Skip to content

yymao/generic-catalog-reader

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Generic Catalog Reader (GCR)

Conda Version PyPIVversion

A ready-to-use, abstract base class for creating a common reader interface that accesses generic table-like catalogs.

This project was started in response to the need of the DESCQA framework. It is now used in the LSST DESC GCR Catalogs.

Installation

You can install GCR from conda-forge:

conda install gcr --channel conda-forge

Or from PyPI:

pip install GCR

Concept

The reader should specify: (1) how to translate (assemble) requested quantities from the native quantities; and (2) how to access native quantities from the underlying data format.

Concept

Usage

You can find API documentation here. However, looking at some real examples is probably more useful.

Basically, you will subclass GCR.BaseGenericCatalog and then set the member dict _quantity_modifiers inside _subclass_init, and implement the member methods _generate_native_quantity_list and _iter_native_dataset. Here's an minimal example.

import h5py
import GCR

class YourCatalogReader(GCR.BaseGenericCatalog):
    
    def _subclass_init(self, **kwargs):
        self._file = kwargs['filename']
        
        self._quantity_modifiers = {
            'galaxy_id' :    'galaxyID',
            'ra':            (lambda x: x/3600.0, 'ra'),
            'dec':           (lambda x: x/3600.0, 'dec'),
            'is_central':    (lambda x, y: x == y, 'haloId', 'parentHaloId'),
        }
        
    def _generate_native_quantity_list(self):
        """
        Must return an iterable of all native quantity names.
        """
        with h5py.File(self._file, 'r') as fh:
            return fh.keys()
        
    def _iter_native_dataset(self, native_filters=None):
        """
        Must be a generator.
        Must yield a callable, *native_quantity_getter*.
        This function must iterate over subsets of rows, not columns!
        Below are specifications of *native_quantity_getter*
        ---
        Must take a single argument of a native quantity name.
        Should assume the argument is valid.
        Must return a numpy 1d array.
        """
        assert not native_filters, '*native_filters* is not supported'
        with h5py.File(self._file, 'r') as fh:
            def native_quantity_getter(native_quantity):
                return fh[native_quantity].value
            yield native_quantity_getter