UVH5Writer#

class pyvisgen.io.datawriters.UVH5Writer(output_path: Path, dataset_type: str, **kwargs)[source]#

Bases: DataWriter

HDF5 file writer for UV-plane simulation data.

This writer saves visibilities, UVW coordinates, LMN coordinates, and the simulated sky to a single HDF5 file per sample. The file layout is:

{dataset_type}_{index}.uvh5
├── visibilities/
│   ├── V_11     (complex128)
│   ├── V_22     (complex128)
│   ├── V_12     (complex128)
│   ├── V_21     (complex128)
│   └── weights  (float64)
├── uvw/
│   ├── u
│   ├── v
│   ├── w
│   └── st_id_pairs  (int64, shape n_baselines x 2)
├── lmn/
│   ├── l
│   ├── m
│   └── n
├── frequency_bands
└── sky/
    └── SI

Parameters:

output_pathstr or Path: Directory path where HDF5 files will be written.
dataset_typestr: Type of dataset being written (e.g., ‘train’, ‘test’, ‘validation’). Used in the output filename pattern.

Examples

>>> writer = UVH5Writer(output_path="./data", dataset_type="train")
>>> writer.write(vis_data, obs, index=0, sky=SI)

Or as a context manager:

>>> with UVH5Writer(output_path="./data", dataset_type="train") as writer:
...     writer.write(vis_data, obs, index=0, sky=SI)

Methods Summary

`get_half_image`(x, y[, overlap])	Extract half height of every image with a small overlap.
`test_shapes`(array, name)	Validate the shape of input arrays.
`write`(vis_data, obs, index[, sky])	Write simulation data to an HDF5 file.

Methods Documentation

get_half_image(x: ndarray, y: ndarray, overlap: int = 5) → tuple[ndarray]#

Extract half height of every image with a small overlap.

Parameters:

xnp.ndarray: Simulated data array with shape (B, C, H, W).
ynp.ndarray: Ground truth array with shape (B, C, H, W).

Returns:

tuple[np.ndarray, np.ndarray]: Tuple containing the cropped x and y arrays.

test_shapes(array: ndarray, name: str) → None#

Validate the shape of input arrays.

Arrays should have the shape (B, C, H, W), where B is the batch size, C the number of channels (2), and W and H the width and height of the images.

Parameters:

arraynp.ndarray: Array to validate.
namestr: Name of the array for error reporting.

Raises:

ValueError: If array axis 1 is not size 2.
ValueError: If array does not have exactly 4 dimensions.

write(vis_data, obs, index: int, sky=None, **kwargs) → None[source]#

Write simulation data to an HDF5 file.

Creates a new HDF5 file for each sample with pattern uvh5_{dataset_type}_{index}.uvh5.

Parameters:

vis_dataVisibilities: Visibilities dataclass object from vis_loop(), containing V_11, V_22, V_12, V_21, u, v, w, and related tensors.
obsObservation: Observation object from Observation. Used to retrieve the LMN coordinate grid via obs.lm.
indexint: Sample index used in the output filename.
skytorch.Tensor or np.ndarray, optional: Sky intensity distribution (SI) passed to the visibility simulation, with shape (C, H, W). If None, the sky/ group is omitted from the output file.

Examples

>>> writer = UVH5Writer(output_path="./data", dataset_type="train")
>>> writer.write(vis_data, obs, index=0, sky=SI)