astropy.config
astropy.constants
- Added
b_wien
to represent Wien wavelength displacement law constant. [#2194]
- Added
astropy.convolution
- Changed the input parameter in
Gaussian1DKernel
andGaussian2DKernel
fromwidth
tostddev
[#2085].
- Changed the input parameter in
astropy.coordinates
- The coordinates package has undergone major changes to implement APE5 . These include backwards-incompatible changes, as the underlying framework has changed substantially. See the APE5 text and the package documentation for more details. [#2422]
- A position_angle method has been added to the new
SkyCoord
. [#2487] - Updated
Angle.dms
andAngle.hms
to returnnamedtuple
-s instead of regular tuples, and addedAngle.signed_dms
attribute that gives the absolute value of thed
,m
, ands
along with the sign. [#1988] - By default,
Distance
objects are now required to be positive. To allow negative values, setallow_negative=True
in theDistance
constructor when creating aDistance
instance. Longitude
(resp.Latitude
) objects cannot be used any more to initialize or setLatitude
(resp.Longitude
) objects. An explicit conversion toAngle
is now required. [#2461]- The deprecated functions for pre-0.3 coordinate object names like ICRSCoordinates have been removed. [#2422]
astropy.cosmology
- Added
z_at_value
function to find the redshift at which a cosmology function matches a desired value. [#1909] - Added
FLRW.differential_comoving_volume
method to give the differential comoving volume at redshift z. [#2103] - The functional interface is now deprecated in favor of the more-explicit use of methods on cosmology objects. [#2343]
- Added
astropy.io.ascii
- The
astropy.io.ascii
output formatslatex
andaastex
accept a dictionary calledlatex_dict
to specify options for LaTeX output. It is now possible to specify the table alignment within the text via thetablealign
keyword. [#1838] - If
header_start
is specified in a call toascii.get_reader
or any method that callsget_reader
(e.g.ascii.read
) butdata_start
is not specified at the same time, thendata_start
is calculated so that the data starts after the header. Before this, the default was that the header line was read again as the first data line [#855 and #1844]. - A new
csv
format was added as a convenience for handling CSV (comma- separated values) data. [#1935] This format also recognises rows with an inconsistent number of elements. [#1562] - An option was added to guess the start of data for CDS format files when they do not strictly conform to the format standard. [#2241]
- Added an HTML reader and writer to the
astropy.io.ascii
package. Parsing requires the installation of BeautifulSoup and is therefore an optional feature. [#2160] - Added support for inputting column descriptions and column units
with the
io.ascii.SExtractor
reader. [#2372] - Allow the use of non-local ReadMe files in the CDS reader. [#2329]
- Provide a mechanism to select how masked values are printed. [#2424]
- The
astropy.io.fits
- Included a new command-line script called
fitsheader
to display the header(s) of a FITS file from the command line. [#2092]
- Included a new command-line script called
astropy.io.misc
astropy.io.registry
astropy.io.votable
astropy.modeling
- Changed
Gaussian2D
model such thattheta
now increases counterclockwise. [#2199] - Replaced the
MatrixRotation2D
model with a new model called simplyRotation2D
which requires only an angle to specify the rotation. The newRotation2D
rotates in a counter-clockwise sense whereas the oldMatrixRotation2D
increased the angle clockwise. [#2266, #2269] - Added a new
AffineTransformation2D
model which serves as a replacement for the capability ofMatrixRotation2D
to accept an arbitrary matrix, while also adding a translation capability. [#2269] - Added
GaussianAbsorption1D
model. [#2215] - New
Redshift
model [#2176].
- Changed
astropy.nddata
astropy.stats
- Added flat prior to binom_conf_interval and binned_binom_proportion
astropy.sphinx
- The
automodapi
andautomodsumm
extensions now include sphinx configuration options to write out whatautomodapi
andautomodsumm
generate, mainly for debugging purposes. [#1975, #2022] - Reference documentation now shows functions/class docstrings at the inteded user-facing API location rather than the actual file where the implementation is found. [#1826]
- The
automodsumm
extension configuration was changed to generate documentation of class__call__
member functions. [#1817, #2135] automodapi
andautomodsumm
now have an:allowed-package-names:
option that make it possible to document functions and classes that are in a different namespace. [#2370]
- The
astropy.table
- Allow row-oriented data input using a new
rows
keyword argument. [#850] - Allow subclassing of
Table
and the component classesRow
,Column
,MaskedColumn
,TableColumns
, andTableFormatter
. [#2287] - Fix to allow numpy integer types as valid indices into tables in Python 3.x [#2477]
- Allow row-oriented data input using a new
astropy.time
- Mean and apparent sidereal time can now be calculated using the
sidereal_time
method [#1418]. - The time scale now defaults to UTC if no scale is provided. [#2091]
TimeDelta
objects can have all scales but UTC, as well as, for consistency with time-like quantities, undefined scale (where the scale is taken from the object one adds to or subtracts from). This allows, e.g., to work consistently in TDB. [#1932]Time
now supports ISO format strings that end in "Z". [#2211, #2203]
- Mean and apparent sidereal time can now be calculated using the
astropy.units
- Support for the unit format Office of Guest Investigator Programs (OGIP) FITS files has been added. [#377]
- The
spectral
equivalency can now handle angular wave number. [#1306 and #1899] - Added
one
as a shorthand fordimensionless_unscaled
. [#1980] - Added
dex
anddB
units. [#1628] - Added
temperature()
equivalencies to support conversion between Kelvin, Celsius, and Fahrenheit. [#2209]
astropy.utils
timer.RunTimePredictor
now usesastropy.modeling
in itsdo_fit()
method. [#1896]
astropy.vo
- A new sub-package,
astropy.vo.samp
, is now available (this was previously the SAMPy package, which has been refactored for use in Astropy). [#1907] - Enhanced functionalities for
VOSCatalog
andVOSDatabase
. [#1206]
- A new sub-package,
astropy.wcs
- astropy now requires wcslib version 4.23 or later. The version of wcslib included with astropy has been updated to version 4.23.
- Bounds checking is now performed on native spherical
coordinates. Any out-of-bounds values will be returned as
NaN
, and marked in thestat
array, if using the low-levelwcslib
interface such asastropy.wcs.Wcsprm.p2s
. [#2107] - A new method,
astropy.wcs.WCS.compare()
, compares two wcsprm structs for equality with varying degrees of strictness. [#2361]
astropy.config
- The configuration system received a major overhaul, as part of APE3. It is no longer possible to save configuration items from Python, but instead users must edit the configuration file directly. The locations of configuration items have moved, and some have been changed to science state values. The old locations should continue to work until astropy 0.5, but deprecation warnings will be displayed. See :ref:`config-0-4-transition` for a detailed description of the changes and how to update existing code. [#2094]
astropy.constants
astropy.convolution
astropy.coordinates
astropy.cosmology
astropy.io.ascii
astropy.io.fits
astropy.io.misc
astropy.io.registry
astropy.io.votable
astropy.modeling
- The method computing the derivative of the model with respect
to parameters was renamed from
deriv
tofit_deriv
. [#1739] ParametricModel
and the associatedParametric1DModel
andParametric2DModel
classes have been renamedFittableModel
,Fittable1DModel
, andFittable2DModel
respectively. The baseModel
class has subsumed the functionality of the oldParametricModel
class so that all models support parameter constraints. The only distinction ofFittableModel
is that anything which subclasses it is assumed "safe" to use with Astropy fitters. [#2276]
- The method computing the derivative of the model with respect
to parameters was renamed from
astropy.nddata
astropy.stats
astropy.table
- The default table printing function now shows a table header row for units if any columns have the unit attribute set. [#1282]
- Before, an unmasked
Table
was automatically converted to a masked table if generated from a masked Table or aMaskedColumn
. Now, this conversion is only done if explicitly requested or if any of the input values is actually masked. [#1185] - The repr() function of
astropy.table.Table
now shows the units if any columns have the unit attribute set. [#2180]
astropy.time
- Correct use of UT in TDB calculation [#1938, #1939].
TimeDelta
objects can have scales other than TAI [#1932].- Location information should now be passed on via an
EarthLocation
instance or anything that initialises it, e.g., a tuple containing either geocentric or geodetic coordinates. [#1928]
astropy.units
Quantity
now converts input to float by default, as this is physically most sensible for nearly all units [#1776].Quantity
comparisons with==
or!=
now always returnTrue
orFalse
, even if units do not match (for which case aUnitsError
used to be raised). [#2328]- Applying
float
orint
to aQuantity
now works for all dimensionless quantities; they are automatically converted to unscaled dimensionless. [#2249] - The exception
astropy.units.UnitException
, which was deprecated in astropy 0.2, has been removed. Useastropy.units.UnitError
instead [#2386]
astropy.utils
astropy.vo
astropy.wcs
calcFootprint
was deprecated. It is replaced bycalc_footprint
. An optional boolean keywordcenter
was added tocalc_footprint
. It controls whether the centers or the corners of the pixels are used in the computation. [#2384]astropy.wcs.WCS.sip_pix2foc
andastropy.wcs.WCS.sip_foc2pix
formerly did not conform to theSIP
standard:CRPIX
was added to thefoc
result so that it could be used as input to "core FITS WCS". As of astropy 0.4,CRPIX
is no longer added to the result, so thefoc
space is correct as defined in the SIP convention. [#2360]astropy.wcs.UnitConverter
, which was deprecated in astropy 0.2, has been removed. Use theastropy.units
module instead. [#2386]- The following methods on
astropy.wcs.WCS
, which were deprecated in astropy 0.1, have been removed [#2386]:all_pix2sky
->all_pix2world
wcs_pix2sky
->wcs_pix2world
wcs_sky2pix
->wcs_world2pix
- The
naxis1
andnaxis2
attributes and theget_naxis
method ofastropy.wcs.WCS
, which were deprecated in astropy 0.2, have been removed. Use the shape of the underlying FITS data array instead. [#2386]
astropy.config
astropy.constants
astropy.convolution
astropy.coordinates
astropy.cosmology
- The distance modulus function in
astropy.cosmology
can now handle negative distances, which can occur in certain closed cosmologies. [#2008] - Removed accidental imports of some extraneous variables in
astropy.cosmology
[#2025]
- The distance modulus function in
astropy.io.ascii
astropy.io.fits
- Use NaN for missing values in FITS when using Table.write for float columns. Earlier the default fill value was close to 1e20.[#2186]
astropy.io.misc
astropy.io.registry
astropy.io.votable
- It is now possible to save an
astropy.table.Table
object as a VOTable with any of the supported data formats,tabledata
,binary
andbinary2
, by using thetabledata_format
kwarg. [#2138]
- It is now possible to save an
astropy.modeling
astropy.nddata
- Change construction of result of arithmetic and unit conversion to allow subclasses to require the presence of attribute like unit. [#2300]
astropy.stats
astropy.table
- The
astropy.table.Column
object can now use both functions and callable objects as formats. [#2313]
- The
astropy.time
- Correct UT1->UTC->UT1 round-trip being off by 1 second if UT1 is on a leap second. [#2077]
astropy.units
Quantity.copy
now behaves identically tondarray.copy
, and thus supports theorder
argument (for numpy >=1.6). [#2284]
astropy.utils
astropy.vo
astropy.wcs
- Astropy now requires wcslib version 4.23 or later. The version of wcslib included with astropy has been updated to version 4.23.
