ENH: support Styler in ExcelFormatter #15530

jnothman · 2017-02-28T11:17:19Z

Pandas is a nice way to export data from Python to Excel. It even has an internal representation of Excel cell styling. It also has a data-driven framework for styling Pandas dataframes with CSS.

This PR connects those features together in a rudimentary and experimental way, for advanced users only. It allows ExcelFormatter to accept a Styler in place of a DataFrame and to provide a style_converter callback for conversion from CSS styles to the styles understood by ExcelWriter.

While I acknowledge that I am extending a hidden feature (to make it more usable) and that high-coverage conversion of CSS to Excel styles is no small feat so ?all prospective usages will be a hack, I would enjoy using this feature and think others would too.

I wasn't sure how to test this as no tests directly instantiate ExcelFormatter. (I could of course alter the to_excel parameters to provide the new feature, but I thought it might best remain hidden. WDYT?)

A test follows:

import pandas as pd
from pandas.formats.format import ExcelFormatter


def test_style_converter():
    df = pd.DataFrame({'a': list(range(3))})

    header_style = '!unset'
    # default Styler provided
    formatter = ExcelFormatter(df)
    for cell in formatter.get_formatted_cells():
        if cell.row < 1 or cell.col < 1:
            if header_style == '!unset':
                header_style = cell.style
                assert header_style is not None
            else:
                assert cell.style == header_style
        else:
            assert cell.style is None

    styler = df.style.background_gradient(cmap='Greys', low=.5, high=0)
    expected = ['c6c6c6', '686868', '000000']
    formatter = ExcelFormatter(styler)
    for cell in formatter.get_formatted_cells():
        if cell.row < 1 or cell.col < 1:
            assert cell.style is header_style
        else:
            assert cell.style == {'fill': {'fgColor': expected[cell.row - 1],
                                           'patternType': 'solid'}}

    styler.to_excel('/tmp/out.xlsx', engine='openpyxl')
    with open('/tmp/out.html', 'w') as f:
        f.write(styler.render())

closes Styling in DataFrame.to_excel #1663
tests added / passed
passes git diff upstream/master | flake8 --diff
whatsnew entry

jnothman · 2017-02-28T11:30:56Z

pandas/formats/format.py

@@ -1902,11 +1907,26 @@ def _format_hierarchical_rows(self):
                                        indexcolval, header_style)
                    gcolidx += 1

+        for cell in self._generate_body(coloffset=gcolidx):


And if uninterested in merging the whole patch, could you consider merging just this refactoring, so that it's easier to extend ExcelFormatter?

jreback · 2017-02-28T12:01:22Z

is this intended to address #1663?

jnothman · 2017-02-28T12:16:05Z

is this intended to address #1663?

I suppose so. And it seems I hacked a bit quickly and need to fix some test failures.

jnothman · 2017-02-28T12:16:30Z

I couldn't tell whether #1663 had anything to do with Styler objects.

jreback · 2017-02-28T12:30:56Z

also have a look at comments #7565

jnothman · 2017-02-28T12:49:46Z

also have a look at comments #7565

I think I've got the gist. Obviously this builds on that support together with Styler

codecov-io · 2017-02-28T13:31:26Z

Codecov Report

Merging #15530 into master will increase coverage by 0.03%.
The diff coverage is 97.51%.

@@            Coverage Diff             @@
##           master   #15530      +/-   ##
==========================================
+ Coverage   90.79%   90.82%   +0.03%     
==========================================
  Files         156      159       +3     
  Lines       50534    50794     +260     
==========================================
+ Hits        45883    46135     +252     
- Misses       4651     4659       +8

Flag	Coverage Δ
#multiple	`88.6% <97.51%> (+0.04%)`	⬆️
#single	`40.36% <17.6%> (-0.09%)`	⬇️

