Add change log for version 0.21 #145

alixdamman · 2017-03-13T12:40:19Z

Based on Git logs.

gdementen

seems to be missing two improvements:

tag multiple groups with an axis in one shot (backward incompatible)
improved LArray.plot to get nicer plots by default.
The axes are transposed compared to what they used to, because the last axis is often used for time series. Also it considers a 1D array like a single series, not N series of 1 point.

gdementen · 2017-03-13T13:12:04Z

doc/source/changes/version_0_21.rst.inc

+New features
+------------
+
+* implemented replaced_axes() method to replace one, several or all axes of an array


typo... replaced -> replace

gdementen · 2017-03-13T13:14:25Z

doc/source/changes/version_0_21.rst.inc

+
+* readme.rst + doc : add installation instructions (closes :issue:`101`).
+
+* allowed to use regex in split_axis method (closes :issue:`106`):


"regex" -> "a regular expression"?

this could also be considered a new feature...

gdementen · 2017-03-13T13:15:45Z

doc/source/changes/version_0_21.rst.inc

+         a0 |  0 |  1 |  2
+         a1 |  3 |  4 |  5
+
+* allowed multiplication (__matmul__ or @ operator) between matrices with dimension != 2 (closes :issue:`122`).


allowed matrix multiplication (@ operator) between arrays with dimension != 2?

gdementen · 2017-03-13T13:16:58Z

doc/source/changes/version_0_21.rst.inc

+
+* implemented Axis.by() which is equivalent to axis[:].by()
+
+* viewer : automatically display plots done in the viewer console in a separate window


clarify "using plot methods"? (and possibly add an example)?

"using plot methods" ? Example ?

by "using plot methods" I mean using array.plot() or array.plot.XXX() (by opposition to using Ctrl-P or the array widget context menu to create the plot)

Concerning the example, something like:

>>> a = ndtest((2, 3)) >>> a.plot()

gdementen · 2017-03-13T13:18:30Z

doc/source/changes/version_0_21.rst.inc

+* viewer: allow changing the number of displayed digits even for integer arrays as that makes sense when using
+  scientific notation (closes :issue:`100`).
+
+* viewer : make shortcuts work even when the focus is not on the array editor widget


I don't consider this a bug fixed, but an improvement, but well this is subjective :)

gdementen · 2017-03-13T13:22:16Z

doc/source/changes/version_0_21.rst.inc

+Miscellaneous improvements
+--------------------------
+
+* readme.rst + doc : add installation instructions (closes :issue:`101`).


not sure mentioning "readme.rst + doc" is worth it.

add -> added to be consistent with the rest.

gdementen · 2017-03-13T13:24:36Z

doc/source/changes/version_0_21.rst.inc

+    nat\sex | M | F
+         BE | 0 | 1
+         FO | 2 | 3
+    >>> from_string('''age,nat\\sex, M, F


Isn't a single example enough to get the point across? Unsure whether we need the 3D or 2D one.

I'll keep only for 3D

gdementen · 2017-03-13T13:26:54Z

doc/source/changes/version_0_21.rst.inc

+
+* allowed multiplication (__matmul__ or @ operator) between matrices with dimension != 2 (closes :issue:`122`).
+
+* implemented Axis.by() which is equivalent to axis[:].by()


this is a new feature, right?

gdementen · 2017-03-13T13:33:47Z

doc/source/changes/version_0_21.rst.inc

+------------
+
+* implemented replaced_axes() method to replace one, several or all axes of an array
+  (instead of using the workaround a.with_axes(a.axes.replace(...)))


probably should mention somewhere that with_axes is deprecated.

gdementen · 2017-03-13T13:36:47Z

doc/source/changes/version_0_21.rst.inc

+
+* implemented replaced_axes() method to replace one, several or all axes of an array
+  (instead of using the workaround a.with_axes(a.axes.replace(...)))
+  (closes :issue:`67`).


Previously, to replace some axes but not all, one had to use a.with_axes(a.axes.replace(...))

alixdamman · 2017-03-13T14:49:24Z

? tag multiple groups with an axis in one shot (backward incompatible) ?
Couldn't find the associated commit.

gdementen · 2017-03-13T15:06:24Z

concerning "tag multiple groups": 70cffbb Gaëtan de Menten [email protected] on 2017-01-24 at 16:50

gdementen · 2017-03-14T08:37:21Z

doc/source/changes/version_0_21.rst.inc

+     a1 |  3 |  4 |  5
+    >>> row = Axis('row', ['r0', 'r1'])
+    >>> column = Axis('column', ['c0', 'c1', 'c2'])
+    Replace one axis (second argument `new_axis` must be provided)


Wouldn't it be better with some blank lines in there like in a real doctest?

gdementen · 2017-03-14T08:40:43Z

doc/source/changes/version_0_21.rst.inc

+    row\\column | c0 | c1 | c2
+            r0 |  0 |  1 |  2
+            r1 |  3 |  4 |  5
+    >>> arr.set_axes({x.a: row, x.b: column})


