About models

The purpose of the data model is to abstract away the peculiarities of the underlying file format. The same data model may be used for data created from scratch in memory, or loaded from FITS or ASDF files or some future file format.

Hierarchy of models

There are different data model classes for different kinds of data.

One model instance, many arrays

Each model instance generally has many arrays that are associated with it. For example, the ImageModel class has the following arrays associated with it:

  • data: The science data

  • dq: The data quality array

  • err: The error array

The shape of these arrays must be broadcast-compatible. If you try to assign an array to one of these members that is not broadcast-compatible with the data array, an exception is raised.

Working with models

Creating a data model from scratch

To create a new ImageModel, just call its constructor. To create a new model where all of the arrays will have default values, simply provide a shape as the first argument:

from stdatamodels.jwst.datamodels import ImageModel
with ImageModel((1024, 1024)) as im:

In this form, the memory for the arrays will not be allocated until the arrays are accessed. This is useful if, for example, you don’t need a data quality array – the memory for such an array will not be consumed:

# Print out the data array.  It is allocated here on first access
# and defaults to being filled with zeros.

If you already have data in a numpy array, you can also create a model using that array by passing it in as a data keyword argument:

data = np.empty((50, 50))
dq = np.empty((50, 50))
with ImageModel(data=data, dq=dq) as im:

Creating a data model from a file

The jwst.datamodels.open function is a convenient way to create a model from a file on disk. It may be passed any of the following:

  • a path to a FITS file

  • a path to an ASDF file

  • a astropy.io.fits.HDUList object

  • a readable file-like object

The file will be opened, and based on the nature of the data in the file, the correct data model class will be returned. For example, if the file contains 2-dimensional data, an ImageModel instance will be returned. You will generally want to instantiate a model using a with statement so that the file will be closed automatically when exiting the with block.

from stdatamodels.jwst import datamodels
with datamodels.open("myimage.fits") as im:
    assert isinstance(im, datamodels.ImageModel)

If you know the type of data stored in the file, or you want to ensure that what is being loaded is of a particular type, use the constructor of the desired concrete class. For example, if you want to ensure that the file being opened contains 2-dimensional image data:

from stdatamodels.jwst.datamodels import ImageModel
with ImageModel("myimage.fits") as im:
    # raises exception if myimage.fits is not an image file

This will raise an exception if the file contains data of the wrong shape.

Saving a data model to a file

Simply call the save method on the model instance. The format to save into will either be deduced from the filename (if provided) or the format keyword argument:



Unlike astropy.io.fits, save always clobbers the output file.

Copying a model

To create a new model based on another model, simply use its copy method. This will perform a deep-copy: that is, no changes to the original model will propagate to the new model:

new_model = old_model.copy()

It is also possible to copy all of the known metadata from one model into a new one using the update method:


History information

Models contain a list of history records, accessed through the history attribute. This is just an ordered list of strings – nothing more sophisticated.

To get to the history:

entries = model.history
for entry in entries:

To add an entry to the history, first create the entry by calling stdatamodels.util.create_history_entry and appending the entry to the model history:

import stdatamodels
entry = stdatamodels.util.create_history_entry("Processed through the frobulator step")

These history entries are stored in HISTORY keywords when saving to FITS format. As an option, history entries can contain a dictionary with a description of the software used. The dictionary must have the following keys:

name: The name of the software author: The author or institution that produced the software homepage: A URI to the homepage of the software version: The version of the software

The calling sequence to create a history entry with the software description is:

entry =  stdatamodels.util.create_history_entry(description, software=software_dict)

where the second argument is the dictionary with the keywords mentioned.

Looking at the contents of a model

Use model.info() to look at the contents of a data model. It renders the underlying ASDF tree starting at the root or a specified node. The number of displayed rows is controlled by the max_row argument:

root.tree (AsdfObject)
├─asdf_library (Software)
│ ├─author (str): Space Telescope Science Institute
│ ├─homepage (str): http://github.com/spacetelescope/asdf
│ ├─name (str): asdf
│ └─version (str): 2.5.2a1.dev12+g12aa460
├─history (dict)
│ └─extensions (list) ...
├─data (ndarray): shape=(2048, 2048), dtype=float32
├─dq (ndarray): shape=(2048, 2048), dtype=uint32
├─err (ndarray): shape=(2048, 2048), dtype=float32
├─meta (dict)
│ ├─aperture (dict) ...
│ ├─bunit_data (str): DN/s
│ ├─bunit_err (str): DN/s
│ ├─cal_step (dict) ...
│ ├─calibration_software_revision (str): 3bfd782b
│ ├─calibration_software_version (str): 0.14.3a1.dev133+g3bfd782b.d20200216
│ ├─coordinates (dict) ...
│ └─28 not shown
├─var_poisson (ndarray): shape=(2048, 2048), dtype=float32
├─var_rnoise (ndarray): shape=(2048, 2048), dtype=float32
└─extra_fits (dict) ...
Some nodes not shown.