- Bug fixes in the projection routines: in
hpxx2s
[the cartesian-to-spherical operation of theHPX
projection] relating to bounds checking, bug introduced at wcslib 4.20; inparx2s
and molx2s`` [the cartesion-to-spherical operation of thePAR
andMOL
projections respectively] relating to setting the stat vector; inhpxx2s
relating to implementation of the vector API; and inxphx2s
relating to setting an out-of-bounds value of phi. - In the
PCO
projection, use alternative projection equations for greater numerical precision near theta == 0. In theCOP
projection, return an exact result for theta at the poles. Relaxed the tolerance for bounds checking a little inSFL
projection. - Fix a bug allocating insufficient memory in
astropy.wcs.WCS.sub
[#2468] - A new method,
Wcsprm.bounds_check
(corresponding to wcslib'swcsbchk
) has been added to control what bounds checking is performed by wcslib. WCS.to_header
will now raise a more meaningful exception when the WCS information is invalid or inconsistent in some way. [#1854]- In
WCS.to_header
,RESTFRQ
andRESTWAV
are no longer rewritten if zero. [#2468] - In
WCS.to_header
, floating point values will now always be written with an exponent or fractional part, i.e..0
being appended if necessary to acheive this. [#2468] - If the C extension for
astropy.wcs
was not built or fails to import for any reason,import astropy.wcs
will result in anImportError
, rather than getting obscure errors once theastropy.wcs
is used. [#2061]
- Misc
- Running the test suite with
python setup.py test
now works if the path to the source contains spaces. [#2488]
- Running the test suite with
python setup.py test --coverage
will now give more accurate results, because the coverage analysis will include early imports of astropy. There doesn't seem to be a way to get this to work when doingimport astropy; astropy.test()
, so thecoverage
keyword toastropy.test
has been removed. Coverage testing now depends only on coverage.py, notpytest-cov
. [#2112]- The included version of py.test has been upgraded to 2.5.1. [#1970]
- The included version of six.py has been upgraded to 1.5.2. [#2006]
- Where appropriate, tests are now run both with and without the
unicode_literals
option to ensure that we support both cases. [#1962] - A monkey patch is performed to fix a bug in Numpy version 1.7 and
earlier where unicode fill values on masked arrays are not
supported. This may cause unintended side effects if your
application also monkey patches
numpy.ma
or relies on the broken behavior. If unicode support of masked arrays is important to your application, upgrade to Numpy 1.8 or later for best results. [#2059] - The developer documentation has been extensively rearranged and rewritten. [#1712]
astropy.coordinates
- if
sep
argument is specified to be a single character insexagisimal_to_string
, it now includes seperators only between items [#2183] - Ensure comparisons involving
Distance
objects do not raise exceptions; also ensure operations that lead to units other than length returnQuantity
. [#2206, #2250] - Multiplication and division of
Angle
objects is now supported. [#2273] - Fixed
Angle.to_string
functionality so that negative angles have the correct amount of padding whenpad=True
. [#2337] - Mixing strings and quantities in the
Angle
constructor now works. For example:Angle(['1d', 1. * u.d])
. [#2398]
- if
astropy.cosmology
- Fixed
format()
compatibility with Python 2.6. [#2129] - Be more carful about converting to floating point internally [#1815, #1818]
- Fixed
astropy.io.ascii
- The CDS reader in
astropy.io.ascii
can now handle multiple description lines in ReadMe files. [#2225] - When reading a table with values that generate an overflow error during type conversion (e.g. overflowing the native C long type), fall through to using string. Previously this generated an exception [#2234].
- Some CDS files mark missing values with
"---"
, others with"--"
. Recognize any string with one to four dashes as null value. [#1335]
- The CDS reader in
astropy.io.fits
- Allow pickling of
FITS_rec
objects. [#1597] - Improved behavior when writing large compressed images on OSX by removing an unncessary check for platform architecture. [#2345]
- Fixed an issue where Astropy
Table
objects containing boolean columns were not correctly written out to FITS files. [#1953] - Several other bug fixes ported from PyFITS v3.2.3 [#2368]
- Fixed a crash on Python 2.x when writing a FITS file directly to a
StringIO.StringIO
object. [#2463]
- Allow pickling of
astropy.io.registry
- Allow readers/writers with the same name to be attached to different classes. [#2312]
astropy.io.votable
- By default, floating point values are now written out using
repr
rather thanstr
to preserve precision [#2137]
- By default, floating point values are now written out using
astropy.modeling
- Fixed the
SIP
andInverseSIP
models both so that they work in the first place, and so that they return results consistent with the SIP functions inastropy.wcs
. [#2177]
- Fixed the
astropy.stats
- Ensure the
axis
keyword inastropy.stats.funcs
can now be used for all axes. [#2173]
- Ensure the
astropy.table
- Ensure nameless columns can be printed, using 'None' for the header. [#2213]
astropy.time
- Fixed pickling of
Time
objects. [#2123]
- Fixed pickling of
astropy.units
Quantity._repr_latex_()
returnsNotImplementedError
for quantity arrays instead of an uninformative formatting exception. [#2258]- Ensure
Quantity.flat
always returnsQuantity
. [#2251] - Angstrom unit renders better in MathJax [#2286]
astropy.utils
- Progress bars will now be displayed inside the IPython qtconsole. [#2230]
data.download_file()
now evaluatesREMOTE_TIMEOUT()
at runtime rather than import time. Previously, settingREMOTE_TIMEOUT
after import had no effect on the function's behavior. [#2302]
astropy.vo
- Fixed
format()
compatibility with Python 2.6. [#2129] - Cone Search validation no longer raises
ConeSearchError
for positive RA. [#2240, #2242]
- Fixed
astropy.wcs
- Fixed a bug where calling
astropy.wcs.Wcsprm.sub
withWCSSUB_CELESTIAL
may cause memory corruption due to underallocation of a temporary buffer. [#2350] - Fixed a memory allocation bug in
astropy.wcs.Wcsprm.sub
andastropy.wcs.Wcsprm.copy
. [#2439]
- Fixed a bug where calling
- Misc
- Fixes for compatibility with Python 3.4. [#1945]
import astropy; astropy.test()
now correctly uses the same test configuration aspython setup.py test
[#1811]
astropy.config
- Fixed a bug where
ConfigurationItem.set_temp()
does not reset to default value when exception is raised withinwith
block. [#2117]
- Fixed a bug where
astropy.convolution
- Fixed a bug where
_truncation
was left undefined forCustomKernel
. [#2016] - Fixed a bug with
_normalization
whenCustomKernel
input array sums to zero. [#2016]
- Fixed a bug where
astropy.coordinates
- Fixed a bug where using
==
on two array coordinates wouldn't work. [#1832] - Fixed bug which caused
len()
not to work for coordinate objects and added a.shape
property to get appropriately array-like behavior. [#1761, #2014] - Fixed a bug where sexagesimal notation would sometimes include exponential notation in the last field. [#1908, #1913]
CompositeStaticMatrixTransform
no longer attempts to reference the undefined variableself.matrix
during instantiation. [#1944]- Fixed pickling of
Longitude
, ensuringwrap_angle
is preserved [#1961] - Allow
sep
argument inAngle.to_string
to be empty (resulting in no separators) [#1989]
- Fixed a bug where using
astropy.io.ascii
- Allow passing unicode delimiters when reading or writing tables. The delimiter must be convertible to pure ASCII. [#1949]
- Fix a problem when reading a table and renaming the columns to names that already exist. [#1991]
astropy.io.fits
- Ported all bug fixes from PyFITS 3.2.1. See the PyFITS changelog at http://pyfits.readthedocs.org/en/v3.2.1/ [#2056]
astropy.io.misc
- Fixed issues in the HDF5 Table reader/writer functions that occurred on Windows. [#2099]
astropy.io.votable
- The
write_null_values
kwarg toVOTable.to_xml
, when set to False (the default) would produce non-standard VOTable files. Therefore, this functionality has been replaced by a better understanding that knows which fields in a VOTable may be left empty (onlychar
,float
anddouble
in VOTable 1.1 and 1.2, and all fields in VOTable 1.3). The kwarg is still accepted but it will be ignored, and a warning is emitted. [#1809] - Printing out a
astropy.io.votable.tree.Table
object using repr or str now uses the pretty formatting inastropy.table
, so it's possible to easily preview the contents of aVOTable
. [#1766]
- The
astropy.modeling
- Fixed bug in computation of model derivatives in
LinearLSQFitter
. [#1903] - Raise a
NotImplementedError
when fitting composite models. [#1915] - Fixed bug in the computation of the
Gaussian2D
model. [#2038] - Fixed bug in the computation of the
AiryDisk2D
model. [#2093]
- Fixed bug in computation of model derivatives in
astropy.sphinx
- Added slightly more useful debug info for AstropyAutosummary. [#2024]
astropy.table
- The column string representation for n-dimensional cells with only one element has been fixed. [#1522]
- Fix a problem that caused
MaskedColumn.__getitem__
to not preserve column metadata. [#1471, #1872] - With Numpy prior to version 1.6.2, tables with Unicode columns now sort correctly. [#1867]
astropy.table
can now print out tables with Unicode columns containing non-ascii characters. [#1864]- Columns can now be named with Unicode strings, as long as they contain only
ascii characters. This makes using
astropy.table
easier on Python 2 whenfrom __future__ import unicode_literals
is used. [#1864] - Allow pickling of
Table
,Column
, andMaskedColumn
objects. [#792] - Fix a problem where it was not possible to rename columns after sorting or adding a row. [#2039]
astropy.time
- Fix a problem where scale conversion problem in TimeFromEpoch was not showing a useful error [#2046]
- Fix a problem when converting to one of the formats
unix
,cxcsec
,gps
orplot_date
when the time scale isUT1
,TDB
orTCB
[#1732] - Ensure that
delta_ut1_utc
gets calculated when accessed directly, instead of failing and giving a rather obscure error message [#1925] - Fix a bug when computing the TDB to TT offset. The transform routine was using meters instead of kilometers for the Earth vector. [#1929]
- Increase
__array_priority__
so thatTimeDelta
can convert itself to aQuantity
also in reverse operations [#1940] - Correct hop list from TCG to TDB to ensure that conversion is possible [#2074]
astropy.units
Quantity
initialisation rewritten for speed [#1775]- Fixed minor string formatting issue for dimensionless quantities. [#1772]
- Fix error for inplace operations on non-contiguous quantities [#1834].
