BUG: on .to_string(index=False) #25000

charlesdong1991 · 2019-01-29T13:37:31Z

closes Bug on .to_string(index=False) #24980
tests added / passed
passes git diff upstream/master -u -- "*.py" | flake8 --diff
whatsnew entry

pep8speaks · 2019-01-29T13:37:40Z

Hello @charlesdong1991! Thanks for updating this PR. We checked the lines you've touched for PEP 8 issues, and found:

There are currently no PEP 8 issues detected in this Pull Request. Cheers! 🍻

Comment last updated at 2019-06-12 19:48:17 UTC

TomAugspurger · 2019-01-29T14:16:39Z

Thanks. Can you add a release note to 0.24.1 under fixed regressions?

codecov · 2019-01-29T14:24:30Z

Codecov Report

Merging #25000 into master will decrease coverage by 49.49%.
The diff coverage is 66.66%.

@@            Coverage Diff             @@
##           master   #25000      +/-   ##
==========================================
- Coverage   92.38%   42.88%   -49.5%     
==========================================
  Files         166      166              
  Lines       52400    52402       +2     
==========================================
- Hits        48409    22473   -25936     
- Misses       3991    29929   +25938

Flag	Coverage Δ
#multiple	`?`
#single	`42.88% <66.66%> (ø)`	⬆️

Impacted Files	Coverage Δ
pandas/io/formats/format.py	`50.76% <66.66%> (-47.23%)`	⬇️
pandas/io/formats/latex.py	`0% <0%> (-100%)`	⬇️
pandas/core/categorical.py	`0% <0%> (-100%)`	⬇️
pandas/io/sas/sas_constants.py	`0% <0%> (-100%)`	⬇️
pandas/tseries/plotting.py	`0% <0%> (-100%)`	⬇️
pandas/tseries/converter.py	`0% <0%> (-100%)`	⬇️
pandas/io/formats/html.py	`0% <0%> (-99.35%)`	⬇️
pandas/core/groupby/categorical.py	`0% <0%> (-95.46%)`	⬇️
pandas/io/sas/sas7bdat.py	`0% <0%> (-91.17%)`	⬇️
pandas/io/sas/sas_xport.py	`0% <0%> (-90.15%)`	⬇️
... and 124 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 6449bc1...9871550. Read the comment docs.

codecov · 2019-01-29T14:24:30Z

Codecov Report

Merging #25000 into master will increase coverage by 0.02%.
The diff coverage is 94.73%.

@@            Coverage Diff             @@
##           master   #25000      +/-   ##
==========================================
+ Coverage   91.86%   91.89%   +0.02%     
==========================================
  Files         179      175       -4     
  Lines       50707    52508    +1801     
==========================================
+ Hits        46583    48250    +1667     
- Misses       4124     4258     +134

Flag	Coverage Δ
#multiple	`90.44% <94.73%> (-0.01%)`	⬇️
#single	`40.72% <15.78%> (-0.47%)`	⬇️

Impacted Files	Coverage Δ
pandas/core/generic.py	`93.54% <ø> (-0.66%)`	⬇️
pandas/io/formats/format.py	`97.81% <94.73%> (-0.1%)`	⬇️
pandas/io/gbq.py	`75% <0%> (-25%)`	⬇️
pandas/plotting/_misc.py	`38.46% <0%> (-21.04%)`	⬇️
pandas/io/gcs.py	`80% <0%> (-20%)`	⬇️
pandas/io/clipboard/clipboards.py	`18.51% <0%> (-16.27%)`	⬇️
pandas/io/clipboard/__init__.py	`39.21% <0%> (-15.14%)`	⬇️
pandas/compat/__init__.py	`77.5% <0%> (-14.5%)`	⬇️
pandas/io/s3.py	`89.47% <0%> (-10.53%)`	⬇️
pandas/io/excel/_xlrd.py	`93.93% <0%> (-6.07%)`	⬇️
... and 121 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 634577e...5a6b867. Read the comment docs.

charlesdong1991 · 2019-01-29T14:31:20Z

added!

TomAugspurger

I notice that this affects DataFrame.to_string as well.

0.23.4

In [7]: pd.DataFrame({"A": [1]}).to_string(index=False)
Out[7]: 'A\n1'

0.24.0

In [2]: pd.DataFrame({"A": [1]}).to_string(index=False)
Out[2]: ' A\n 1'

can you add a test and fix for that as well?

