Importing Data#

Note

jdaviz is undergoing constant development. The latest enhancement is deconfigged mode where a user can load data into jdaviz - as shown below - and specify in what format the data should be parsed. Viewers are created based on the data loaded and user input. If you uncover any issues or bugs, you can open a GitHub issue.

Jdaviz provides several ways to import data into the application. This guide will walk through the available methods and their use cases.

Using load()#

The recommended way to load data is to use the _load() method on the configuration helper, whether that’s the deconfigged mode as shown below, or a specific configuration such as Specviz, Cubeviz, etc. This provides a simple interface for loading data:

import jdaviz
jdaviz.show()
jdaviz.load('myspectrum.fits', format='1D Spectrum')

# Alternatively
from jdaviz import Specviz
specviz = Specviz()
specviz.show()
specviz.load('myspectrum.fits')

Using Loaders API (Interactive)#

For more control over the loading process, you can use the Loaders API. This provides an interactive way to:

  • Select specific loaders

  • Configure import options

  • Inspect available options from dropdowns, etc.

  • View API hints for available options

  • Preview data before importing

The loaders are accessed through the loaders property on the helper:

# Access loaders
loaders = jdaviz.loaders

# Get list of available loaders
print(loaders)

# Use a specific loader
ldr = loaders['object']  # Or file, file drop, or url

# Configure loading options
ldr.object = my_data  # or ldr.filename, ldr.url
ldr.format = '1D Spectrum'

# Import the data
ldr.load()

The loaders interface provides access to parameters and options that control how the data is loaded and processed.

Available Loaders#

The main loaders available include:

  • 'file' - Load data from local files

  • 'url' - Load data from remote URLs/URIs

  • 'object' - Load Python objects directly (Spectrum1D, SpectrumList, NDData, etc)

Each loader has its own set of parameters that can be discovered through Using API Hints or by accessing the loader object’s attributes directly. For example, to see available formats for the object loader:

ldr.format.choices

Available formats#

  • '1D Spectrum' - For individual 1D Spectrum objects or files containing a single spectrum

  • '1D Spectrum List' - For SpectrumList objects or files containing multiple spectra

  • '1D Spectrum Concatenated' - For combining multiple spectra into a single spectrum

  • 'Image' - For 2D image data

  • '2D Spectrum' - For 2D spectral data

  • '3D Spectrum' - For 3D spectral cube data

  • 'Footprint' - For footprint overlays on images

  • 'Subset' - For loading subsets/regions

  • 'Trace' - For loading spectral traces

Using API Hints#

Jdaviz provides an API hints feature that helps you discover available attributes and parameters for loaders. You can enable API hints:

jdaviz.toggle_api_hints()

Or in the UI, click the API Hints button in the top right.

When API hints are enabled, you’ll see Python code snippets showing how to access and set various loader attributes. For example:

# When selecting a file loader, you might see:
ldr = jdaviz.loaders['file']
ldr.format = '1D Spectrum'  # Set the format
ldr.filename = 'myfile.fits'  # Set the filename

The hints update as you interact with the UI, showing you the exact Python code needed to reproduce your actions programmatically.

Differences from load_data()#

The load() method replaces the older load_data() method. Key differences:

  1. Loader Auto-detection - load() attempts to auto-detect the appropriate loader based on the input

  2. Format Selection - Provides more control over data format via the format parameter

  3. Target Specification - Can direct data to specific viewers using target

  4. Consistent Interface - Works consistently across different data types

Differences from the Loaders API#

For most use cases, load() provides a simpler interface. Use the Loaders API when you need:

  • More control over loading options

  • Interactive configuration

  • Data preview before loading

  • Access to specialized loader features

While load_data() is still supported for backwards compatibility, load() is recommended for new code.