- The definition of the unit
bar
has been corrected to "1e5 Pascal" from "100 Pascal" [#1910] - For units that are close to known units, but not quite, for example due to differences in case, the exception will now include recommendations. [#1870]
- The generic and FITS unit parsers now accept multiple slashes in the unit string. There are multiple ways to interpret them, but the approach taken here is to convert "m/s/kg" to "m s-1 kg-1". Multiple slashes are accepted, but discouraged, by the FITS standard, due to the ambiguity of parsing, so a warning is raised when it is encountered. [#1911]
- The use of "angstrom" (with a lower case "a") is now accepted in FITS unit strings, since it is in common usage. However, since it is not officially part of the FITS standard, a warning will be issued when it is encountered. [#1911]
- Pickling unrecognized units will not raise a
AttributeError
. [#2047] astropy.units
now correctly preserves the precision of fractional powers. [#2070]- If a
Unit
orQuantity
is raised to a floating point power that is very close to a rational number with a denominator less than or equal to 10, it is converted to aFraction
object to preserve its precision through complex unit conversion operations. [#2070]
astropy.utils
- Fixed crash in
timer.RunTimePredictor.do_fit
. [#1905] - Fixed
astropy.utils.compat.argparse
for Python 3.1. [#2017]
- Fixed crash in
astropy.wcs
astropy.wcs.WCS
,astropy.wcs.WCS.fix
andastropy.wcs.find_all_wcs
now have atranslate_units
keyword argument that is passed down toastropy.wcs.Wcsprm.fix
. This can be used to specify any unsafe translations of units from rarely used ones to more commonly used ones.Although
"S"
is commonly used to represent seconds, its translation to"s"
is potentially unsafe since the standard recognizes"S"
formally as Siemens, however rarely that may be used. The same applies to"H"
for hours (Henry), and"D"
for days (Debye).When these sorts of changes are performed, a warning is emitted. [#1854]
When a unit is "fixed" by
astropy.wcs.WCS.fix
orastropy.wcs.Wcsprm.unitfix
, it now correctly reports theCUNIT
field that was changed. [#1854]astropy.wcs.Wcs.printwcs
will no longer warn thatcdelt
is being ignored when none was present in the FITS file. [#1845]astropy.wcs.Wcsprm.set
is called from within theastropy.wcs.WCS
constructor, therefore any invalid information in the keywords will be raised from the constructor, rather than on a subsequent call to a transformation method. [#1918]Fix a memory corruption bug when using
astropy.wcs.Wcs.sub
withastropy.wcs.WCSSUB_CELESTIAL
. [#1960]Fixed the
AttributeError
exception that was raised when usingastropy.wcs.WCS.footprint_to_file
. [#1912]Fixed a
NameError
exception that was raised when usingastropy.wcs.validate
or thewcslint
script. [#2053]Fixed a bug where named WCSes may be erroneously reported as
' '
when usingastropy.wcs.validate
or thewcslint
script. [#2053]Fixed a bug where error messages about incorrect header keywords may not be propagated correctly, resulting in a "NULL error object in wcslib" message. [#2106]
Misc
- There are a number of improvements to make Astropy work better on big endian platforms, such as MIPS, PPC, s390x and SPARC. [#1849]
- The test suite will now raise exceptions when a deprecated feature of Python or Numpy is used. [#1948]
- A new function,
astropy.wcs.get_include
, has been added to get the location of theastropy.wcs
C header files. [#1755] - The doctests in the
.rst
files in thedocs
folder are now tested along with the other unit tests. This is in addition to the testing of doctests in docstrings that was already being performed. Seedocs/development/testguide.rst
for more information. [#1771] - Fix a problem where import fails on Python 3 if setup.py exists in current directory. [#1877]
General
A top-level configuration item,
unicode_output
has been added to control whether the Unicode string representation of certain objects will contain Unicode characters. For example, whenuse_unicode
is False (default):>>> from astropy import units as u >>> print(unicode(u.degree)) deg
When
use_unicode
is True:>>> from astropy import units as u >>> print(unicode(u.degree)) °
See handling-unicode for more information. [#1441]
astropy.utils.misc.find_api_page
is now imported into the top-level. This allows usage likeastropy.find_api_page(astropy.units.Quantity)
. [#1779]
astropy.convolution
- New class-based system for generating kernels, replacing
make_kernel
. [#1255] Theastropy.nddata.convolution
sub-package has now been moved toastropy.convolution
. [#1451]
- New class-based system for generating kernels, replacing
astropy.coordinates
- Two classes
astropy.coordinates.Longitude
andastropy.coordinates.Latitude
have been added. These are derived from the newAngle
class and used for all longitude-like (RA, azimuth, galactic L) and latitude-like coordinates (Dec, elevation, galactic B) respectively. TheLongitude
class provides auto-wrapping capability andLatitude
performs bounds checking. astropy.coordinates.Distance
supports conversion to and from distance modulii. [#1472]astropy.coordinates.SphericalCoordinateBase
and derived classes now support arrays of coordinates, enabling large speed-ups for some operations on multiple coordinates at the same time. These coordinates can also be indexed using standard slicing or any Numpy-compatible indexing. [#1535, #1615]- Array coordinates can be matched to other array coordinates, finding the
closest matches between the two sets of coordinates (see the
astropy.coordinates.matching.match_coordinates_3d
andastropy.coordinates.matching.match_coordinates_sky
functions). [#1535]
- Two classes
astropy.cosmology
- Added support for including massive Neutrinos in the cosmology classes. The Planck (2013) cosmology has been updated to use this. [#1364]
- Calculations now use and return
Quantity
objects where appropriate. [#1237]
astropy.io.ascii
- Added support for writing IPAC format tables [#1152].
astropy.io.fits
- Added initial support for table columns containing pseudo-unsigned
integers. This is currently enabled by using the
uint=True
option when opening files; any table columns with the correct BZERO value will be interpreted and returned as arrays of unsigned integers. [#906] - Upgraded vendored copy of CFITSIO to v3.35, though backwards compatibility back to version v3.28 is maintained.
- Added support for reading and writing tables using the Q format for columns. The Q format is identical to the P format (variable-length arrays) except that it uses 64-bit integers for the data descriptors, allowing more than 4 GB of variable-length array data in a single table.
- Some refactoring of the table and
FITS_rec
modules in order to better separate the details of the FITS binary and ASCII table data structures from the HDU data structures that encapsulate them. Most of these changes should not be apparent to users (but see API Changes below).
- Added initial support for table columns containing pseudo-unsigned
integers. This is currently enabled by using the
astropy.io.votable
- Updated to support the VOTable 1.3 draft. [#433]
- Added the ability to look up and group elements by their utype attribute. [#622]
- The format of the units of a VOTable file can be specified using the
unit_format
parameter. Note that units are still always written out using the CDS format, to ensure compatibility with the standard.