TomAugspurger · 2019-01-29T15:20:20Z

doc/source/whatsnew/v0.24.1.rst

@@ -22,6 +22,7 @@ Fixed Regressions

 - Bug in :meth:`DataFrame.itertuples` with ``records`` orient raising an ``AttributeError`` when the ``DataFrame`` contained more than 255 columns (:issue:`24939`)
 - Bug in :meth:`DataFrame.itertuples` orient converting integer column names to strings prepended with an underscore (:issue:`24940`)
+- Bug in :meth:`to_string` with ``index`` set to ``False``, leading space is added. (:issue:`24980`)


Suggested change

- Bug in :meth:`to_string` with ``index`` set to ``False``, leading space is added. (:issue:`24980`)

- Bug in :meth:`Series.to_string` adding a leading space when ``index=False`` (:issue:`24980`)

mroeschke · 2019-01-29T16:25:49Z

pandas/io/formats/format.py

+            return format_array(values_to_format, None,
+                                float_format=self.float_format,
+                                na_rep=self.na_rep,
+                                leading_space=False)


Just leading_space=self.index then you don't need to duplicate the call?

hmm... not sure about this... self.index by default is True, not None, right?

You can do leading_space=bool(self.index) if you want to not allow passing None.

charlesdong1991 · 2019-01-29T16:56:55Z

hmm... i am doubting if this solution is correct... I just find out that for numeric value, this is still incorrect:
pd.Series(1).to_string(index=False), the output still contain leading space
but if it's a mixed array, then leading space goes away, e.g. pd.Series([1, "a"]).to_string(index=False).
The same behaviour remains in DataFrame, only works if column value is not numeric...
e.g. pd.DataFrame({"A": ["a", "b"], "B": [1, "dd"]}).to_string(index=False) works fine
but pd.DataFrame({"A": ["a", "b"], "B": [1, 2]}).to_string(index=False) has leading space...
any thoughts? @TomAugspurger

TomAugspurger · 2019-01-29T17:08:05Z

Indeed, the changes here are for GenericArrayFormater. We would maybe need something similar for the other formatters...

I'm re-reading #22505 now...

TomAugspurger · 2019-01-29T17:10:02Z

@gshiba can you confirm that Series.to_string is supposed to have a leading space, even when index=False?

TomAugspurger · 2019-01-29T17:14:22Z

This was apparently to fix an alignment bug with a mixture of positive / negative values

# 0.23.4
In [3]: print(pd.Series([0, -1]).to_string(index=False))
0
-1

# 0.24.0
In [2]: print(pd.Series([0, -1]).to_string(index=False))
 0
-1

I suppose that's fine, but the change to non-negative series surprises

In [3]: print(pd.Series([0, 1]).to_string(index=False))
 0
 1

Can we apply the leading space just when necessary (any values less than 0?)

charlesdong1991 · 2019-01-29T19:14:42Z

@TomAugspurger btw, i tried to fix it while keeping the original as much as possible, probably it's not very efficient, but looks working... look forward to your feedback and suggestion!
FYI, i haven't updated whatsnew since i am not sure if my solution is correct way to go, will update later once this is settled.
now, the output looks like:

# 0.24.0
In [1]: print(pd.Series([0, -1]).to_string(index=False))
 0
-1

In [2]: print(pd.Series([0, 1]).to_string(index=False))
0
1

gshiba · 2019-01-29T21:31:16Z

It seems there has been many PRs and Issues for this that it's hard to summarize succinctly, but here goes my understanding.

(My investigation only concerned numerical columns -- integers specifically; I did not look at string columns)

Old history

The leading space originates from the % d in IntArrayFormatter.get_result (106fe99; this was in year 2012. it is now in IntArrayFormatter._format_strings).

So as far as I can tell, the following behavior is intended:

there is a leading space for positive values in Int columns
there is no leading space (but a negative sign instead) for negative values in Int columns

And thus, when

an Int column is the first column in a DataFrame (or we have a Integer Series) and
it only contains zero or positive values and
index=False,

it looks like there's a redundant white space.