Impacted Files	Coverage Δ
pandas/io/formats/format.py	`95.02% <100%> (-0.1%)`	⬇️
pandas/io/formats/style.py	`96.36% <100%> (+0.07%)`	⬆️
pandas/core/frame.py	`97.64% <100%> (-0.02%)`	⬇️
pandas/io/formats/css.py	`100% <100%> (ø)`
pandas/io/formats/common.py	`94.44% <94.44%> (ø)`
pandas/io/formats/excel.py	`96.64% <96.64%> (ø)`
pandas/plotting/_converter.py	`63.54% <0%> (-1.82%)`	⬇️
pandas/core/series.py	`94.97% <0%> (ø)`	⬆️
... and 5 more

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update f114af0...c7a51ca. Read the comment docs.

jreback · 2017-02-28T16:11:48Z

can you show a picture of what this does? e.g. the excel picture.

chris-b1 · 2017-02-28T21:07:25Z

While I like the idea of using Styler to define Excel styles, I'm not sure about this api. I do understand that this would be intended to advanced / experimental.

I would prefer it if a default style_converter was provided, and extensible in some way. Handle simple things like background color, text color, etc by default.

jnothman · 2017-02-28T22:21:05Z

okay. I was hoping to pose this as an intermediate step, seeing as we can surely only support a subset of CSS, ideally corresponding to what is currently generated through Styler's declarative API. (There are many ways to specify border styles, colours, etc in CSS.) I am not sure of the time frame in which I can make a solid attempt at that. I see that providing a default converter has the advantage of not exposing the internal excel styling language to users. Yes, @jreback there seem to be problems with rendering. Will get back to you shortly.

…

On 1 Mar 2017 8:07 am, "chris-b1" ***@***.***> wrote: While I like the idea of using Styler to define Excel styles, I'm not sure about this api. I do understand that this would be intended to advanced / experimental. I would prefer it if a default style_converter was provided, and extensible in some way. Handle simple things like background color, text color, etc by default. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#15530 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAEz63qaoL5UHJpxwZYSryYveorRv4zHks5rhIyWgaJpZM4MORU3> .

jnothman · 2017-02-28T23:10:12Z

Sorted. I needed a patternType and to use fgColor rather than bgColor.

jnothman · 2017-02-28T23:16:18Z

I have added writing to the test in the PR description above... with paths in /tmp/ so not actually to be used in a test. Output is as follows:

@chris-b1, in the PR description I have also drafted a CSSToExcelStyle class with an architecture for a more full-fledged converter. WDYT?

jnothman · 2017-03-01T00:34:00Z

Note that openpyxl cannot support conversion of the linear-gradient used for bar chart backgrounds without https://bitbucket.org/openpyxl/openpyxl/issues/771 being fixed.

jnothman · 2017-03-01T01:09:47Z

If we go down @c123w's path, can we have a dependency on tinycss or tinycss2 when this feature is used to avoid replicating a full-fledged CSS parser here?

Also, what would be the public API? df.style.apply(...).to_excel() seems most natural.

jnothman · 2017-03-01T01:13:09Z

I do have some concern that exporting to HTML and using Excel's import may be appropriate and much simpler in many cases.

jreback · 2017-03-01T01:36:44Z

@jnothman would not be a problem to have an optional dep like that

chris-b1 · 2017-03-01T01:38:03Z

@jnothman - thanks - a couple thoughts.

Your CSSToExcelStyle mockup is basically what I was thinking - though I would probably make the handlers return a dict of style rather than mutating. My concern was that a relatively complex callback is too hard to think about (at least for me!) - something like that, which can be used as an example / extended is probably friendlier.

And to be clear, I personally would be fine initially supporting a tiny subset of CSS (even just font color and backround). Could take the dependency if needed, although I'm not sure how much it would really add in this case?

jreback · 2017-03-01T01:38:30Z

df.style.apply(...).to_excel() would be nice.

we de-facto have .to_html() now via _repr_html_

you might want to define this new style sheet as _repr_excel_

@TomAugspurger

jnothman · 2017-03-01T02:33:52Z

Thanks @jreback for the feedback on API.