astropy.modeling
- Added a new framework for representing and evaluating mathematical models and for fitting data to models. See "What's New in Astropy 0.3" in the documentation for further details. [#493]
astropy.stats
- Added robust statistics functions
astropy.stats.funcs.median_absolute_deviation
,astropy.stats.funcs.biweight_location
, andastropy.stats.funcs.biweight_midvariance
. [#621] - Added
astropy.stats.funcs.signal_to_noise_oir_ccd
for computing the signal to noise ratio for source being observed in the optical/IR using a CCD. [#870] - Add
axis=int
option tostropy.stats.funcs.sigma_clip
to allow clipping along a given axis for multidimensional data. [#1083]
- Added robust statistics functions
astropy.table
- New columns can be added to a table via assignment to a non-existing column by name. [#726]
- Added
join
function to perform a database-like join on two tables. This includes support for inner, left, right, and outer joins as well as metadata merging. [#903] - Added
hstack
andvstack
functions to stack two or more tables. [#937] - Tables now have a
.copy
method and include support forcopy
anddeepcopy
. [#1208] - Added support for selecting and manipulating groups within a table with
a database style
group_by
method. [#1424] - Table
read
andwrite
functions now include rudimentary support reading and writing of FITS tables via the unified reading/writing interface. [#591] - The
units
anddtypes
attributes and keyword arguments in Column, MaskedColumn, Row, and Table are now deprecated in favor of the single-tenseunit
anddtype
. [#1174] - Setting a column from a Quantity now correctly sets the unit on the Column object. [#732]
- Add
remove_row
andremove_rows
to remove table rows. [#1230] - Added a new
Table.show_in_browser
method that opens a web browser and displays the table rendered as HTML. [#1342] - New tables can now be instantiated using a single row from an existing table. [#1417]
astropy.time
- New
Time
objects can be instantiated from existingTime
objects (but with different format, scale, etc.) [#889] - Added a
Time.now
classmethod that returns the current UTC time, similarly to Python'sdatetime.now
. [#1061] - Update internal time manipulations so that arithmetic with Time and TimeDelta objects maintains sub-nanosecond precision over a time span longer than the age of the universe. [#1189]
- Use
astropy.utils.iers
to providedelta_ut1_utc
, so that automatic calculation of UT1 becomes possible. [#1145] - Add
datetime
format which allows converting to and from standard librarydatetime.datetime
objects. [#860] - Add
plot_date
format which allows converting to and from the date representation used when plotting dates with matplotlib via thematplotlib.pyplot.plot_date
function. [#860] - Add
gps
format (seconds since 1980-01-01 00:00:00 UTC, including leap seconds) [#1164] - Add array indexing to Time objects [#1132]
- Allow for arithmetic of multi-element and single-element Time and TimeDelta objects. [#1081]
- Allow multiplication and division of TimeDelta objects by constants and arrays, as well as changing sign (negation) and taking the absolute value of TimeDelta objects. [#1082]
- Allow comparisons of Time and TimeDelta objects. [#1171]
- Support interaction of Time and Quantity objects that represent a time interval. [#1431]
- New
astropy.units
- Added parallax equivalency for length-angle. [#985]
- Added mass-energy equivalency. [#1333]
- Added a new-style format method which will use format specifiers
(like
0.03f
) in new-style format strings for the Quantity's value. Specifiers which can't be applied to the value will fall back to the entire string representation of the quantity. [#1383] - Added support for complex number values in quantities. [#1384]
- Added new spectroscopic equivalencies for velocity conversions (relativistic, optical, and radio conventions are supported) [#1200]
- The
spectral
equivalency now also handles wave number. - The
spectral_density
equivalency now also accepts a Quantity for the frequency or wavelength. It also handles additional flux units. - Added Brightness Temperature (antenna gain) equivalency for conversion between T_B and flux density. [#1327]
- Added percent unit, and allowed any string containing just a number to be interpreted as a scaled dimensionless unit. [#1409]
- New-style format strings can be used to set the unit output format. For
example,
"{0:latex}".format(u.km)
will print with the latex formatter. [#1462] - The
Unit.is_equivalent
method can now take a tuple. In this case, the method returnsTrue
if the unit is equivalent to any of the units listed in the tuple. [#1521] def_unit
can now take a 2-tuple of names of the form (short, long), where each entry is a list. This allows for handling strange units that might have multiple short names. [#1543]- Added
dimensionless_angles
equivalency, which allows conversion of any power of radian to dimensionless. [#1161] - Added the ability to enable set of units, or equivalencies that are used by default. Also provided context managers for these cases. [#1268]
- Imperial units are disabled by default. [#1593, #1662]
- Added an
astropy.units.add_enabled_units
context manager, which allows creating a temporary context with additional units temporarily enabled in the global units namespace. [#1662] Unit
instances now have.si
and.cgs
properties a laQuantity
. These serve as shortcuts forUnit.to_system(cgs)[0]
etc. [#1610]
astropy.vo
- New package added to support Virtual Observatory Simple Cone Search query and service validation. [#552]
astropy.wcs
- Fixed attribute error in
astropy.wcs.Wcsprm
(lattype->lattyp) [#1463] - Included a new command-line script called
wcslint
and accompanying API for validating the WCS in a given FITS file or header. [#580] - Upgraded included version of WCSLIB to 4.19.
- Fixed attribute error in
astropy.utils
- Added a new set of utilities in
astropy.utils.timer
for analyzing the runtime of functions and making runtime predections for larger inputs. [#743] ProgressBar
andSpinner
classes can now be used directly to return generator expressions. [#771]- Added
astropy.utils.iers
which allows reading in of IERS A or IERS B bulletins and interpolation in UT1-UTC. - Added a function
astropy.utils.find_api_page
--given a class or object from theastropy
package, this will open that class's API documentation in a web browser. [#663] - Data download functions such as
download_file
now accept ashow_progress
argument to suppress console output, and atimeout
argument. [#865, #1258]
- Added a new set of utilities in
astropy.extern.six
- Added six for python2/python3 compatibility
Astropy now uses the ERFA library instead of the IAU SOFA library for fundamental time transformation routines. The ERFA library is derived, with permission, from the IAU SOFA library but is distributed under a BSD license. See
license/ERFA.rst
for details. [#1293]astropy.logger
- The Astropy logger now no longer catches exceptions by default, and also
only captures warnings emitted by Astropy itself (prior to this change,
following an import of Astropy, any warning got re-directed through the
Astropy logger). Logging to the Astropy log file has also been disabled by
default. However, users of Astropy 0.2 will likely still see the previous
behavior with Astropy 0.3 for exceptions and logging to file since the
default configuration file installed by 0.2 set the exception logging to be
on by default. To get the new behavior, set the
log_exceptions
andlog_to_file
configuration items toFalse
in theastropy.cfg
file. [#1331]
- The Astropy logger now no longer catches exceptions by default, and also
only captures warnings emitted by Astropy itself (prior to this change,
following an import of Astropy, any warning got re-directed through the
Astropy logger). Logging to the Astropy log file has also been disabled by
default. However, users of Astropy 0.2 will likely still see the previous
behavior with Astropy 0.3 for exceptions and logging to file since the
default configuration file installed by 0.2 set the exception logging to be
on by default. To get the new behavior, set the
General
- The configuration option
utils.console.use_unicode
has been moved to the top level and renamed tounicode_output
. It now not only affects console widgets, such as progress bars, but also controls whether calling unicode on certain classes will return a string containing unicode characters.
- The configuration option
astropy.coordinates
- The
astropy.coordinates.Angle
class is now a subclass ofastropy.units.Quantity
. This means it has all of the methods of a numpy.ndarray. [#1006] - The
astropy.coordinates.Distance
class is now a subclass ofastropy.units.Quantity
. This means it has all of the methods of a numpy.ndarray. [#1472]- All angular units are now supported, not just
radian
,degree
andhour
, but nowarcsecond
andarcminute
as well. The object will retain its native unit, so when printing out a value initially provided in hours, itsto_string()
will, by default, also be expressed in hours. - The
Angle
class now supports arrays of angles. - To be consistent with
units.Unit
,Angle.format
has been deprecated and renamed toAngle.to_string
. - To be consistent with
astropy.units
, all plural forms of unit names have been removed. Therefore, the following properties ofastropy.coordinates.Angle
should be renamed:radians
->radian
degrees
->degree
hours
->hour
- Multiplication and division of two
Angle
objects used to raiseNotImplementedError
. Now they raiseTypeError
.
- All angular units are now supported, not just
- The
astropy.coordinates.Angle
class no longer has abounds
attribute so there is no bounds-checking or auto-wrapping at this level. This allowsAngle
objects to be used in arbitrary arithmetic expressions (e.g. coordinate distance computation). - The
astropy.coordinates.RA
andastropy.coordinates.Dec
classes have been removed and replaced withastropy.coordinates.Longitude
andastropy.coordinates.Latitude
respectively. These are now used for the components of Galactic and Horizontal (Alt-Az) coordinates as well instead of plainAngle
objects. astropy.coordinates.angles.rotation_matrix
andastropy.coordinates.angles.angle_axis
now take aunit
kwarg instead ofdegrees
kwarg to specify the units of the angles.rotation_matrix
will also take the unit from the givenAngle
object if no unit is provided.- The
AngularSeparation
class has been removed. The output of the coordinatesseparation()
method is now anastropy.coordinates.Angle
. [#1007] - The coordinate classes have been renamed in a way that remove the
Coordinates
at the end of the class names. E.g.,ICRSCoordinates
from previous versions is now calledICRS
. [#1614] HorizontalCoordinates
are now namedAltAz
, to reflect more common terminology.
- The
astropy.cosmology
- The Planck (2013) cosmology will likely give slightly different (and more accurate) results due to the inclusion of Neutrino masses. [#1364]
- Cosmology class properties now return
Quantity
objects instead of simple floating-point values. [#1237] - The names of cosmology instances are now truly optional, and are set to
None
rather than the name of the class if the user does not provide them. [#1705]
astropy.io.ascii
- In the
read
method ofastropy.io.ascii
, empty column values in an ASCII table are now treated as missing values instead of the previous treatment as a zero-length string "". This now corresponds to the behavior of other table readers likenumpy.genfromtxt
. To restore the previous behavior setfill_values=None
in the call toascii.read()
. [#919] - The
read
andwrite
methods ofastropy.io.ascii
now have aformat
argument for specifying the file format. This is the preferred way to choose the format instead of theReader
andWriter
arguments. [#961] - The
include_names
andexclude_names
arguments were removed from theBaseHeader
initializer, and now instead handled by the reader and writer classes directly. [#1350] - Allow numeric and otherwise unusual column names when reading a table
where the
format
argument is specified, but other format details such as the delimiter or quote character are being guessed. [#1692] - When reading an ASCII table using the
Table.read()
method, the default has changed fromguess=False
toguess=True
to allow auto-detection of file format. This matches the default behavior ofascii.read()
.
