Style display format #12162

TomAugspurger · 2016-01-28T03:29:49Z

Closes #11692
Closes #12134
Closes #12125

This adds a .format method to Styler for formatting the display value
(the actual text) of each scalar value.

In the processes of cleaning up the template, I close #12134 (spurious 0)
and #12125 (KeyError from using iloc improperly)

TomAugspurger · 2016-01-28T03:35:33Z

Still a WIP for now, need to add handling for subset for some cases. Wanted to get this up since it's come up a couple times.

TomAugspurger · 2016-01-31T14:56:46Z

Ok, ready for review whenever someone gets a chance. These changes should be entirely backwards compatible.

jreback · 2016-01-31T15:36:05Z

pandas/core/style.py

+
+        Returns
+        -------
+        self: Styler


I think maybe update the other doc-strings (.apply, .applymap) etc to the same Returns (e.g. self : Styler) is pretty clear

TomAugspurger · 2016-01-31T21:11:24Z

Thanks Jeff.

I fixed everything except the select_dtypes enhancement. Do that as a followup since presumably it would be for all the style methods, not just format?

jreback · 2016-01-31T21:16:50Z

did u update the docs with an example for this?

TomAugspurger · 2016-01-31T21:36:14Z

Yeah, I've uploaded a gist. I'm not sure why the diff was so badly mangled.

I looked into nbsphinx a bit and it looks promising. We'll need to tweak a few things to get it working with our site. I plan on doing that sometime.

jreback · 2016-02-06T19:52:11Z

pandas/core/style.py

+    if not callable(formatter):
+        return lambda x: formatter.format(x)
+    else:
+        return formatter


doesn't this have to be a string?

e.g. df.style.format(5) is a failure, yes? I suppose that is like .set_precision(5)?

maybe raise here if its not a callable or string?

Good call. I now raise if it's not a callable or string.

This is probably something we should do more of. Since we compute the styles lazily, it'd be best to raise early when we can.

TomAugspurger · 2016-02-08T13:20:22Z

All green here.

jreback · 2016-02-08T15:12:50Z

pandas/core/style.py

+def _maybe_wrap_formatter(formatter):
+    if not (callable(formatter) or com.is_string_like(formatter)):
+        msg = "Expected a template string or callable, got {} instead".format(
+            formatter)


ok, though I would have simply done an if/elsif/raise

if you want to fix, then go ahead and merge (use the scripts/merge-pr.py tool. note that I will occasionaly edit the generated commit message if it includes too much stuff (e.g. it by default includes everything at the top of the PR). you can do it before it pushes upstream.

Will fix and merge tonight.

TomAugspurger · 2016-02-09T02:51:22Z

I need to figure out why the doc build failed here. Haven't been able to reproduce it locally yet.

jreback · 2016-02-10T18:25:12Z

lgtm. merge when ready.

TomAugspurger · 2016-02-10T18:28:47Z

I'll get all my PRs finished up tonight. If I can't figure out why this is causing the sphinx build to fail I'll drop the changes to the notebook and do that in a followup PR.

jreback · 2016-02-12T03:14:01Z

still bombing sphinx?

TomAugspurger · 2016-02-12T12:23:39Z

Surprisingly, yes. Even though I removed my changes to the docs. I've got one more thing to try out this morning.

On Feb 11, 2016, at 21:14, Jeff Reback [email protected] wrote:

still bombing sphinx?

—
Reply to this email directly or view it on GitHub.

jorisvandenbossche · 2016-02-12T13:12:52Z

@TomAugspurger Looking at the doc build log, it gives an error with "ImportError: No module named abc" (for the from collections.abc import MutableMapping that you added). That can possibly explain also all the other problems with sphinx.
I get that same error message if I just try the import, but strange that the other tests do not fail. (a python 2 vs 3 issue?)

jorisvandenbossche · 2016-02-12T13:14:18Z

pandas/core/style.py

@@ -255,6 +272,68 @@ def _translate(self):
                    precision=precision, table_styles=table_styles,
                    caption=caption, table_attributes=self.table_attributes)

+    def format(self, formatter, subset=None):
+        '''