@chris-b1:

though I would probably make the handlers return a dict of style rather than mutating.

I considered this. I mutate in order to take advantage of setdefault. This allows setting the patternType to 'solid' only if it was not set to something else. On the other hand, this may be construed as ExcelWriter having a bug in that it doesn't default to patternType='solid'. I would rather not have to modify ExcelWriter every time something like this is needed, so mutating is best, IMO.

And to be clear, I personally would be fine initially supporting a tiny subset of CSS (even just font color and backround). Could take the dependency if needed, although I'm not sure how much it would really add in this case?

I'm not yet certain about the need for a CSS parser, but as is the way of these things, you can implement something that works for the limited cases you've thought of, but then someone will break it, even just by specifying "red" or "hsl(0, 100%, 50%)" or "rgb(255, 0, 0)" or "rgba(255, 0, 0, 100)" instead of "#ff0000", let alone by inserting a comment or unexpected use of whitespace, etc.

I personally would be fine initially supporting a tiny subset of CSS (even just font color and backround)

Font and background colour seem a good place to start, followed by borders. Ideally we'd support the bar charting provided by Styler, but will have to wait on openpyxl for that.

jnothman · 2017-03-01T04:29:48Z

I've attempted to implement a few handlers. Tricky parsing cases include font which can be composed of identifiers, strings for font name (comma separated for a back-off strategy which won't be handled in Excel except to set font family) and numeric quantities (for size). Functions may also be present, at least in theory, and while we would ignore them, splitting on whitespace becomes insufficent. The many expressions of font would still need handling with tinycss, but at least it would identify the token types and handle string escaping.

Tricky not in terms of parsing but in terms of conversion are elements that require state to be stored that is not naturally represented in the Excel style dicts. One (not very useful) example is that text-decoration-style may indicate single, double etc, but it applies to underline, overline (not supported in Excel) or line-through (double not supported in openpyxl) depending on text-decoration-line. So we can provide limited support, or need to maintain state that is not part of the output.

jnothman · 2017-03-01T07:58:23Z

I have wasted far too much time on this, but I think the principled way to go about CSS to Excel conversion for a list of CSS declarations:

remove comments, normalize whitespace, ?tokenize and parse function-argument structures
expand CSS shorthands (e.g. font, border-width) to atomic properties
override earlier values for same property with later values
resolve values relative to parent/initial settings (does not need to be fully fledged for our application, but font-size: 1.5em needs to be interpreted)
compute excel styles with reference to calculated CSS properties (rather than by processing each CSS declaration in turn)

A full CSS interpreter would also deal with matching selectors to contexts and ordering declarations by specificity.

xhtml2pdf does some of these things in a different context, but I can't at a glance commend its implementation. I haven't found other Python implementations of this CSS resolution process.

jreback · 2017-04-03T15:15:50Z

status on this @chris-b1

chris-b1 · 2017-04-03T17:28:34Z

@jnothman - thoughts? I'm not opposed to exposing something small / experimental, like your top comment.

jnothman · 2017-04-03T22:43:40Z

I'll see if i can give this a few minutes today, but I'm generally quite time poor atm

…

On 4 Apr 2017 3:28 am, "chris-b1" ***@***.***> wrote: @jnothman <https://github.com/jnothman> - thoughts? I'm not opposed to exposing something small / experimental, like your top comment. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#15530 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAEz63fcI5va_oL5GpFZwJ5B_srjetU5ks5rsSxMgaJpZM4MORU3> .

jreback · 2017-04-19T10:07:28Z

doc/source/style.ipynb

@@ -858,6 +847,53 @@
   "cell_type": "markdown",
   "metadata": {},
   "source": [
+    "## Export to Excel\n",
+    "\n",
+    "*New in version 0.20.0*\n",


you might want to add an Experimental statement somehwere

There's one right below, isn't there?

oh i c, ok then.

jreback · 2017-04-19T13:54:00Z

@jnothman ok if you'd make those small changes. ping on green.

jnothman · 2017-04-19T14:28:09Z

There were larger changes that came out of smoke testing xlwt with test_styler_to_excel. See commit 6465913 (`and False` there is removed later in de53808). Let me know if you'd rather leave xlwt alone.

