A few PEP8 tweaks, plus remove linefeeds from a config field doc
string and standardize the indentation for one block of code.

Use standard alias for numpy and move the import
to the standard location

Change TextReaderTask._DefaultName to 'catalogReader'
a name sufficiently generic to be used by variant readers.

Add a new catalog reader FitsReaderTask and associated config
to read catalogs from FITS files.
Add a test for the new task in a new test file
tests/testIndexReference.py (which presently on has the one test)

Apply autopep8, remove unused imports,
and stop setting unused variables.

import numpy -> import numpy as np
import testFindAstrometryNetDataDir ->
    from . import testFindAstrometryNetDataDir
standardize grouping of imports

Rename TextReaderTask/Config to ReadTextCatalogTask/Config
Rename FitsReaderTask/Config to ReadFitsCatalogTask/Config
Rename "readFile" method to "run"
Update unit tests accordingly

Move catalog reader tasks to new module readCatalog.py
Rename tests/testIndexReference.py to tests/testReadCatalog.py
since it only tests a reader

Add missing
  from __future__ import absolute_import, division, print_function
to all three index catalog reader modules

Also put table columns into a more reasonable order,
since the code no longer rearranges them.