Can you use """ here instead of '''? Does not really matter of course, but I just like some uniformity on this :-) (and I think we mainly use the double ones for docstrings)

Closes pandas-dev#11692 Closes pandas-dev#12134 Closes pandas-dev#12125 This adds a `.format` method to Styler for formatting the display value (the actual text) of each scalar value. In the processes of cleaning up the template, I close pandas-dev#12134 (spurious 0) and pandas-dev#12125 (KeyError from using iloc improperly) cherry pick test from pandas-dev#12126 only allow str formatting for now fix tests for new spec formatter callable update notebook

TomAugspurger · 2016-02-12T13:23:59Z

Thanks Joris! Hopefully that was it. It is strange that the tests didn't fail on py27. We do skip on the 27_slow_nnet_LOCALE job, since we had a strange interaction between Jinja and matplotlib. Not sure why the other ones didn't fail.

jorisvandenbossche · 2016-02-12T13:36:17Z

Yeah, very strange .. I checked and they are certainly not skipped on some of the python 2.7 builds

TomAugspurger · 2016-02-12T20:28:44Z

@jreback passed without breaking sphinx, if you want to merge.

jreback · 2016-02-12T20:40:54Z

gr8!

thanks!

and of course can just join the travis queue :)

jreback · 2016-02-12T20:43:53Z

thanks @TomAugspurger

I realized after I pressed Y that we don't have all of the issues listed in the whatsnew. Can you add them when you have a chance.

TomAugspurger · 2016-02-12T20:51:19Z

Ahh yeah, sorry. I've made a todo item.

@TomAugspurger

Addresses #12090 (comment) by making the `Styler` behaviour match regular `to_html` behaviour. This PR is based from #12162. ##### New behaviour <img width="596" alt="screen shot 2016-02-08 at 10 35 23 am" src="https://cloud.githubusercontent.com/assets/3064019/12890011 /b183e81e-ce4f-11e5-9b9f-c021fcb33c5a.png"> cc @TomAugspurger Author: Tom Augspurger <[email protected]> Author: Henry Hammond <[email protected]> Closes #12260 from HHammond/style-remove-col-index-none and squashes the following commits: 48c2f4c [Henry Hammond] BUG: Ignore style column headers when None a15248a [Tom Augspurger] ENH: display_format for style

Also redoes some of the changes that were inadvertently revereted in pandas-dev#12260. That PR had two commits, one of which was an older version of what was eventually merged in pandas-dev#12162. The stale commit incorrectly merged was a15248a. It should have been a3c38fe. For the most part the changes were just style, but there was a python2 import error, which our tests didn't fail on because I was trying to catch a jinja ImportError, and instead caught all ImportErrors.

Also redoes some of the changes that were inadvertently revereted in #12260. That PR had two commits, one of which was an older version of what was eventually merged in #12162. The stale commit incorrectly merged was a15248a. It should have been a3c38fe. For the most part the changes were just style, but there was a python2 import error, which our tests didn't fail on because I was trying to catch a jinja ImportError, and instead caught all ImportErrors.

TomAugspurger changed the title ~~Style display format~~ WIP: Style display format Jan 28, 2016

jreback added Output-Formatting __repr__ of pandas objects, to_string IO HTML read_html, to_html, Styler.apply, Styler.applymap labels Jan 28, 2016

jreback added this to the 0.18.0 milestone Jan 28, 2016

TomAugspurger force-pushed the style-display-format branch from ea8ddc6 to 75b2570 Compare January 31, 2016 14:55

jreback reviewed Jan 31, 2016
View reviewed changes

TomAugspurger force-pushed the style-display-format branch from 75b2570 to ecf6362 Compare January 31, 2016 21:09

TomAugspurger force-pushed the style-display-format branch from ecf6362 to 849ae28 Compare February 6, 2016 03:24

jreback reviewed Feb 6, 2016
View reviewed changes

TomAugspurger changed the title ~~WIP: Style display format~~ Style display format Feb 6, 2016

TomAugspurger force-pushed the style-display-format branch 2 times, most recently from 192080b to aee0d90 Compare February 7, 2016 15:58