…

On 19 April 2017 at 23:54, Jeff Reback ***@***.***> wrote: @jnothman <https://github.com/jnothman> ok if you'd make those small changes. ping on green. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#15530 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAEz61PxdGhFf61clcdCmSbWlKu048wUks5rxhIEgaJpZM4MORU3> .

I had misunderstood the intention of the reviewer.

jreback · 2017-04-20T01:04:00Z

thanks, very nice patch!.

watch the dev docs: http://pandas-docs.github.io/pandas-docs-travis/ to see what the built looks like; if needed please issue a followup. (this could take an hour or so).

TomAugspurger · 2017-04-20T01:23:46Z

Thanks @jnothman!

jnothman · 2017-04-20T01:40:40Z

thanks, very nice patch!.

Oh I hope so! I kept wondering if maybe we should just export HTML and ask Google to convert it to Excel... But I'd much rather this interface, personally!

jnothman · 2017-04-20T01:40:48Z

Thanks a lot for the reviews and approvals.

jnothman · 2017-04-20T02:26:13Z

doc/source/whatsnew/v0.20.0.txt

+    import os
+    os.remove('styled.xlsx')
+
+See the :ref:`Style documentation <style>` for more detail.


This ref isn't working. @TomAugspurger, I assume this is related to the nbsphinx changes. There is another ref to style in 0.17.1

closes pandas-dev#1663 Author: Joel Nothman <joel.nothman@gmail.com> Closes pandas-dev#15530 from jnothman/excel_style and squashes the following commits: c7a51ca [Joel Nothman] Test currently fails on openpyxl1 due to version incompatibilities 836f39e [Joel Nothman] Revert changes to xlwt de53808 [Joel Nothman] Remove debug code a5d51f9 [Joel Nothman] Merge branch 'master' into excel_style 934df06 [Joel Nothman] Display df, not styled 6465913 [Joel Nothman] More pytest-like test_styler_to_excel; enhancements to xlwt 6168765 [Joel Nothman] Recommended changes to what's new 9669d7d [Joel Nothman] Require jinja in test with df.style 14035c5 [Joel Nothman] Merge branch 'master' into excel_style 3071bac [Joel Nothman] Complete tests ceb9171 [Joel Nothman] reasons for xfails e2cfa77 [Joel Nothman] Test Styler.to_excel d5db0ac [Joel Nothman] Remove obsolete TODO 0256fc6 [Joel Nothman] Return after unhandled font size warning 60d6a3b [Joel Nothman] add doc/source/styled.xlsx to the gitignore 4e72993 [Joel Nothman] Fix what's new heading d144fdf [Joel Nothman] Font name strings 61fdc69 [Joel Nothman] Complete testing basic CSS -> Excel conversions 6ff8a46 [Joel Nothman] Fix loose character; sorry 6d3ffc6 [Joel Nothman] Lint 79eae41 [Joel Nothman] Documentation tweaks c4f59c6 [Joel Nothman] Doc tweaks 2c3d015 [Joel Nothman] Fix JSON syntax in IPynb b1d774b [Joel Nothman] What's new heading 096f26c [Joel Nothman] Merge remote-tracking branch 'upstream/master' into excel_style 433be03 [Joel Nothman] Documentation 9a62699 [Joel Nothman] Fix tests and add TODOs to tests 7c54a69 [Joel Nothman] Fix test failures; avoid hair border which renders strangely 8e9a567 [Joel Nothman] Fixes from integration testing c1fc232 [Joel Nothman] Remove debugging print statements a43d6b7 [Joel Nothman] Cleaner imports a1127f6 [Joel Nothman] Merge branch 'master' into excel_style 306eebe [Joel Nothman] Module-level docstring 350eab5 [Joel Nothman] remove spurious blank line efce9b6 [Joel Nothman] More CSS to Excel testing; define ExcelFormatter.write f17a0f4 [Joel Nothman] Some border style tests 1a8818f [Joel Nothman] Lint 9a5b791 [Joel Nothman] Fix testing ImportError 1984cab [Joel Nothman] Fix making get_level_lengths non-private eb02cc1 [Joel Nothman] Fix testing ImportError 3b26087 [Joel Nothman] Make get_level_lengths non-private f62f02d [Joel Nothman] File restructure dc953d4 [Joel Nothman] Font size and border width 7db59c0 [Joel Nothman] Test inherited styles in converter d103f61 [Joel Nothman] Refactoring and initial tests for CSS to Excel 176e51c [Joel Nothman] Fix NameError c589c35 [Joel Nothman] Fix some lint errors (yes, the code needs testing) cb5cf02 [Joel Nothman] Fix bug where inherited not being passed; avoid classmethods 0ce72f9 [Joel Nothman] Use inherited font size for em_pt 8780076 [Joel Nothman] Merge branch 'master' into excel_style 96680f9 [Joel Nothman] Largely complete CSSToExcelConverter and Styler.to_excel() f1cde08 [Joel Nothman] FIX column offset incorrect in refactor ada5101 [Joel Nothman] ENH: support Styler in ExcelFormatter

