Skip to main content

Package to efficiently read large fits arrays in object by object

Project description

Package to sequentially efficiently read large fits arrays in object by object

  • Free software: MIT license


  • Sequentially read in large fits files, within a given fixed memory limit

Quick usage

The following example shows an example of reading in a large fits hdu within a memory limit of 2048MB, assuming light curves are stored in rows:

import numpy as np
import fitsio
import fitsiochunked as fc

with fitsio.FITS(filename) as infile:
    hdu = infile['flux']
    napertures = hdu.get_info()['ndim'][0]
    mean_flux = np.zeros(napertures)

    for chunk in fc.chunks(hdu, memory_limit_mb=2048):

        # `chunk` is a namedtuple with `.data` and `.slice` properties
        chunk_data =
        print('Data shape:', chunk_data.shape)
        print('Data dtype:', chunk_data.dtype)

        chunk_slice = chunk.slice
        print('Chunk starting from aperture:', chunk_slice.start)
        print('Chunk up to:', chunk_slice.stop)

        chunk_mean = np.average(chunk_data, axis=1)
        mean_flux[chunk_slice] = chunk_mean

The library copes with an aribtrary number of hdus:

import numpy as np
import fitsio
import fitsiochunked as fc

with fitsio.FITS(filename) as infile:
    hjd_hdu = infile['hjd']
    flux_hdu = infile['flux']
    fluxerr_hdu = infile['fluxerr']

    napertures = flux_hdu.get_info()['ndim'][0]
    mean_flux = np.zeros(napertures)

    for chunks in fc.chunks(hjd_hdu, flux_hdu, fluxerr_hdu, memory_limit_mb=2048):
        # chunks is a tuple of chunks
        hjd_chunk, flux_chunk, fluxerr_chunk = chunks

        # `chunk` is a namedtuple with `.data` and `.slice` properties
        flux_chunk_data =
        print('Data shape:', flux_chunk_data.shape)
        print('Data dtype:', flux_chunk_data.dtype)

        # and so on

Note: if multiple hdus are supplied, then the memory_limit_mb and chunksize arguments to chunks apply to each HDU i.e. three HDUs and a memory limit of 2048MB will lead to 3x2048 = 6144MB of memory used.


Install with pip:

pip install fitsiochunked
# or get the latest development version from github
pip install git+

or download and run the setup file:

git clone
cd fitsiochunked
python install


The high level interface is the chunks function, which builds a ChunkedAdapter object wrapping a fitsio.ImageHDU object.

The ChunkedAdapter wraps a fitsio HDU object. When constructed, it becomes a callable which yields the image data in that hdu in chunks.

The chunksize can be set either with with the parameter chunksize which simply yields chunksize rows each time, or with memory_limit_mb which tries (no promises!) to automatically calculate the number of lightcurves that will fit into memory_limit_mb megabytes of memory.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

fitsiochunked-0.2.1.tar.gz (6.3 kB view hashes)

Uploaded source

Built Distribution

fitsiochunked-0.2.1-py2.py3-none-any.whl (5.9 kB view hashes)

Uploaded py2 py3

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Huawei Huawei PSF Sponsor Microsoft Microsoft PSF Sponsor NVIDIA NVIDIA PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page