ASDF - Advanced Scientific Data Format

The Advanced Scientific Data Format (ASDF) is a next-generation interchange format for scientific data. This package contains the Python implementation of the ASDF Standard. More information on the ASDF Standard itself can be found here.

The ASDF format has the following features:

• A hierarchical, human-readable metadata format (implemented using YAML)
• Numerical arrays are stored as binary data blocks which can be memory mapped. Data blocks can optionally be compressed.
• The structure of the data can be automatically validated using schemas (implemented using JSON Schema)
• Native Python data types (numerical types, strings, dicts, lists) are serialized automatically
• ASDF can be extended to serialize custom data types

ASDF is under active development on github. More information on contributing can be found below.

Overview

This section outlines basic use cases of the ASDF package for creating and reading ASDF files.

Creating a file

We're going to store several numpy arrays and other data to an ASDF file. We do this by creating a "tree", which is simply a dict, and we provide it as input to the constructor of AsdfFile:

import asdf
import numpy as np

# Create some data
sequence = np.arange(100)
squares = sequence**2
random = np.random.random(100)

# Store the data in an arbitrarily nested dictionary
tree = {
    "foo": 42,
    "name": "Monty",
    "sequence": sequence,
    "powers": {"squares": squares},
    "random": random,
}

# Create the ASDF file object from our data tree
af = asdf.AsdfFile(tree)

# Write the data to a new file
af.write_to("example.asdf")

If we open the newly created file's metadata section, we can see some of the key features of ASDF on display:

#ASDF 1.0.0
#ASDF_STANDARD 1.2.0
%YAML 1.1
%TAG ! tag:stsci.edu:asdf/
--- !core/asdf-1.1.0
asdf_library: !core/software-1.0.0 {author: The ASDF Developers, homepage: 'http://github.com/asdf-format/asdf',
  name: asdf, version: 2.0.0}
history:
  extensions:
  - !core/extension_metadata-1.0.0
    extension_class: asdf.extension.BuiltinExtension
    software: {name: asdf, version: 2.0.0}
foo: 42
name: Monty
powers:
  squares: !core/ndarray-1.0.0
    source: 1
    datatype: int64
    byteorder: little
    shape: [100]
random: !core/ndarray-1.0.0
  source: 2
  datatype: float64
  byteorder: little
  shape: [100]
sequence: !core/ndarray-1.0.0
  source: 0
  datatype: int64
  byteorder: little
  shape: [100]
...

The metadata in the file mirrors the structure of the tree that was stored. It is hierarchical and human-readable. Notice that metadata has been added to the tree that was not explicitly given by the user. Notice also that the numerical array data is not stored in the metadata tree itself. Instead, it is stored as binary data blocks below the metadata section (not shown above).

It is possible to compress the array data when writing the file:

af.write_to("compressed.asdf", all_array_compression="zlib")

The built-in compression algorithms are 'zlib', and 'bzp2'. The 'lz4' algorithm becomes available when the lz4 package is installed. Other compression algorithms may be available via extensions.

Reading a file

To read an existing ASDF file, we simply use the top-level open function of the asdf package:

import asdf

af = asdf.open("example.asdf")

The open function also works as a context handler:

with asdf.open("example.asdf") as af:
    ...

Warning

The memmap argument replaces copy_arrays as of ASDF 4.0 (memmap == not copy_arrays).

To get a quick overview of the data stored in the file, use the top-level AsdfFile.info() method:

>>> import asdf
>>> af = asdf.open("example.asdf")
>>> af.info()
root (AsdfObject)
├─asdf_library (Software)
│ ├─author (str): The ASDF Developers
│ ├─homepage (str): http://github.com/asdf-format/asdf
│ ├─name (str): asdf
│ └─version (str): 2.8.0
├─history (dict)
│ └─extensions (list)
│   └─[0] (ExtensionMetadata)
│     ├─extension_class (str): asdf.extension.BuiltinExtension
│     └─software (Software)
│       ├─name (str): asdf
│       └─version (str): 2.8.0
├─foo (int): 42
├─name (str): Monty
├─powers (dict)
│ └─squares (NDArrayType): shape=(100,), dtype=int64
├─random (NDArrayType): shape=(100,), dtype=float64
└─sequence (NDArrayType): shape=(100,), dtype=int64

The AsdfFile behaves like a Python dict, and nodes are accessed like any other dictionary entry:

>>> af["name"]
'Monty'
>>> af["powers"]
{'squares': <array (unloaded) shape: [100] dtype: int64>}

Array data remains unloaded until it is explicitly accessed:

>>> af["powers"]["squares"]
array([   0,    1,    4,    9,   16,   25,   36,   49,   64,   81,  100,
        121,  144,  169,  196,  225,  256,  289,  324,  361,  400,  441,
        484,  529,  576,  625,  676,  729,  784,  841,  900,  961, 1024,
       1089, 1156, 1225, 1296, 1369, 1444, 1521, 1600, 1681, 1764, 1849,
       1936, 2025, 2116, 2209, 2304, 2401, 2500, 2601, 2704, 2809, 2916,
       3025, 3136, 3249, 3364, 3481, 3600, 3721, 3844, 3969, 4096, 4225,
       4356, 4489, 4624, 4761, 4900, 5041, 5184, 5329, 5476, 5625, 5776,
       5929, 6084, 6241, 6400, 6561, 6724, 6889, 7056, 7225, 7396, 7569,
       7744, 7921, 8100, 8281, 8464, 8649, 8836, 9025, 9216, 9409, 9604,
       9801])

>>> import numpy as np
>>> expected = [x**2 for x in range(100)]
>>> np.equal(af["powers"]["squares"], expected).all()
True

By default, uncompressed data blocks are memory mapped for efficient access. Memory mapping can be disabled by using the memmap option of open when reading:

af = asdf.open("example.asdf", memmap=False)

For more information and for advanced usage examples, see the documentation.

Extending ASDF

Out of the box, the asdf package automatically serializes and deserializes native Python types. It is possible to extend asdf by implementing custom tags that correspond to custom user types. More information on extending ASDF can be found in the official documentation.

Installation

Stable releases of the ASDF Python package are registered at PyPi. The latest stable version can be installed using pip:

$ pip install asdf

The latest development version of ASDF is available from the main branch on github. To clone the project:

$ git clone https://github.com/asdf-format/asdf

To install:

$ cd asdf
$ pip install .

To install in development mode:

$ pip install -e .

Testing

To install the test dependencies from a source checkout of the repository:

$ pip install -e ".[tests]"

To run the unit tests from a source checkout of the repository:

$ pytest

It is also possible to run the test suite from an installed version of the package.

$ pip install "asdf[tests]"
$ pytest --pyargs asdf

It is also possible to run the tests using tox.

$ pip install tox

To list all available environments:

$ tox -va

To run a specific environment:

$ tox -e <envname>

Documentation

More detailed documentation on this software package can be found here.

More information on the ASDF Standard itself can be found here.

There are two mailing lists for ASDF:

• asdf-users

• asdf-developers

  If you are looking for the Adaptable Seismic Data Format, information can be found here.

License

ASDF is licensed under a BSD 3-clause style license. See LICENSE.rst for the licenses folder for licenses for any included software.

Contributing

We welcome feedback and contributions to the project. Contributions of code, documentation, or general feedback are all appreciated. Please follow the contributing guidelines to submit an issue or a pull request.

We strive to provide a welcoming community to all of our users by abiding to the Code of Conduct.