saved under /arc/home/[username] and /arc/projects/[project].

For a local or scripted check, list the result before opening it:

canfar data ls -lh arc:/home/[username]/alma/results
canfar data info arc:/home/[username]/alma/results/result.fits

5. Save results and clean up

Copy final products out of /scratch before stopping the Session. From inside the Session, use ordinary filesystem commands:

mkdir -p /arc/home/[username]/alma/results
cp -a /scratch/alma-reduction/results/. \
    /arc/home/[username]/alma/results/

From the machine where the files are local, use the data command instead:

canfar data cp local:/path/to/result.fits \
    arc:/home/[username]/alma/results/result.fits
canfar data ls -lh arc:/home/[username]/alma/results

When the results are safely stored, inspect and remove the Session if it is no longer needed:

canfar info [session-id]
canfar delete [session-id] --force

Expected results

At the end of the workflow:

  1. You can sign in to the Science Portal and access your project storage. If you use the CLI, canfar auth show and canfar server ls confirm your setup.
  2. The selected Session starts and its application opens from the Science Portal.
  3. Raw data and final products are visible under the chosen /arc path from the Desktop, Notebook, or CARTA Session.
  4. CASA scripts complete using a compatible image, and the reduction products are stored outside /scratch.
  5. Deleting the Session does not remove the files saved under /arc.

Troubleshooting

Symptom Check or fix
Login fails or credentials are stale Run canfar --log-level debug login cadc --force, then canfar auth show.
No suitable image appears Run canfar image ls --kind desktop (or --kind notebook/--kind carta) and choose an image that includes the required software.
Session remains pending Run canfar ps --all, canfar events [session-id], and canfar stats; reduce requested resources if the Science Platform Server has no capacity.
A data copy fails Confirm the local:, arc:, or vault: source and destination with canfar data ls -lh IDENTIFIER:/path; check the credentials for the storage identifier's owning server. Default arc and vault use CADC credentials independently of the active compute server.
Files disappear after the Session ends Move them from /scratch to /arc/home/[username] or /arc/projects/[project] before deleting the Session.
A shared project path is denied Ask the project administrator to add your CADC account to the project group; see Permissions.
CASA is missing or incompatible Select a Desktop image that lists the required CASA version and match the version to the archive scripts. If CASA 6.5.0 to 6.5.2 opens with display errors, exit CASA and start it again.
A transfer reports expired credentials Renew the credentials used by the storage identifier's owning server, then retry. See support if the record or certificate cannot be refreshed.
The browser cannot connect to a new Session Wait for the Session to reach Running, then retry canfar open [session-id]; inspect canfar info [session-id] and canfar logs [session-id] if it still fails.