jnothman · 2017-04-26T10:52:57Z

I'm interested in making a separate package of the CSS handling here, in part so that it can be extended without small changes to Pandas, and in part so that it is available to be used elsewhere. Is there a nice way to avoid synchrony issues?

jnothman · 2017-04-26T11:34:09Z

css.py and test_css.py are now living (under different names) at https://github.com/jnothman/cssdecl

jreback · 2017-04-26T13:21:32Z

what you would do is release a version of this package

then in pandas we would do an import of this package and use its api

this would have to be for 0.21.0 (which is starting soon actually)

jnothman · 2017-04-26T16:12:17Z

I don't want to call the package stable enough to import without pinned versions. Could we just keep a parallel copy but recommend against changes in Pandas?

…

On 26 April 2017 at 23:21, Jeff Reback ***@***.***> wrote: what you would do is release a version of this package then in pandas we would do an import of this package and use its api this would have to be for 0.21.0 (which is starting soon actually) — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#15530 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAEz62KTYiKnjBoyhqi84ePHFefhgOsYks5rz0TogaJpZM4MORU3> .

jreback · 2017-04-26T16:45:31Z

no that's exactly what u so
release 0.1 and we can pin in pandas for now

jnothman · 2017-04-26T16:49:37Z

Okay. There's version 0.0.0 released... https://pypi.org/project/cssdecl/#history If you'd rather 0.1 I can do that.

…

On 27 April 2017 at 02:45, Jeff Reback ***@***.***> wrote: no that's exactly what u so release 0.1 and we can pin in pandas for now — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#15530 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAEz6xwKItsO7rG740OMI6tNhnf2uArQks5rz3S2gaJpZM4MORU3> .

jnothman · 2017-04-27T05:49:56Z

I assume you don't want me to include a hard dependency on pypi's cssdecl, but make it an optional dependency like openpyxl is? If so it seems hard to require a particular version when I'm expecting backwards incompatible changes.

