Page History
| Table of Contents |
|---|
The Pandeia engine of the Exposure Time Calculator is released to the community to support users who wish to script their calculations, run more extensive parameter space studies, and have more direct control of their scenes. We also recognize that the community has developed more extensive wrappers and public tools that depend on the Pandeia engine.
This page is intended to facilitate communication with developers in the community with Pandeia engine dependencies.
The latest release of the Pandeia engine is 1.7.
- Get the latest engine software software, installable with pip (https://pypi.org/project/pandeia.engine/)
- See the installation instructions
- Get the Synphot data files that support certain target spectra manipulations (https://archive.stsci.edu/hlsp/reference-atlases)
- Get the required mission-specific items:
- View the Input API documentation
- View the Output API documentation
Next Planned Release
The next release of the Pandeia Engine will be after JWST commissioning.
The next release will require Python 3.8+, as a result of Astropy 5.0
| Warning | ||
|---|---|---|
| ||
This alpha release contains numerous changes to APIs in the engine internals and the reference data structures; there are no changes to the regular external input API apart from a change to the meaning of the aperture_size and sky_annulus parameters. They now refer to the full-height of the aperture and a full-span distance; see JETC-2678 for more details. * If you are using the Pandeia Engine by passing it an input dictionary via perform_calculation.perform_calculation(), you MAY not be affected by the changes. * If your code uses the Pandeia Engine as a library and uses (in particular, but not limited to) the Scene, Source, ConvolvedSceneCube, PSFLibrary, or AdvancedPSF classes, you may need to make significant changes to your code. THIS CODE AND DATA PACKAGE DOES NOT CONTAIN ANY POST-COMMISSIONING DATA. This code is being distributed to the community for development purposes only and is not approved for scientific use. Developers should not make releases based on this alpha release. This code is not feature-complete, and there will be other major and minor changes prior to the actual 2.0 release. Engine: v2.0alpha (or |
In the next release: (Bolded items are new and NOT in the v2.0 alpha)
| Release Notes | Description |
| JETC-1957 | The engine no longer accepts non-integer values for groups, integrations, and exposures |
| JETC-2093 | Users can now specify "pixel" as an option for extraction units. |
| JETC-2354 | New normalization bandpasses: johnson_u, johnson_b, johnson_r, cousins_r, galex_fuv, galex_nuv. |
| JETC-2358 | The next release will not support versions of Python older than 3.8 |
| JETC-2594 | Changes have been made to background handling for calculations with no background; their results should be more consistent when compared to calculations with background. |
| JETC-2593 | The Source and Scene class initializers, and calc_utils.build_default_source() function, now require a "telescope" parameter, which must be the "roman" string. calc_utils.build_empty_source() does NOT accept "telescope", and calc_utils.build_default_calculation() is unchanged. |
| JETC-1857 | The Pandeia Engine now warns if the instrument setup has a filter leak, in the form of substantial flux from a given source coming from wavelengths outside the defined instrument bandpass. Such flux will not be accounted for in the Engine's results. |
| JETC-2678 | SpecApPhot aperture size now defined by the full-height instead of the half-height. Sky Annulus size (which still defines two rectangular regions above and below the aperture) now indicates the distance in arcseconds between the inner edges of the rectangular regions, and between the outer edges of the rectangular regions. |
| JETC-2812 | Spectra with sharp features were previously being incorrectly resampled, resulting in an erroneous reduction in flux. A persistent warning about source spectrum extrapolation was removed. |
| JETC-1909 | The ConvolvedSceneCube creation and AdvancedPSF class have been refactored to remove redundant operations. Though each AdvancedPSF instance should be identical, it (and its construction) are now very different. The Pandeia Engine should be ~10% faster as a result. |
| JETC-2812 | Spectra with sharp features were previously being incorrectly resampled, resulting in an erroneous reduction in flux. A persistent warning about source spectrum extrapolation was removed. |
| JETC-1280 | Synphot source errors have been shortened, and the error for normalizing a zero-flux source has been made more descriptive. |
| JETC-2781 | The PSFLibrary class has been refactored to load only the PSFs required for the calculation. |
What support is available?
Questions about the Pandeia engine may be directed to email help@stsci.edu with Roman and/or WFIRST in the subject line or body. However, due to the complexity of the engine, support will be limited and response times may be longer than for other tools.
We welcome comments and feature requests, and these will be considered along with other ETC work.
What is the Pandeia Engine?
The Pandeia engine uses a pixel-based 3-dimensional approach to perform calculations on small (typically a few arcseconds) 2-dimensional user-created astronomical scenes. It models both the spatial and the wavelength dimensions, using realistic point spread functions for each instrument mode. It natively handles correlated read noise, inter-pixel capacitance, and saturation. Since the signal and noise are modeled for individual detector pixels, the ETC is able to replicate many of the steps that observers will perform when calibrating and reducing their Roman data. This simplifies interpretation of the extracted signal-to-noise ratio (SNR) calculated by the ETC.
| Info |
|---|
| While the Pandeia engine includes many effects not typically included in other ETCs, it is not an observation simulator. It does not simulate the full detector, nor does it include 2-dimensional effects such as distortion. |
Details on the algorithms used to compute signal and noise on the detector and the strategies used to compute the extracted products can be found in Pontoppidan et al. 2016.