HHammond mentioned this pull request Feb 7, 2016

Show Index Headers on DataFrames with Style #12090

Merged

jreback reviewed Feb 8, 2016
View reviewed changes

HHammond mentioned this pull request Feb 8, 2016

BUG: Make Styler ignore None index names to match to_html behaviour #12260

Closed

TomAugspurger force-pushed the style-display-format branch from aee0d90 to b8e9d05 Compare February 9, 2016 01:40

TomAugspurger force-pushed the style-display-format branch from b8e9d05 to 60eb3ef Compare February 11, 2016 03:46

jorisvandenbossche reviewed Feb 12, 2016
View reviewed changes

TomAugspurger force-pushed the style-display-format branch 2 times, most recently from b2870a4 to 1c52491 Compare February 12, 2016 13:17

TomAugspurger force-pushed the style-display-format branch from 1c52491 to a3c38fe Compare February 12, 2016 13:19

jreback closed this in cf8b7f8 Feb 12, 2016

TomAugspurger mentioned this pull request Feb 13, 2016

DOC: whatsnews for style fixes #12317

Closed

xflr6 mentioned this pull request Feb 13, 2016

BUG: style.set_precision(0) displays spurious .0 #12137

Closed

TomAugspurger mentioned this pull request Mar 4, 2016

TST: Rework style test skip #12525

Closed

TomAugspurger mentioned this pull request Mar 4, 2016

Expose Scheduler Status dask/distributed#161

Merged

TomAugspurger deleted the style-display-format branch November 3, 2016 12:38

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Style display format #12162

Style display format #12162

TomAugspurger commented Jan 28, 2016

TomAugspurger commented Jan 28, 2016

TomAugspurger commented Jan 31, 2016

jreback Jan 31, 2016

TomAugspurger commented Jan 31, 2016

jreback commented Jan 31, 2016

TomAugspurger commented Jan 31, 2016

jreback Feb 6, 2016

TomAugspurger Feb 6, 2016

TomAugspurger commented Feb 8, 2016

jreback Feb 8, 2016

jreback Feb 8, 2016

TomAugspurger Feb 8, 2016

TomAugspurger commented Feb 9, 2016

jreback commented Feb 10, 2016

TomAugspurger commented Feb 10, 2016

jreback commented Feb 12, 2016

TomAugspurger commented Feb 12, 2016

jorisvandenbossche commented Feb 12, 2016

jorisvandenbossche Feb 12, 2016

TomAugspurger commented Feb 12, 2016

jorisvandenbossche commented Feb 12, 2016

TomAugspurger commented Feb 12, 2016

jreback commented Feb 12, 2016

jreback commented Feb 12, 2016

TomAugspurger commented Feb 12, 2016

Style display format #12162

Style display format #12162

Conversation

TomAugspurger commented Jan 28, 2016

TomAugspurger commented Jan 28, 2016

TomAugspurger commented Jan 31, 2016

jreback Jan 31, 2016

Choose a reason for hiding this comment

TomAugspurger commented Jan 31, 2016

jreback commented Jan 31, 2016

TomAugspurger commented Jan 31, 2016

jreback Feb 6, 2016

Choose a reason for hiding this comment

TomAugspurger Feb 6, 2016

Choose a reason for hiding this comment

TomAugspurger commented Feb 8, 2016

jreback Feb 8, 2016

Choose a reason for hiding this comment

jreback Feb 8, 2016

Choose a reason for hiding this comment

TomAugspurger Feb 8, 2016

Choose a reason for hiding this comment

TomAugspurger commented Feb 9, 2016

jreback commented Feb 10, 2016

TomAugspurger commented Feb 10, 2016

jreback commented Feb 12, 2016

TomAugspurger commented Feb 12, 2016

jorisvandenbossche commented Feb 12, 2016

jorisvandenbossche Feb 12, 2016

Choose a reason for hiding this comment

TomAugspurger commented Feb 12, 2016

jorisvandenbossche commented Feb 12, 2016

TomAugspurger commented Feb 12, 2016

jreback commented Feb 12, 2016

jreback commented Feb 12, 2016

TomAugspurger commented Feb 12, 2016