jreback · 2017-04-27T11:00:22Z

yes

normally u start versions > 0 iow 0.1

jnothman · 2017-04-28T00:17:04Z

Fyi, I'm having trouble with conda build: https://groups.google.com/a/continuum.io/forum/m/#!topic/conda/yKDGIMZZl2o

closes pandas-dev#1663 Author: Joel Nothman <joel.nothman@gmail.com> Closes pandas-dev#15530 from jnothman/excel_style and squashes the following commits: c7a51ca [Joel Nothman] Test currently fails on openpyxl1 due to version incompatibilities 836f39e [Joel Nothman] Revert changes to xlwt de53808 [Joel Nothman] Remove debug code a5d51f9 [Joel Nothman] Merge branch 'master' into excel_style 934df06 [Joel Nothman] Display df, not styled 6465913 [Joel Nothman] More pytest-like test_styler_to_excel; enhancements to xlwt 6168765 [Joel Nothman] Recommended changes to what's new 9669d7d [Joel Nothman] Require jinja in test with df.style 14035c5 [Joel Nothman] Merge branch 'master' into excel_style 3071bac [Joel Nothman] Complete tests ceb9171 [Joel Nothman] reasons for xfails e2cfa77 [Joel Nothman] Test Styler.to_excel d5db0ac [Joel Nothman] Remove obsolete TODO 0256fc6 [Joel Nothman] Return after unhandled font size warning 60d6a3b [Joel Nothman] add doc/source/styled.xlsx to the gitignore 4e72993 [Joel Nothman] Fix what's new heading d144fdf [Joel Nothman] Font name strings 61fdc69 [Joel Nothman] Complete testing basic CSS -> Excel conversions 6ff8a46 [Joel Nothman] Fix loose character; sorry 6d3ffc6 [Joel Nothman] Lint 79eae41 [Joel Nothman] Documentation tweaks c4f59c6 [Joel Nothman] Doc tweaks 2c3d015 [Joel Nothman] Fix JSON syntax in IPynb b1d774b [Joel Nothman] What's new heading 096f26c [Joel Nothman] Merge remote-tracking branch 'upstream/master' into excel_style 433be03 [Joel Nothman] Documentation 9a62699 [Joel Nothman] Fix tests and add TODOs to tests 7c54a69 [Joel Nothman] Fix test failures; avoid hair border which renders strangely 8e9a567 [Joel Nothman] Fixes from integration testing c1fc232 [Joel Nothman] Remove debugging print statements a43d6b7 [Joel Nothman] Cleaner imports a1127f6 [Joel Nothman] Merge branch 'master' into excel_style 306eebe [Joel Nothman] Module-level docstring 350eab5 [Joel Nothman] remove spurious blank line efce9b6 [Joel Nothman] More CSS to Excel testing; define ExcelFormatter.write f17a0f4 [Joel Nothman] Some border style tests 1a8818f [Joel Nothman] Lint 9a5b791 [Joel Nothman] Fix testing ImportError 1984cab [Joel Nothman] Fix making get_level_lengths non-private eb02cc1 [Joel Nothman] Fix testing ImportError 3b26087 [Joel Nothman] Make get_level_lengths non-private f62f02d [Joel Nothman] File restructure dc953d4 [Joel Nothman] Font size and border width 7db59c0 [Joel Nothman] Test inherited styles in converter d103f61 [Joel Nothman] Refactoring and initial tests for CSS to Excel 176e51c [Joel Nothman] Fix NameError c589c35 [Joel Nothman] Fix some lint errors (yes, the code needs testing) cb5cf02 [Joel Nothman] Fix bug where inherited not being passed; avoid classmethods 0ce72f9 [Joel Nothman] Use inherited font size for em_pt 8780076 [Joel Nothman] Merge branch 'master' into excel_style 96680f9 [Joel Nothman] Largely complete CSSToExcelConverter and Styler.to_excel() f1cde08 [Joel Nothman] FIX column offset incorrect in refactor ada5101 [Joel Nothman] ENH: support Styler in ExcelFormatter