I wonder if the later two syntaxes shouldn't be put behind a "equivalent syntaxes" comment so that we can remove the result and make the changelog shorter to not loose our users interest right at the first change :)

gdementen · 2017-03-14T11:26:53Z

doc/source/changes/version_0_21.rst.inc

+* added possibility to tag multiple groups with an axis in one shot (backward incompatible)
+
+    >>> year  = Axis('year', '2001 .. 2010')
+    >>> year.group('2001,2002;2003:2008;2009,2010')


I think we should deprecate .group (see #148), and the commit actually included two distinct changes:

interpret string syntax in axis[a_string], which means you can do:

>>> year = Axis('year', '2001..2010') >>> year['2001:2004'] year[2001:2004]

change the axis of multiple groups via a single axis[groups]

>>> groups = year[2001:2004], year[2008,2009] >>> groups (year[2001:2004], year[2008, 2009]) >>> x.time[groups] (x.time[2001:2004], x.time[2008, 2009])

However (and this is a bit silly that I did not make it work -- but to my defense, I did not expect you to merge that PR), what you used as an example is broken at this point, because the result is not a valid index in an array:

>>> year['2001:2004;2008;2009'] year[slice(2001, 2004, None), [2008, 2009]] >>> arr = ndrange(year) >>> arr.sum(year['2001:2004;2008,2009']) TypeError Traceback (most recent call last) ... TypeError: unhashable type: 'slice' >>> # this works though ... arr.sum(tuple(year['2001:2004;2008,2009'])) year | 2001:2004 | 2008,2009 | 6 | 15

gdementen · 2017-03-14T11:27:38Z

larray/core.py

@@ -3976,14 +3976,16 @@ def replace_axes(self, axes_to_replace=None, new_axis=None, **kwargs):
         a1 |  3 |  4 |  5
        >>> row = Axis('row', ['r0', 'r1'])
        >>> column = Axis('column', ['c0', 'c1', 'c2'])
+
+        Replace one axis (expect 2 arguments)


could you sync this line with the corresponding one in the release notes?

gdementen · 2017-03-14T14:02:08Z

doc/source/changes/version_0_21.rst.inc

+  The axes are transposed compared to what they used to, because the last axis is often used for time series.
+  Also it considers a 1D array like a single series, not N series of 1 point.
+
+* added possibility to tag multiple groups with an axis in one shot (backward incompatible)


this paragraph should be killed

gdementen · 2017-03-15T08:51:27Z

LGTM

gdementen reviewed Mar 13, 2017

View reviewed changes

alixdamman force-pushed the add_change_log_version_0_21 branch from 1a1152d to 5f62f04 Compare March 13, 2017 15:50

gdementen reviewed Mar 14, 2017

View reviewed changes

alixdamman force-pushed the add_change_log_version_0_21 branch from 5f62f04 to 72a33d6 Compare March 14, 2017 13:06

gdementen mentioned this pull request Mar 14, 2017

make axis['multiple;groups'] work #151

Open

gdementen reviewed Mar 14, 2017

View reviewed changes

gdementen mentioned this pull request Mar 14, 2017

fix xticks in plot viewer when zooming (issue #143) #144

Merged

Add change log for version 0.21

9cebd44

alixdamman force-pushed the add_change_log_version_0_21 branch from 72a33d6 to 9cebd44 Compare March 15, 2017 08:32

gdementen merged commit f0fa4aa into larray-project:master Mar 15, 2017

alixdamman deleted the add_change_log_version_0_21 branch March 15, 2017 11:12

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add change log for version 0.21 #145

Add change log for version 0.21 #145

alixdamman commented Mar 13, 2017

gdementen left a comment

gdementen Mar 13, 2017

gdementen Mar 13, 2017

gdementen Mar 13, 2017

gdementen Mar 13, 2017

gdementen Mar 13, 2017

alixdamman Mar 13, 2017

gdementen Mar 13, 2017 •

edited

Loading

gdementen Mar 13, 2017

gdementen Mar 13, 2017

gdementen Mar 13, 2017

alixdamman Mar 13, 2017

gdementen Mar 13, 2017

gdementen Mar 13, 2017

gdementen Mar 13, 2017

alixdamman commented Mar 13, 2017

gdementen commented Mar 13, 2017

gdementen Mar 14, 2017

gdementen Mar 14, 2017

gdementen Mar 14, 2017

gdementen Mar 14, 2017

gdementen Mar 14, 2017

gdementen commented Mar 15, 2017


		* readme.rst + doc : add installation instructions (closes :issue:`101`).

		* allowed to use regex in split_axis method (closes :issue:`106`):


		* implemented Axis.by() which is equivalent to axis[:].by()

		* viewer : automatically display plots done in the viewer console in a separate window


		* allowed multiplication (__matmul__ or @ operator) between matrices with dimension != 2 (closes :issue:`122`).

		* implemented Axis.by() which is equivalent to axis[:].by()

Add change log for version 0.21 #145

Add change log for version 0.21 #145

Conversation

alixdamman commented Mar 13, 2017

gdementen left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

gdementen Mar 13, 2017 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

alixdamman commented Mar 13, 2017

gdementen commented Mar 13, 2017

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

gdementen commented Mar 15, 2017

gdementen Mar 13, 2017 •

edited

Loading