---
license: cc-by-4.0
pretty_name: Space-based (JWST) 3d data cubes
tags:
- astronomy
- compression
- images
dataset_info:
  config_name: tiny
  features:
  - name: image
    dtype:
      array3_d:
        shape:
        - 2048
        - 2048
        dtype: uint8
  - name: ra
    dtype: float64
  - name: dec
    dtype: float64
  - name: pixscale
    dtype: float64
  - name: ntimes
    dtype: int64
  - name: image_id
    dtype: string
  splits:
  - name: train
    num_bytes: 100761802
    num_examples: 2
  - name: test
    num_bytes: 75571313
    num_examples: 1
  download_size: 201496920
  dataset_size: 176333115
---

# SBI-16-3D Dataset

SBI-16-3D is a dataset which is part of the AstroCompress project. It contains data assembled from the James Webb Space Telescope (JWST). <TODO>Describe data format</TODO>

# Usage

You first need to install the `datasets` and `astropy` packages:

```bash
pip install datasets astropy
```

There are two datasets: `tiny` and `full`, each with `train` and `test` splits. The `tiny` dataset has 2 4D images in the `train` and 1 in the `test`. The `full` dataset contains all the images in the `data/` directory.


## Local Use (RECOMMENDED)

You can clone this repo and use directly without connecting to hf:

```bash
git clone https://huggingface.co/datasets/AstroCompress/SBI-16-3D
```

```bash
git lfs pull
```

Then `cd SBI-16-3D` and start python like:

```python
from datasets import load_dataset
import numpy
dataset = load_dataset("./SBI-16-3D.py", "tiny", data_dir="./data/", writer_batch_size=1, trust_remote_code=True)
ds = dataset.with_format("np", dtype=numpy.uint16)
```

Now you should be able to use the `ds` variable like:

```python
ds["test"][0]["image"].shape # -> (5, 2048, 2048)
```

Note of course that it will take a long time to download and convert the images in the local cache for the `full` dataset. Afterward, the usage should be quick as the files are memory-mapped from disk.



## Use from Huggingface Directly

This method may only be an option when trying to access the "tiny" version of the dataset.

To directly use from this data from Huggingface, you'll want to log in on the command line before starting python:

```bash
huggingface-cli login
```

or

```
import huggingface_hub
huggingface_hub.login(token=token)
```

Then in your python script:

```python
from datasets import load_dataset
import numpy
dataset = load_dataset("AstroCompress/SBI-16-3D", "tiny", writer_batch_size=1, trust_remote_code=True)
ds = dataset.with_format("np", columns=["image"], dtype=numpy.uint16)

# or torch
import torch
dst = dataset.with_format("torch", columns=["image"], dtype=torch.uint16)

# or pandas
dsp = dataset.with_format("pandas", columns=["image"], dtype=numpy.uint16)
```


## Demo Colab Notebook
We provide a demo collab notebook to get started on using the dataset [here](https://colab.research.google.com/drive/1SuFBPZiYZg9LH4pqypc_v8Sp99lShJqZ?usp=sharing).

## Utils scripts
Note that utils scripts such as `eval_baselines.py` must be run from the parent directory of `utils`, i.e. `python utils/eval_baselines.py`.