- In the
astropy.io.fits
- The
astropy.io.fits.new_table
function is marked "pending deprecation". This does not mean it will be removed outright or that its functionality has changed. It will likely be replaced in the future for a function with similar, if not subtly different functionality. A better, if not slightly more verbose approach is to usepyfits.FITS_rec.from_columns
to create a newFITS_rec
table--this has the same interface aspyfits.new_table
. The difference is that it returns a planFITS_rec
array, and not an HDU instance. ThisFITS_rec
object can then be used as the data argument in the constructors forBinTableHDU
(for binary tables) orTableHDU
(for ASCII tables). This is analogous to creating anImageHDU
by passing in an image array.pyfits.FITS_rec.from_columns
is just a simpler way of creating a FITS-compatible recarray from a FITS column specification. - The
updateHeader
,updateHeaderData
, andupdateCompressedData
methods of theCompDataHDU
class are pending deprecation and moved to internal methods. The operation of these methods depended too much on internal state to be used safely by users; instead they are invoked automatically in the appropriate places when reading/writing compressed image HDUs. - The
CompDataHDU.compData
attribute is pending deprecation in favor of the clearer and more PEP-8 compatibleCompDataHDU.compressed_data
. - The constructor for
CompDataHDU
has been changed to accept new keyword arguments. The new keyword arguments are essentially the same, but are in underscore_separated format rather than camelCase format. The old arguments are still pending deprecation. - The internal attributes of HDU classes
_hdrLoc
,_datLoc
, and_datSpan
have been replaced with_header_offset
,_data_offset
, and_data_size
respectively. The old attribute names are still pending deprecation. This should only be of interest to advanced users who have created their own HDU subclasses. - The following previously deprecated functions and methods have been removed
entirely:
createCard
,createCardFromString
,upperKey
,ColDefs.data
,setExtensionNameCaseSensitive
,_File.getfile
,_TableBaseHDU.get_coldefs
,Header.has_key
,Header.ascardlist
. - Interfaces that were pending deprecation are now fully deprecated. These
include:
create_card
,create_card_from_string
,upper_key
,Header.get_history
, andHeader.get_comment
. - The
.name
attribute on HDUs is now directly tied to the HDU's header, so that if.header['EXTNAME']
changes so does.name
and vice-versa.
- The
astropy.io.registry
- Identifier functions for reading/writing Table and NDData objects should
now accept
(origin, *args, **kwargs)
instead of(origin, args, kwargs)
. [#591] - Added a new
astropy.io.registry.get_formats
function for listing registered I/O formats and details about the their readers/writers. [#1669]
- Identifier functions for reading/writing Table and NDData objects should
now accept
astropy.io.votable
- Added a new option
use_names_over_ids
option to use when converting from VOTable objects to Astropy Tables. This can prevent a situation where column names are not preserved when converting from a VOTable. [#609]
- Added a new option
astropy.nddata
- The
astropy.nddata.convolution
sub-package has now been moved toastropy.convolution
, and themake_kernel
function has been removed. (the kernel classes should be used instead) [#1451]
- The
astropy.stats.funcs
- For
sigma_clip
, themaout
optional parameter has been removed, and the function now always returns a masked array. A new boolean parametercopy
can be used to indicated whether the input data should be copied (copy=True
, default) or used by reference (copy=False
) in the output masked array. [#1083]
- For
astropy.table
- The first argument to the
Column
andMaskedColumn
classes is now the data array--thename
argument has been changed to an optional keyword argument. [#840] - Added support for instantiating a
Table
from a list of dict, each one representing a single row with the keys mapping to column names. [#901] - The plural 'units' and 'dtypes' have been switched to 'unit' and 'dtype' where appropriate. The original attributes are still present in this version as deprecated attributes, but will be removed in the next version. [#1174]
- The
copy
methods ofColumn
andMaskedColumn
were changed so that the first argument is noworder='C'
. This is required for compatibility with Numpy 1.8 which is currently in development. [#1250] - Comparing a column (with == or !=) to a scalar, an array, or another column now always returns a boolean Numpy array (which is a masked array if either of the arguments in the comparison was masked). This is in contrast to the previous behavior, which in some cases returned a boolean Numpy array, and in some cases returned a boolean Column object. [#1446]
- The first argument to the
astropy.time
- For consistency with
Quantity
, the attributesval
andis_scalar
have been renamed tovalue
andisscalar
, respectively, and the attributevals
has been dropped. [#767] - The double-float64 internal representation of time is used more efficiently to enable better accuracy. [#366]
- Format and scale arguments are now allowed to be case-insensitive. [#1128]
- For consistency with
astropy.units
- The
Quantity
class now inherits from the Numpy array class, and includes the following API changes [#929]:- Using
float(...)
,int(...)
, andlong(...)
on a quantity will now only work if the quantity is dimensionless and unscaled. - All Numpy ufuncs should now treat units correctly (or raise an exception if not supported), rather than extract the value of quantities and operate on this, emitting a warning about the implicit loss of units.
- When using relevant Numpy ufuncs on dimensionless quantities (e.g.
np.exp(h * nu / (k_B * T))
), or combining dimensionless quantities with Python scalars or plain Numpy arrays1 + v / c
, the dimensionless Quantity will automatically be converted to an unscaled dimensionless Quantity. - When initializing a quantity from a value with no unit, it is now set to be dimensionless and unscaled by default. When initializing a Quantity from another Quantity and with no unit specified in the initializer, the unit is now taken from the unit of the Quantity being initialized from.
- Using
- Strings are no longer allowed as the values for Quantities. [#1005]
- Quantities are always comparable with zero regardless of their units. [#1254]
- The exception
astropy.units.UnitsException
has been renamed toastropy.units.UnitsError
to be more consistent with the naming of built-in Python exceptions. [#1406] - Multiplication with and division by a string now always returns a Unit (rather than a Quantity when the string was first) [#1408]
- Imperial units are disabled by default.
- The
astropy.wcs
For those including the
astropy.wcs
C headers in their project, they should now include it as:#include "astropy_wcs/astropy_wcs_api.h"
instead of:
#include "astropy_wcs_api.h"
[#1631]
The
--enable-legacy
option forsetup.py
has been removed. [#1493]
astropy.io.ascii
- The
write()
function was ignoring thefill_values
argument. [#910] - Fixed an issue in
DefaultSplitter.join
where the delimiter attribute was ignored when writing the CSV. [#1020] - Fixed writing of IPAC tables containing null values. [#1366]
- When a table with no header row was read without specifying the format and
using the
names
argument, then the first row could be dropped. [#1692]
- The
astropy.io.fits
- Binary tables containing compressed images may, optionally, contain other columns unrelated to the tile compression convention. Although this is an uncommon use case, it is permitted by the standard.
- Reworked some of the file I/O routines to allow simpler, more consistent mapping between OS-level file modes ('rb', 'wb', 'ab', etc.) and the more "PyFITS-specific" modes used by PyFITS like "readonly" and "update". That is, if reading a FITS file from an open file object, it doesn't matter as much what "mode" it was opened in so long as it has the right capabilities (read/write/etc.) Also works around bugs in the Python io module in 2.6+ with regard to file modes.
- Fixed a long-standing issue where writing binary tables did not correctly write the TFORMn keywords for variable-length array columns (they omitted the max array length parameter of the format). This was thought fixed in an earlier version, but it was only fixed for compressed image HDUs and not for binary tables in general.
astropy.nddata
- Fixed crash when trying to multiple or divide
NDData
objects with uncertainties. [#1547]
- Fixed crash when trying to multiple or divide
astropy.table
- Using a list of strings to index a table now correctly returns a new table with the columns named in the list. [#1454]
- Inequality operators now work properly with
Column
objects. [#1685]
astropy.time
Time
scale and format attributes are now shown when callingdir()
on aTime
object. [#1130]
astropy.wcs
- Fixed assignment to string-like WCS attributes on Python 3. [#956]
astropy.units
- Fixed a bug that caused the order of multiplication/division of plain Numpy arrays with Quantities to matter (i.e. if the plain array comes first the units were not preserved in the output). [#899]
- Directly instantiated
CompositeUnits
were made printable without crashing. [#1576]
- Misc
- Fixed various modules that hard-coded
sys.stdout
as default arguments to functions at import time, rather than using the runtime value ofsys.stdout
. [#1648] - Minor documentation fixes and enhancements [#922, #1034, #1210, #1217, #1491, #1492, #1498, #1582, #1608, #1621, #1646, #1670, #1756]
- Fixed a crash that could sometimes occur when running the test suite on systems with platform names containing non-ASCII characters. [#1698]
- Fixed various modules that hard-coded
- General
- Astropy now follows the PSF Code of Conduct. [#1216]
- Astropy's test suite now tests all doctests in inline docstrings. Support for running doctests in the reST documentation is planned to follow in v0.3.1.
- Astropy's test suite can be run on multiple CPUs in parallel, often
greatly improving runtime, using the
--parallel
option. [#1040] - A warning is now issued when using Astropy with Numpy < 1.5--much of Astropy may still work in this case but it shouldn't be expected to either. [#1479]
- Added automatic download/build/installation of Numpy during Astropy installation if not already found. [#1483]
- Handling of metadata for the
NDData
andTable
classes has been unified by way of a commonMetaData
descriptor--it allows instantiating an object with metadata of any mapping type, and subsequently prevents replacing the mapping stored in the.meta
attribute (only direct updates to that object are allowed). [#1686]
astropy.coordinates
- Angles containing out of bounds minutes or seconds (eg. 60) can be parsed--the value modulo 60 is used with carry to the hours/minutes, and a warning is issued rather than raising an exception. [#990]
astropy.io.fits
- The new compression code also adds support for the ZQUANTIZ and ZDITHER0
keywords added in more recent versions of this FITS Tile Compression spec.
This includes support for lossless compression with GZIP. (#198) By default
no dithering is used, but the
SUBTRACTIVE_DITHER_1
andSUBTRACTIVE_DITHER_2
methods can be enabled by passing the correct constants to thequantize_method
argument to theCompImageHDU
constructor. A seed can be manually specified, or automatically generated using either the system clock or checksum-based methods via thedither_seed
argument. See the documentation forCompImageHDU
for more details. - Images compressed with the Tile Compression standard can now be larger than 4 GB through support of the Q format.