More recent history

In an attempt to fix the redundant white space, PR #11942 was merged, which in turn broke alignment with column headers, depending on the max width of values vs width of column name, etc (as described in Issue #13032)

I had two PRs going:

PR Fix DataFrame.to_string() justification #22437 attempted to ~~revert~~ change % d to %d as suggested here: Justification is broken with to_string(index=False) #13032 (comment) That turned out to affect a lot of tests, and was rejected in favor of PR Fix DataFrame.to_string() justification (2) #22505.
PR Fix DataFrame.to_string() justification (2) #22505 reverted parts of PR Leading space in to_string(index=False), #11833 #11942 and embraced the leading space even when index=False. With this PR, df.to_string(index=False) will print the leading space when the first column is positive only, as well as fix the Issue Justification is broken with to_string(index=False) #13032 by preserving leading/trailing spaces on first/last lines.

PR #22505 was merged in 30b942a (and released in 0.24.0), and alignment looks good now, but the leading white space is still there.

Going forward

I understand leading white spaces are unexpected when only positive values are in a Series.

If this is going to be fixed, care should be taken not to break alignment among the values, and also with the column header which could be wider/shorter than the values for DataFrames.

(I hope this doesn't add to the confusion!)

charlesdong1991 · 2019-01-29T21:45:21Z

very detailed and thorough explanation, awesome!!! @gshiba

jreback · 2019-01-30T12:36:10Z

pandas/io/formats/format.py

+        if self.index:
+            leading_space = None
+        else:
+            leading_space = False


is this really the idiom? These are both False values, so this is extra confusing.

shouldn't this just be leading_space = self.index?

somehow it looks like there is a slight difference between False and None, and if default for leading_space is None which means it's implicit, then a leading_space is added.

if leading_space is False: # False specifically, so that the default is # to include a space if we get here. tpl = u'{v}' else: tpl = u' {v}'

not sure who added this. It is just ripe for errors. Are there really 3 cases or just 2? If there are 3, then the api for leading_space should be a string that is checked and much more readable.

should have 2 cases... leading_space=None basically is equivalent to True IIUC

leading_space=None preserves the 0.23.4 and earlier behavior "add a leading space when there are any floating point and no missing values".

then let's make this more explicit and use a str argname then like leading_space='compat' or somesuch

@charlesdong1991 can you fix this up

TomAugspurger

Looking back at the PR, the relevant change seems to be right at https://github.com/pandas-dev/pandas/pull/22505/files#diff-425b4da47d01dc33d86c5c697e196b70L291.

So, how about ... something like

result = self.adj.adjoin(3, fmt_values)

if all(x[0].startswith(' ') for x in fmt_values):
    result = result..replace('\n ', '').strip()

There's an edge-case there on a series where all the strings start with a ' '...

TomAugspurger · 2019-01-31T20:22:26Z

pandas/core/generic.py

-        '\\begin{tabular}{lll}\n\\toprule\n      name &    mask &    weapon
-        \\\\\n\\midrule\n   Raphael &     red &       sai \\\\\n Donatello &
-         purple &  bo staff \\\\\n\\bottomrule\n\\end{tabular}\n'
+        '\\begin{tabular}{lll}\n\\toprule\n     name &   mask &   weapon


These test changes are concerning. Is this reverting an API change between 0.23.4 and 0.24.0?

i think this is the leading space that is wrongly added due to this bug

TomAugspurger · 2019-01-31T20:29:17Z

pandas/io/formats/format.py

+        if self.index:
+            leading_space = None
+        else:
+            leading_space = False


leading_space=None preserves the 0.23.4 and earlier behavior "add a leading space when there are any floating point and no missing values".

TomAugspurger · 2019-01-31T20:35:36Z

Though that edge case is there in 0.23.4...

>>> print(pd.Series([' a', ' b']).to_string(index=False))
a
 b

We could exclude that erroneous strip in the case of all strings starting with a ' ' with self.series.str.startswith(' ').all(), with suitable type-checking before doing the self.series.str so that doesn't raise.

gshiba · 2019-01-31T23:40:07Z

Applying strip may have unintended consequences.

Current behavior

In [1]: import pandas as pd

In [2]: print_ = lambda s: print('\n'.join([f'^{l}$' for l in s.split('\n')]))

In [3]: print_(pd.Series(['a', ' ']).to_string(index=False))
^ a$
^  $

In [4]: pd.__version__
Out[4]: '0.24.0'

If strip is used, I think we'd end up with something like:

In [X]: print_(pd.Series(['a', ' ']).to_string(index=False))
^a$

TomAugspurger · 2019-02-01T00:42:40Z

Yeah, strip is not a good solution here.