Searching a model

model.search() can be used to search the ASDF tree by key or value:


root.tree (AsdfObject)
└─meta (dict)
├─instrument (dict)
│ └─filter (str): F170LP
└─ref_file (dict)
  └─filteroffset (dict)

Converting from astropy.io.fits

This section describes how to port code that uses astropy.io.fits to use jwst.datamodels.

Opening a file

Instead of:



from stdatamodels.jwst.datamodels import ImageModel
with ImageModel("myfile.fits") as model:

In place of ImageModel, use the type of data one expects to find in the file. For example, if spectrographic data is expected, use SpecModel. If it doesn’t matter (perhaps the application is only sorting FITS files into categories) use the base class JwstDataModel.

An alternative is to use:

from stdatamodels.jwst import datamodels
with datamodels.open("myfile.fits") as model:

The datamodels.open() method checks if the DATAMODL FITS keyword has been set, which records the DataModel that was used to create the file. If the keyword is not set, then datamodels.open() does its best to guess the best DataModel to use.

Accessing data

Data should be accessed through one of the pre-defined data members on the model (data, dq, err). There is no longer a need to hunt through the HDU list to find the data.

Instead of:




Accessing keywords

The data model hides direct access to FITS header keywords. Instead, use the Metadata tree.

There is a convenience method, find_fits_keyword to find where a FITS keyword is used in the metadata tree:

>>> from stdatamodels.jwst.datamodels import JwstDataModel
>>> # First, create a model of the desired type
>>> model = JwstDataModel()
>>> model.find_fits_keyword('DATE-OBS')

This information shows that instead of:




Extra FITS keywords

When loading arbitrary FITS files, there may be keywords that are not listed in the schema for that data model. These “extra” FITS keywords are put under the model in the _extra_fits namespace.

Under the _extra_fits namespace is a section for each header data unit, and under those are the extra FITS keywords. For example, if the FITS file contains a keyword FOO in the primary header, its value can be obtained using:


This feature is useful to retain any extra keywords from input files to output products.

To get a list of everything in _extra_fits:


returns a dictionary of of the instance at the model._extra_fits node.

_instance can be used at any node in the tree to return a dictionary of rest of the tree structure at that node.

Environment Variables

There are a number of environment variables that affect how models are read.


Used by ~jwst.datamodels.JwstDataModel when instantiating a model from a file. If True, values that do not validate the schema will still be added to the metadata. If False, they will be set to None. Default is False.


Used by ~jwst.datamodels.JwstDataModel when instantiating a model from a file. If True, schema validation errors will generate an exception. If False, they will generate a warning. Default is False.


DEPRECATED: In the future the FITS header will always be used. Used by ~jwst.datamodels.JwstDataModel when instantiating a model from a FITS file. When False, models opened from FITS files will proceed and load the FITS header values into the model. When True and the FITS file has an ASDF extension, the loading/validation of the FITS header will be skipped, loading the model only from the ASDF extension. If not defined, the instantiation routines will determine whether the loading/validation of the FITS header can be skipped or not.


Implemented by the utility function jwst.datamodels.util.check_memory_allocation and used by ~jwst.outlier_detection.OutlierDetectionStep and ~jwst.resample.ResampleStep. When defined, determines how much of currently available memory should be used to instantiated an output resampled image. If not defined, no check is made.

Examples would be: 1.0 would allow all available memory to be used. 0.5 would allow only half the available memory to be used.

For flag or boolean variables, any value in ('true', 't', 'yes', 'y') or a non-zero number, will evaluate as True. Any value in ('false', 'f', 'no', 'n', '0') will evaluate as False. The values are case-insensitive.

All of the environment variables have equivalent function arguments in the API for the relevant code. The environment variables are used only if explicit values had not been used in a script. In other words, values in code override environment variables.