The Substamp Class
The Substamp class is a data container that holds all information about a single substamp used for kernel fitting.
Data models for the HOTPANTS wrapper.
- class hotpants.models.Substamp(substamp_id: int, stamp_group_id: int, x: int, y: int)
Bases:
objectData class to hold all information about a single substamp.
A substamp is a small cutout of an image centered on a bright, isolated star, used to model the convolution kernel. This class tracks the data and status of each substamp throughout the pipeline.
- id
A unique identifier for the substamp.
- Type:
int
- stamp_group_id
The identifier for the parent stamp region.
- Type:
int
- x
The x-coordinate of the substamp center.
- Type:
float
- y
The y-coordinate of the substamp center.
- Type:
float
- status
The current processing status of the substamp.
- Type:
- image_cutout
The pixel data from the science image.
- Type:
np.ndarray
- template_cutout
The pixel data from the template image.
- Type:
np.ndarray
- noise_variance_cutout
The noise variance in the cutout region.
- Type:
np.ndarray
- fit_results
A dictionary containing results from local fits.
- Type:
dict
- local_kernel_solution
The kernel solution coefficients derived from this substamp alone.
- Type:
np.ndarray
- convolved_model_local
The convolved model of this substamp using its local solution.
- Type:
np.ndarray
- convolved_model_global
The convolved model of this substamp using the final global solution.
- Type:
np.ndarray
- basis_vectors
A 3D array representing the kernel basis functions convolved with the image data at this location. This is the core component for the linear fit, as described in the Alard & Lupton (1998) paper. Each 2D slice of this array is one of the Gaussian-Laguerre basis functions convolved with the local image data, forming the design matrix for the least-squares fit.
- Type:
np.ndarray