- All HDUs now have a
.ver
.level
attribute that returns the value of the EXTVAL and EXTLEVEL keywords from that HDU's header, if the exist. This was added for consistency with the.name
attribute which returns the EXTNAME value from the header. - Then
Column
andColDefs
classes have new.dtype
attributes which give the Numpy dtype for the column data in the first case, and the full Numpy compound dtype for each table row in the latter case. - There was an issue where new tables created defaulted the values in all string columns to '0.0'. Now string columns are filled with empty strings by default--this seems a less surprising default, but it may cause differences with tables created with older versions of PyFITS or Astropy.
- The new compression code also adds support for the ZQUANTIZ and ZDITHER0
keywords added in more recent versions of this FITS Tile Compression spec.
This includes support for lossless compression with GZIP. (#198) By default
no dithering is used, but the
astropy.io.misc
- The HDF5 reader can now refer to groups in the path as well as datasets; if given a group, the first dataset in that group is read. [#1159]
astropy.nddata
NDData
objects have more helpful, though still rudimentary__str__` and ``__repr__
displays. [#1313]
astropy.units
- Added 'cycle' unit. [#1160]
- Extended units supported by the CDS formatter/parser. [#1468]
- Added unicode an LaTeX symbols for liter. [#1618]
astropy.wcs
- Redundant SCAMP distortion parameters are removed with SIP distortions are also present. [#1278]
- Added iterative implementation of
all_world2pix
that can be reliably inverted. [#1281]
astropy.coordinates
- Fixed incorrect string formatting of Angles using
precision=0
. [#1319] - Fixed string formatting of Angles using
decimal=True
which ignored theprecision
argument. [#1323] - Fixed parsing of format strings using appropriate unicode characters
instead of the ASCII
-
for minus signs. [#1429]
- Fixed incorrect string formatting of Angles using
astropy.io.ascii
- Fixed a crash in the IPAC table reader when the
include/exclude_names
option is set. [#1348] - Fixed writing AASTex tables to honor the
tabletype
option. [#1372]
- Fixed a crash in the IPAC table reader when the
astropy.io.fits
- Improved round-tripping and preservation of manually assigned column
attributes (
TNULLn
,TSCALn
, etc.) in table HDU headers. (Note: This issue was previously reported as fixed in Astropy v0.2.2 by mistake; it is not fixed until v0.3.) [#996] - Fixed a bug that could cause a segfault when trying to decompress an compressed HDU whose contents are truncated (due to a corrupt file, for example). This still causes a Python traceback but better that than a segfault. [#1332]
- Newly created
CompImageHDU
HDUs use the correct value of theDEFAULT_COMPRESSION_TYPE
module-level constant instead of hard-coding "RICE_1" in the header. - Fixed a corner case where when extra memory is allocated to compress an image, it could lead to unnecessary in-memory copying of the compressed image data and a possible memory leak through Numpy.
- Fixed a bug where assigning from an mmap'd array in one FITS file over the old (also mmap'd) array in another FITS file failed to update the destination file. Corresponds to PyFITS issue 25.
- Some miscellaneous documentation fixes.
- Improved round-tripping and preservation of manually assigned column
attributes (
astropy.io.votable
- Added a warning for when a VOTable 1.2 file contains no
RESOURCES
elements (at least one should be present). [#1337] - Fixed a test failure specific to MIPS architecture caused by an errant floating point warning. [#1179]
- Added a warning for when a VOTable 1.2 file contains no
astropy.nddata.convolution
- Prevented in-place modification of the input arrays to
convolve()
. [#1153]
- Prevented in-place modification of the input arrays to
astropy.table
- Added HTML escaping for string values in tables when outputting the table as HTML. [#1347]
- Added a workaround in a bug in Numpy that could cause a crash when
accessing a table row in a masked table containing
dtype=object
columns. [#1229] - Fixed an issue similar to the one in #1229, but specific to unmasked tables. [#1403]
astropy.units
- Improved error handling for unparseable units and fixed parsing CDS units without mantissas in the exponent. [#1288]
- Added a physical type for spectral flux density. [#1410]
- Normalized conversions that should result in a scale of exactly 1.0 to round off slight floating point imprecisions. [#1407]
- Added support in the CDS unit parser/formatter for unusual unit prefixes that are nonetheless required to be supported by that convention. [#1426]
- Fixed the parsing of
sqrt()
in unit format strings which was returningunit ** 2
instead ofunit ** 0.5
. [#1458]
astropy.wcs
- When passing a single array to the wcs transformation functions,
(
astropy.wcs.Wcs.all_pix2world
, etc.), its second dimension must now exactly match the number of dimensions in the transformation. [#1395] - Improved error message when incorrect arguments are passed to
WCS.wcs_world2pix
. [#1394] - Fixed a crash when trying to read WCS from FITS headers on Python 3.3 in Windows. [#1363]
- Only headers that are required as part of the WCSLIB C API are installed by the package, per request of system packagers. [#1666]
- When passing a single array to the wcs transformation functions,
(
- Misc
- Fixed crash when the
COLUMNS
environment variable is set to a non-integer value. [#1291] - Fixed a bug in
ProgressBar.map
wheremultiprocess=True
could cause it to hang on waiting for the process pool to be destroyed. [#1381] - Fixed a crash on Python 3.2 when affiliated packages try to use the
astropy.utils.data.get_pkg_data_*
functions. [#1256] - Fixed a minor path normalization issue that could occur on Windows in
astropy.utils.data.get_pkg_data_filename
. [#1444] - Fixed an annoyance where configuration items intended only for testing showed up in users' astropy.cfg files. [#1477]
- Prevented crashes in exception logging in unusual cases where no traceback is associated with the exception. [#1518]
- Fixed a crash when running the tests in unusual environments where
sys.stdout.encoding
isNone
. [#1530] - Miscellaneous documentation fixes and improvements [#1308, #1317, #1377, #1393, #1362, #1516]
- Fixed crash when the
- Astropy installation now requests setuptools >= 0.7 during build/installation
if neither distribute or setuptools >= 0.7 is already installed. In other
words, if
import setuptools
fails,ez_setup.py
is used to bootstrap the latest setuptools (rather than usingdistribute_setup.py
to bootstrap the now obsolete distribute package). [#1197] - When importing Astropy from a source checkout without having built the
extension modules first an
ImportError
is raised rather than aSystemExit
exception. [#1269]
astropy.coordinates
- Fixed the angle parser to support parsing the string "1 degree". [#1168]
astropy.cosmology
- Fixed a crash in the
comoving_volume
method on non-flat cosmologies when passing it an array of redshifts.
- Fixed a crash in the
astropy.io.ascii
- Fixed a bug that prevented saving changes to the comment symbol when writing changes to a table. [#1167]
astropy.io.fits
- Added a workaround for a bug in 64-bit OSX that could cause truncation when writing files greater than 2^32 bytes in size. [#839]
astropy.io.votable
- Fixed incorrect reading of tables containing multiple
<RESOURCE>
elements. [#1223]
- Fixed incorrect reading of tables containing multiple
astropy.table
- Fixed a bug where
Table.remove_column
andTable.rename_column
could cause a masked table to lose its masking. [#1120] - Fixed bugs where subclasses of
Table
did not preserver their class in certain operations. [#1142] - Fixed a bug where slicing a masked table did not preserve the mask. [#1187]
- Fixed a bug where
astropy.units
- Fixed a bug where the
.si
and.cgs
properties of dimensionlessQuantity
objects raised aZeroDivisionError
. [#1150] - Fixed a bug where multiple subsequent calls to the
.decompose()
method on array quantities applied a scale factor each time. [#1163]
- Fixed a bug where the
- Misc
- Fixed an installation crash that could occur sometimes on Debian/Ubuntu
and other *NIX systems where
pkg_resources
can be installed without installingsetuptools
. [#1150] - Updated the
distribute_setup.py
bootstrapper to use setuptools >= 0.7 when installing on systems that don't already have an up to date version of distribute/setuptools. [#1180] - Changed the
version.py
template so that Astropy affiliated packages can (and they should) use their owncython_version.py
andutils._compiler
modules where appropriate. This issue only pertains to affiliated package maintainers. [#1198] - Fixed a corner case where the default config file generation could crash if building with matplotlib but not Sphinx installed in a virtualenv. [#1225]
- Fixed a crash that could occur in the logging module on systems that don't have a default preferred encoding (in particular this happened in some versions of PyCharm). [#1244]
- The Astropy log now supports passing non-string objects (and calling
str()
on them by default) to the logging methods, in line with Python's standard logging API. [#1267] - Minor documentation fixes [#582, #696, #1154, #1194, #1212, #1213, #1246, #1252]
- Fixed an installation crash that could occur sometimes on Debian/Ubuntu
and other *NIX systems where
astropy.cosmology
- Added a new
Plank13
object representing the Plank 2013 results. [#895]
- Added a new
astropy.units
- Performance improvements in initialization of
Quantity
objects with a large number of elements. [#1231]
- Performance improvements in initialization of
astropy.time
- Fixed inaccurate handling of leap seconds when converting from UTC to UNIX timestamps. [#1118]
- Tightened required accuracy in many of the time conversion tests. [#1121]
- Misc
- Fixed a regression that was introduced in v0.2.2 by the fix to issue #992 that was preventing installation of Astropy affiliated packages that use Astropy's setup framework. [#1124]
astropy.io
- Fixed issues in both the
fits
andvotable
sub-packages where array byte order was not being handled consistently, leading to possible crashes especially on big-endian systems. [#1003]
- Fixed issues in both the
astropy.io.fits
- When an error occurs opening a file in fitsdiff the exception message will now at least mention which file had the error.
- Fixed a couple cases where creating a new table using TDIMn in some of the columns could cause a crash.