…

On Jan 31, 2019, at 17:40, Gosuke Shibahara ***@***.***> wrote: Applying strip may have unintended consequences. Current behavior In [1]: import pandas as pd In [2]: print_ = lambda s: print('\n'.join([f'^{l}$' for l in s.split('\n')])) In [3]: print_(pd.Series(['a', ' ']).to_string(index=False)) ^ a$ ^ $ In [4]: pd.__version__ Out[4]: '0.24.0' If strip is used, I think we'd end up with something like: In [X]: print_(pd.Series(['a', ' ']).to_string(index=False)) ^a$ — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or mute the thread.

TomAugspurger · 2019-02-01T12:44:08Z

I don't see a straightforward solution to this, that isn't going to regress in other ways.

It seems to me like we need to push down some of the column formatting to the value formatting. When deciding the values in the array

[10, -10]

we need to know that the array contains negative values, to decide how the first value, 10, should be formatted.

Going to push to 0.24.2.

@charlesdong1991 feel free to keep working on this if you want, and let us know if you get stuck (though I won't be able to take a look for a week or two).

jreback

this is looking like too much is changing for a backport, but hard to tell

jreback · 2019-02-04T13:08:27Z

doc/source/whatsnew/v0.24.1.rst

@@ -22,7 +22,7 @@ Fixed Regressions

 - Bug in :meth:`DataFrame.itertuples` with ``records`` orient raising an ``AttributeError`` when the ``DataFrame`` contained more than 255 columns (:issue:`24939`)
 - Bug in :meth:`DataFrame.itertuples` orient converting integer column names to strings prepended with an underscore (:issue:`24940`)
- Fixed regression in :class:`Index.intersection` incorrectly sorting the values by default (:issue:`24959`).
+- Bug in :meth:`to_string` with ``index`` set to ``False``, leading space is added. (:issue:`24980`)


move to 0.24.2

jreback · 2019-02-04T13:09:24Z

pandas/io/formats/format.py

+        if self.index:
+            leading_space = None
+        else:
+            leading_space = False


then let's make this more explicit and use a str argname then like leading_space='compat' or somesuch

charlesdong1991 · 2019-02-11T22:13:13Z

@TomAugspurger I just added more edge cases in pytest as you mentioned above with my current fix, result looks as expected... e.g.

>>> print(pd.Series([' a', ' b']).to_string(index=False))
 a
 b

>>> print(pd.Series(['10', '-10']).to_string(index=False))
 10
-10

>>> print(pd.Series(['1', '.1']).to_string(index=False))
 1
.1

have to say this fix is not very elegant... but if you agree on this fix, I could make leading_space more explicit as @jreback suggested.

charlesdong1991 · 2019-02-15T20:13:53Z

Hi, sorry to bother you @TomAugspurger @jreback master seemed broken last a couple days, so the test didn't pass... today when i am trying to merge the master to get latest changes, and resubmit the PR, i still get an error which seems unrelated to my PR, could you give any hint where I should change in my PR to fix this issue? thank you very much!

Traceback (most recent call last):
  File "C:\Miniconda\lib\site-packages\pkg_resources\__init__.py", line 357, in get_provider
    module = sys.modules[moduleOrReq]
KeyError: 'numpy'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "setup.py", line 738, in <module>
    ext_modules=maybe_cythonize(extensions, compiler_directives=directives),
  File "setup.py", line 481, in maybe_cythonize
    numpy_incl = pkg_resources.resource_filename('numpy', 'core/include')
  File "C:\Miniconda\lib\site-packages\pkg_resources\__init__.py", line 1142, in resource_filename
    return get_provider(package_or_requirement).get_resource_filename(
  File "C:\Miniconda\lib\site-packages\pkg_resources\__init__.py", line 359, in get_provider
    __import__(moduleOrReq)
ModuleNotFoundError: No module named 'numpy'
Obtaining file:///D:/a/1/s
    Complete output from command python setup.py egg_info:
    Traceback (most recent call last):
      File "C:\Miniconda\lib\site-packages\pkg_resources\__init__.py", line 357, in get_provider
        module = sys.modules[moduleOrReq]
    KeyError: 'numpy'

jreback · 2019-02-16T17:33:31Z

pandas/io/formats/format.py

+        if self.index:
+            leading_space = None
+        else:
+            leading_space = False