ENH: support Styler in ExcelFormatter

ada5101

jnothman commented Feb 28, 2017

View reviewed changes

FIX column offset incorrect in refactor

f1cde08

jreback added IO Excel read_excel, to_excel Output-Formatting __repr__ of pandas objects, to_string labels Feb 28, 2017

jnothman added 2 commits April 5, 2017 12:06

Largely complete CSSToExcelConverter and Styler.to_excel()

96680f9

Merge branch 'master' into excel_style

8780076

jreback added this to the 0.20.0 milestone Apr 19, 2017

jreback reviewed Apr 19, 2017

View reviewed changes

TomAugspurger approved these changes Apr 19, 2017

View reviewed changes

jnothman added 2 commits April 20, 2017 00:25

Recommended changes to what's new

6168765

More pytest-like test_styler_to_excel; enhancements to xlwt

6465913

jnothman added 5 commits April 20, 2017 00:49

Display df, not styled

934df06

I had misunderstood the intention of the reviewer.

Merge branch 'master' into excel_style

a5d51f9

Remove debug code

de53808

Revert changes to xlwt

836f39e

Test currently fails on openpyxl1 due to version incompatibilities

c7a51ca

jreback closed this in 1b52b12 Apr 20, 2017

jnothman commented Apr 20, 2017

View reviewed changes

gordol mentioned this pull request Jan 25, 2018

excel header_style broken in 0.22 #19386

Closed

ENH: support Styler in ExcelFormatter #15530

ENH: support Styler in ExcelFormatter #15530

Conversation

jnothman commented Feb 28, 2017 • edited Loading

jnothman Feb 28, 2017

Choose a reason for hiding this comment

jreback commented Feb 28, 2017

jnothman commented Feb 28, 2017

jnothman commented Feb 28, 2017

jreback commented Feb 28, 2017

jnothman commented Feb 28, 2017

codecov-io commented Feb 28, 2017 • edited by codecov bot Loading

Codecov Report

jreback commented Feb 28, 2017

chris-b1 commented Feb 28, 2017

jnothman commented Feb 28, 2017 via email

jnothman commented Feb 28, 2017

jnothman commented Feb 28, 2017

jnothman commented Mar 1, 2017

jnothman commented Mar 1, 2017 • edited Loading

jnothman commented Mar 1, 2017

jreback commented Mar 1, 2017

chris-b1 commented Mar 1, 2017

jreback commented Mar 1, 2017 • edited Loading

jnothman commented Mar 1, 2017

jnothman commented Mar 1, 2017

jnothman commented Mar 1, 2017

jreback commented Apr 3, 2017

chris-b1 commented Apr 3, 2017

jnothman commented Apr 3, 2017 via email

jreback Apr 19, 2017

Choose a reason for hiding this comment

jnothman Apr 19, 2017

Choose a reason for hiding this comment

jnothman Apr 19, 2017

Choose a reason for hiding this comment

jreback Apr 19, 2017

Choose a reason for hiding this comment

jreback commented Apr 19, 2017

jnothman commented Apr 19, 2017 via email • edited Loading

jreback commented Apr 20, 2017

TomAugspurger commented Apr 20, 2017

jnothman commented Apr 20, 2017

jnothman commented Apr 20, 2017

jnothman Apr 20, 2017

Choose a reason for hiding this comment

jnothman commented Apr 26, 2017

jnothman commented Apr 26, 2017

jreback commented Apr 26, 2017

jnothman commented Apr 26, 2017 via email

jreback commented Apr 26, 2017

jnothman commented Apr 26, 2017 via email

jnothman commented Apr 27, 2017 • edited Loading

jreback commented Apr 27, 2017

jnothman commented Apr 28, 2017

jnothman commented Feb 28, 2017 •

edited

Loading

codecov-io commented Feb 28, 2017 •

edited by codecov bot

Loading

jnothman commented Mar 1, 2017 •

edited

Loading

jreback commented Mar 1, 2017 •

edited

Loading

jnothman commented Apr 19, 2017 via email •

edited

Loading

jnothman commented Apr 27, 2017 •

edited

Loading