- Slightly refactored how tables containing variable-length array columns are
handled to add two improvements: Fixes an issue where accessing the data
after a call to the
astropy.io.fits.getdata
convenience function caused an exception, and allows the VLA data to be read from an existing mmap of the FITS file. - Fixed a bug on Python 3 where attempting to open a non-existent file on Python 3 caused a seemingly unrelated traceback.
- Fixed an issue in the tests that caused some tests to fail if Astropy is installed with read-only permissions.
- Fixed a bug where instantiating a
BinTableHDU
from a numpy array containing boolean fields converted all the values toFalse
. - Fixed an issue where passing an array of integers into the constructor of
Column()
when the column type is floats of the same byte width caused the column array to become garbled. - Fixed inconsistent behavior in creating CONTINUE cards from byte strings versus unicode strings in Python 2--CONTINUE cards can now be created properly from unicode strings (so long as they are convertable to ASCII).
- Fixed a bug in parsing HIERARCH keywords that do not have a space after the first equals sign (before the value).
- Prevented extra leading whitespace on HIERARCH keywords from being treated as part of the keyword.
- Fixed a bug where HIERARCH keywords containing lower-case letters was
mistakenly marked as invalid during header validation along with an
ancillary issue where the
Header.index()
method id not work correctly with HIERARCH keywords containing lower-case letters. - Disallowed assigning NaN and Inf floating point values as header values, since the FITS standard does not define a way to represent them in. Because this is undefined, the previous behavior did not make sense and produced invalid FITS files. [#954]
- Fixed an obscure issue that can occur on systems that don't have flush to memory-mapped files implemented (namely GNU Hurd). [#968]
astropy.io.votable
- Stopped deprecation warnings from the
astropy.io.votable
package that could occur during setup. [#970] - Fixed an issue where INFO elements were being incorrectly dropped when occurring inside a TABLE element. [#1000]
- Fixed obscure test failures on MIPS platforms. [#1010]
- Stopped deprecation warnings from the
astropy.nddata.convolution
- Fixed an issue in
make_kernel()
when using an Airy function kernel. Also removed the superfluous 'brickwall' option. [#939]
- Fixed an issue in
astropy.table
- Fixed a crash that could occur when adding a row to an empty (rowless) table with masked columns. [#973]
- Made it possible to assign to one table row from the value of another row, effectively making it easier to copy rows, for example. [#1019]
astropy.time
- Added appropriate
__copy__
and__deepcopy__
behavior; this omission caused a seemingly unrelated error in FK5 coordinate separation. [#891]
- Added appropriate
astropy.units
- Fixed an issue where the
isiterable()
utility returnedTrue
for quantities with scalar values. Added an__iter__
method for theQuantity
class and fixedisiterable()
to catch false positives. [#878] - Fixed previously undefined behavior when multiplying a unit by a string. [#949]
- Added 'time' as a physical type--this was a simple omission. [#959]
- Fixed issues with pickling unit objects so as to play nicer with the multiprocessing module. [#974]
- Made it more difficult to accidentally override existing units with a new unit of the same name. [#1070]
- Added several more physical types and units that were previously omitted, including 'mass density', 'specific volume', 'molar volume', 'momentum', 'angular momentum', 'angular speed', 'angular acceleration', 'electric current', 'electric current density', 'electric field strength', 'electric flux density', 'electric charge density', 'permittivity', 'electromagnetic field strength', 'radiant intensity', 'data quantity', 'bandwidth'; and 'knots', 'nautical miles', 'becquerels', and 'curies' respectively. [#1072]
- Fixed an issue where the
- Misc
- Fixed a permission error that could occur when running
astropy.test()
on Python 3 when Astropy is installed as root. [#811] - Made it easier to filter warnings from the
convolve()
function and fromQuantity
objects. [#853] - Fixed a crash that could occur in Python 3 when generation of the default config file fails during setup. [#952]
- Fixed an unrelated error message that could occur when trying to import astropy from a source checkout without having build the extension modules first. This issue was claimed to be fixed in v0.2.1, but the fix itself had a bug. [#971]
- Fixed a crash that could occur when running the
build_sphinx
setup command in Python 3. [#977] - Added a more helpful error message when trying to run the
setup.py build_sphinx
command when Sphinx is not installed. [#1027] - Minor documentation fixes and restructuring. [#935, #967, #978, #1004, #1028, #1047]
- Fixed a permission error that could occur when running
- Some performance improvements to the
astropy.units
package, in particular improving the time it takes to import the sub-package. [#1015]
astropy.coordinates
- Fixed encoding errors that could occur when formatting coordinate objects
in code using
from __future__ import unicode_literals
. [#817] - Fixed a bug where the minus sign was dropped when string formatting dms coordinates with -0 degrees. [#875]
- Fixed encoding errors that could occur when formatting coordinate objects
in code using
astropy.io.fits
- Properly supports the ZQUANTIZ keyword used to support quantization level--this includes working support for lossless GZIP compression of images.
- Fixed support for opening gzipped FITS files in a writeable mode. [#256]
- Added a more helpful exception message when trying to read invalid values
from a table when the required
TNULLn
keyword is missing. [#309] - More refactoring of the tile compression handling to work around a potential memory access violation that was particularly prevalent on Windows. [#507]
- Fixed an integer size mismatch in the compression module that could affect 32-bit systems. [#786]
- Fixed malformatting of the
TFORMn
keywords when writing compressed image tables (they omitted the max array length parameter from the variable-length array format). - Fixed a crash that could occur when writing a table containing multi- dimensional array columns from an existing file into a new file.
- Fixed a bug in fitsdiff that reported two header keywords containing NaN as having different values.
astropy.io.votable
- Fixed links to the
astropy.io.votable
documentation in the VOTable validator output. [#806] - When reading VOTables containing integers that are out of range for their column type, display a warning rather than raising an exception. [#825]
- Changed the default string format for floating point values for better round-tripping. [#856]
- Fixed opening VOTables through the
Table.read()
interface for tables that have no names. [#927] - Fixed creation of VOTables from an Astropy table that does not have a data mask. [#928]
- Minor documentation fixes. [#932]
- Fixed links to the
astropy.nddata.convolution
- Added better handling of
inf
values to theconvolve_fft
family of functions. [#893]
- Added better handling of
astropy.table
- Fixed silent failure to assign values to a row on multiple columns. [#764]
- Fixed various buggy behavior when viewing a table after sorting by one of its columns. [#829]
- Fixed using
numpy.where()
with table indexing. [#838] - Fixed a bug where opening a remote table with
Table.read()
could cause the entire table to be downloaded twice. [#845] - Fixed a bug where
MaskedColumn
no longer worked if the column being masked is renamed. [#916]
astropy.units
- Added missing capability for array
Quantity
s to be initializable by a list ofQuantity
s. [#835] - Fixed the definition of year and lightyear to be in terms of Julian year per the IAU definition. [#861]
- "degree" was removed from the list of SI base units. [#863]
- Added missing capability for array
astropy.wcs
- Fixed
TypeError
when callingWCS.to_header_string()
. [#822] - Added new method
WCS.all_world2pix
for converting from world coordinates to pixel space, including inversion of the astrometric distortion correction. [#1066, #1281]
- Fixed
- Misc
- Fixed a minor issue when installing with
./setup.py develop
on a fresh git clone. This is likely only of interest to developers on Astropy. [#725] - Fixes a crash with
ImportError: No module named 'astropy.version'
when running setup.py from a source checkout for the first time on OSX with Python 3.3. [#820] - Fixed an installation issue where running
./setup.py install
or when installing with pip the.astropy
directory gets created in the home directory of the user running the command. The user's.astropy
directory should only be created when they use Astropy, not when they install it. [#867] - Fixed an exception when creating a
ProgressBar
with a "total" of 0. [#752] - Added better documentation of behavior that can occur when trying to import the astropy package from within a source checkout without first building the extension modules. [#795, #864]
- Added link to the installation instructions in the README. [#797]
- Catches segfaults in xmllint which can occur sometimes and is otherwise out of our control. [#803]
- Minor changes to the documentation template. [#805]
- Fixed a minor exception handling bug in
download_file()
. [#808] - Added cleanup of any temporary files if an error occurs in
download_file()
. [#857] - Filesystem free space is checked for before attempting to download a file
with
download_file()
. [#858] - Fixed package data locating to work across symlinks--required to work with some OS packaging layouts. [#827]
- Fixed a bug when building Cython extensions where hidden files containing
.pyx
extensions could cause the build to crash. This can be an issue with software and filesystems that autogenerate hidden files. [#834] - Fixed bug that could cause a "script" called README.rst to be installed in a bin directory. [#852]
- Fixed some miscellaneous and mostly rare reference leaks caught by cpychecker. [#914]
- Fixed a minor issue when installing with
- Added logo and branding for Windows binary installers. [#741]
- Upgraded included version libexpat to 2.1.0. [#781]
- ~25% performance improvement in unit composition/decomposition. [#836]
- Added previously missing LaTeX formatting for
L_sun
andR_sun
. [#841] ConfigurationItem
s now have a more useful and informative__repr__
and improved documentation for how to use them. [#855]- Added a friendlier error message when trying to import astropy from a source checkout without first building the extension modules inplace. [#864]
py.test
now outputs more system information for help in debugging issues from users. [#869]- Added unit definitions "mas" and "uas" for "milliarcsecond" and "microarcsecond" respectively. [#892]
This is a brief overview of the new features included in Astropy 0.2--please see the "What's New" section of the documentation for more details.
astropy.coordinates
- This new subpackage contains a representation of celestial coordinates, and provides a wide range of related functionality. While fully-functional, it is a work in progress and parts of the API may change in subsequent releases.
astropy.cosmology
- Update to include cosmologies with variable dark energy equations of state. (This introduces some API incompatibilities with the older Cosmology objects).
- Added parameters for relativistic species (photons, neutrinos) to the astropy.cosmology classes. The current treatment assumes that neutrinos are massless. [#365]
- Add a WMAP9 object using the final (9-year) WMAP parameters from Hinshaw et al. 2013. It has also been made the default cosmology. [#629, #724]
astropy.table
I/O infrastructure for custom readers/writers implemented. [#305]- Added support for reading/writing HDF5 files [#461]
- Added support for masked tables with missing or invalid data [#451]
New
astropy.time
sub-package. [#332]New
astropy.units
sub-package that includes a class for units (astropy.units.Unit
) and scalar quantities that have units (astropy.units.Quantity
). [#370, #445]This has the following effects on other sub-packages:
- In
astropy.wcs
, thewcs.cunit
list now takes and returnsastropy.units.Unit
objects. [#379] - In
astropy.nddata
, units are now stored asastropy.units.Unit
objects. [#382] - In
astropy.table
, units on columns are now stored asastropy.units.Unit
objects. [#380] - In
astropy.constants
, constants are now stored asastropy.units.Quantity
objects. [#529]
- In
astropy.io.ascii
- Improved integration with the
astropy.table
Table class so that table and column metadata (e.g. keywords, units, description, formatting) are directly available in the output table object. The CDS, DAOphot, and IPAC format readers now provide this type of integrated metadata. - Changed to using astropy.table masked tables instead of NumPy masked arrays for tables with missing values.
- Added SExtractor table reader to
astropy.io.ascii
[#420] - Removed the Memory reader class which was used to convert data input
passed to the
write
function into an internal table. Insteadwrite
instantiates an astropy Table object using the data input towrite
. - Removed the NumpyOutputter as the output of reading a table is now
always a
Table
object. - Removed the option of supplying a function as a column output formatter.
- Added a new
strip_whitespace
keyword argument to thewrite
function. This controls whether whitespace is stripped from the left and right sides of table elements before writing. Default is True. - Fixed a bug in reading IPAC tables with null values.
- Improved integration with the
Generalized I/O infrastructure so that
astropy.nddata
can also have custom readers/writers [#659]astropy.wcs
- From updating the the underlying wcslib 4.16:
- When
astropy.wcs.WCS
constructs a default coordinate representation it will give it the special name "DEFAULTS", and will not report "Found one coordinate representation".
- When
- From updating the the underlying wcslib 4.16:
- A configuration file with all options set to their defaults is now generated
when astropy is installed. This file will be pulled in as the users'
astropy configuration file the first time they
import astropy
. [#498] - Astropy doc themes moved into
astropy.sphinx
to allow affiliated packages to access them. - Added expanded documentation for the
astropy.cosmology
sub-package. [#272] - Added option to disable building of "legacy" packages (pyfits, vo, etc.).
- The value of the astronomical unit (au) has been updated to that adopted by IAU 2012 Resolution B2, and the values of the pc and kpc constants have been updated to reflect this. [#368]
- Added links to the documentation pages to directly edit the documentation on GitHub. [#347]
- Several updates merged from
pywcs
intoastropy.wcs
[#384]:- Improved the reading of distortion images.
- Added a new option to choose whether or not to write SIP coefficients.
- Uses the
relax
option by default so that non-standard keywords are allowed. [#585]
- Added HTML representation of tables in IPython notebook [#409]
- Rewrote CFITSIO-based backend for handling tile compression of FITS files. It now uses a standard CFITSIO instead of heavily modified pieces of CFITSIO as before. Astropy ships with its own copy of CFITSIO v3.30, but system packagers may choose instead to strip this out in favor of a system-installed version of CFITSIO. This corresponds to PyFITS ticket 169. [#318]
- Moved
astropy.config.data
toastropy.utils.data
and re-factored the I/O routines to separate out the generic I/O code that can be used to open any file or resource from the code used to access Astropy-related data. The 'core' I/O routine is nowget_readable_fileobj
, which can be used to access any local as well as remote data, supports caching, and can decompress gzip and bzip2 files on-the-fly. [#425] - Added a classmethod to
astropy.coordinates.coordsystems.SphericalCoordinatesBase
that performs a name resolve query using Sesame to retrieve coordinates for the requested object. This works for any subclass ofSphericalCoordinatesBase
, but requires an internet connection. [#556] astropy.nddata.convolution
removed requirement of PyFFTW3; uses Numpy's FFT by default instead with the added ability to specify an FFT implementation to use. [#660]
astropy.io.ascii
- Fixed crash when pprinting a row with INDEF values. [#511]
- Fixed failure when reading DAOphot files with empty keyword values. [#666]
astropy.io.fits
- Improved handling of scaled images and pseudo-unsigned integer images in
compressed image HDUs. They now work more transparently like normal image
HDUs with support for the
do_not_scale_image_data
anduint
options, as well asscale_back
andsave_backup
. The.scale()
method works better too. Corresponds to PyFITS ticket 88. - Permits non-string values for the EXTNAME keyword when reading in a file, rather than throwing an exception due to the malformatting. Added verification for the format of the EXTNAME keyword when writing. Corresponds to PyFITS ticket 96.
- Added support for EXTNAME and EXTVER in PRIMARY HDUs. That is, if EXTNAME
is specified in the header, it will also be reflected in the
.name
attribute and infits.info()
. These keywords used to be verboten in PRIMARY HDUs, but the latest version of the FITS standard allows them. Corresponds to PyFITS ticket 151. - HCOMPRESS can again be used to compress data cubes (and higher-dimensional arrays) so long as the tile size is effectively 2-dimensional. In fact, compatible tile sizes will automatically be used even if they're not explicitly specified. Corresponds to PyFITS ticket 171.
- Fixed a bug that could cause a deadlock in the filesystem on OSX when reading the data from certain types of FITS files. This only occurred when used in conjunction with Numpy 1.7. [#369]
- Added support for the optional
endcard
parameter in theHeader.fromtextfile()
andHeader.totextfile()
methods. Althoughendcard=False
was a reasonable default assumption, there are still text dumps of FITS headers that include the END card, so this should have been more flexible. Corresponds to PyFITS ticket 176. - Fixed a crash when running fitsdiff on two empty (that is, zero row) tables. Corresponds to PyFITS ticket 178.
- Fixed an issue where opening a FITS file containing a random group HDU in update mode could result in an unnecessary rewriting of the file even if no changes were made. This corresponds to PyFITS ticket 179.
- Fixed a crash when generating diff reports from diffs using the
ignore_comments
options. Corresponds to PyFITS ticket 181. - Fixed some bugs with WCS Paper IV record-valued keyword cards:
- Cards that looked kind of like RVKCs but were not intended to be were over-permissively treated as such--commentary keywords like COMMENT and HISTORY were particularly affected. Corresponds to PyFITS ticket 183.
- Looking up a card in a header by its standard FITS keyword only should always return the raw value of that card. That way cards containing values that happen to valid RVKCs but were not intended to be will still be treated like normal cards. Corresponds to PyFITS ticket 184.
- Looking up a RVKC in a header with only part of the field-specifier (for example "DP1.AXIS" instead of "DP1.AXIS.1") was implicitly treated as a wildcard lookup. Corresponds to PyFITS ticket 184.
- Fixed a crash when diffing two FITS files where at least one contains a compressed image HDU which was not recognized as an image instead of a table. Corresponds to PyFITS ticket 187.
- Fixed a bug where opening a file containing compressed image HDUs in 'update' mode and then immediately closing it without making any changes caused the file to be rewritten unnecessarily.
- Fixed two memory leaks that could occur when writing compressed image data, or in some cases when opening files containing compressed image HDUs in 'update' mode.
- Fixed a bug where
ImageHDU.scale(option='old')
wasn't working at all--it was not restoring the image to its original BSCALE and BZERO values. - Fixed a bug when writing out files containing zero-width table columns, where the TFIELDS keyword would be updated incorrectly, leaving the table largely unreadable.
- Fixed a minor string formatting issue.
- Fixed bugs in the backwards compatibility layer for the
CardList.index
andCardList.count
methods. Corresponds to PyFITS ticket 190. - Improved
__repr__
and text file representation of cards with long values that are split into CONTINUE cards. Corresponds to PyFITS ticket 193. - Fixed a crash when trying to assign a long (> 72 character) value to blank ('') keywords. This also changed how blank keywords are represented--there are still exactly 8 spaces before any commentary content can begin; this may affect the exact display of header cards that assumed there could be fewer spaces in a blank keyword card before the content begins. However, the current approach is more in line with the requirements of the FITS standard. Corresponds to PyFITS ticket 194.
- Improved handling of scaled images and pseudo-unsigned integer images in
compressed image HDUs. They now work more transparently like normal image
HDUs with support for the
astropy.io.votable
- The
Table
class now maintains a single array object which is a Numpy masked array. For variable-length columns, the object that is stored there is also a Numpy masked array. - Changed the
pedantic
configuration option to beFalse
by default due to the vast proliferation of non-compliant VO Tables. [#296] - Renamed
astropy.io.vo
toastropy.io.votable
.
- The
astropy.table
- Added a workaround for an upstream bug in Numpy 1.6.2 that could cause a maximum recursion depth RuntimeError when printing table rows. [#341]
astropy.wcs
- Updated to wcslib 4.15 [#418]
- Fixed a problem with handling FITS headers on locales that do not use dot as a decimal separator. This required an upstream fix to wcslib which is included in wcslib 4.14. [#313]
- Fixed some tests that could fail due to missing/incorrect logging configuration--ensures that tests don't have any impact on the default log location or contents. [#291]
- Various minor documentation fixes [#293 and others]
- Fixed a bug where running the tests with the
py.test
command still tried to replace the system-installed pytest with the one bundled with Astropy. [#454] - Improved multiprocessing compatibility for file downloads. [#615]
- Fixed handling of Cython modules when building from a source checkout of a tagged release version. [#594]
- Added a workaround for a bug in Sphinx that could occur when using the
:tocdepth:
directive. [#595] - Minor VOTable fixes [#596]
- Fixed how
setup.py
usesdistribute_setup.py
to prevent possibleVersionConflict
errors when an older version of distribute is already installed on the user's system. [#616][#640] - Changed use of
log.warn
in the logging module tolog.warning
since the former is deprecated. [#624]
- Initial release.