@charlesdong1991 can you fix this up

jreback · 2019-02-16T17:34:10Z

doc/source/whatsnew/v0.24.2.rst

@@ -56,7 +56,7 @@ Bug Fixes
 - Bug in reading a HDF5 table-format ``DataFrame`` created in Python 2, in Python 3 (:issue:`24925`)
 - Bug in reading a JSON with ``orient='table'`` generated by :meth:`DataFrame.to_json` with ``index=False`` (:issue:`25170`)
 - Bug where float indexes could have misaligned values when printing (:issue:`25061`)
-
+- Bug in :meth:`Series.to_string` adding a leading space when ``index=False`` (:issue:`24980`)


move to 0.25.0

charlesdong1991 · 2019-02-16T19:04:20Z

I changed the default value for leading_space to 'compat'as you suggested. Not sure if it's what you mean... feel free to leave your comments and advice. thanks @jreback

jreback · 2019-03-20T02:04:19Z

if you can merge master i can have a look

charlesdong1991 · 2019-03-24T14:27:02Z

@jreback sorry for my late reply. I have been with other things recently. I just fixed conflicts and merge the latest master. pls take a look if you have time

charlesdong1991 · 2019-04-12T17:23:07Z

Hi, @jreback i fixed the conflicts and merged to master, pls take a look if you get some time, thanks very much!

jreback · 2019-06-08T20:21:40Z

can you merge master and will look again.

charlesdong1991 · 2019-06-12T20:34:35Z

Hi, @jreback i merged to master and found this type validation failure, do you have idea why/how it is caused?

WillAyd · 2019-06-12T20:48:31Z

Hmm most likely caused by the merging of #26802 - thought it was green when I merged but apparently not. Let me take a look

WillAyd · 2019-06-27T03:23:47Z

@charlesdong1991 merging master again should resolve prior issue

WillAyd · 2019-07-15T01:08:49Z

@charlesdong1991 I think this is stale so closing for now but ping if you want to pick back up

TomAugspurger added this to the 0.24.1 milestone Jan 29, 2019

TomAugspurger added the Output-Formatting __repr__ of pandas objects, to_string label Jan 29, 2019

charlesdong1991 force-pushed the issue_24980 branch from 015fade to 9871550 Compare January 29, 2019 14:24

TomAugspurger requested changes Jan 29, 2019

View reviewed changes

mroeschke reviewed Jan 29, 2019

View reviewed changes

charlesdong1991 force-pushed the issue_24980 branch from d79a39c to d21255a Compare January 30, 2019 09:48

jreback requested changes Jan 30, 2019

View reviewed changes

charlesdong1991 changed the title ~~Bug: on .to_string(index=False)~~ BUG: on .to_string(index=False) Jan 31, 2019

TomAugspurger reviewed Jan 31, 2019

View reviewed changes

TomAugspurger modified the milestones: 0.24.1, 0.24.2 Feb 1, 2019

jreback requested changes Feb 4, 2019

View reviewed changes

charlesdong1991 force-pushed the issue_24980 branch from d21255a to 89bbd3b Compare February 11, 2019 22:06

charlesdong1991 force-pushed the issue_24980 branch from 000de58 to dee05ac Compare February 15, 2019 19:48

jreback requested changes Feb 16, 2019

View reviewed changes

jreback modified the milestones: 0.24.2, 0.25.0 Feb 16, 2019

jreback reviewed Feb 16, 2019

View reviewed changes

charlesdong1991 closed this Feb 17, 2019

charlesdong1991 force-pushed the issue_24980 branch from f0dd021 to 4a20d5b Compare February 17, 2019 11:23

charlesdong1991 reopened this Feb 17, 2019

charlesdong1991 force-pushed the issue_24980 branch from 781584d to ac6f1fc Compare February 18, 2019 08:43

charlesdong1991 closed this Mar 24, 2019

charlesdong1991 force-pushed the issue_24980 branch from ac6f1fc to 6e0f9a9 Compare March 24, 2019 14:23

charlesdong1991 reopened this Mar 24, 2019

charlesdong1991 force-pushed the issue_24980 branch from f34d4e7 to 439fd18 Compare April 12, 2019 12:45

merge master

5a6b867

charlesdong1991 force-pushed the issue_24980 branch from 23adc5d to 5a6b867 Compare June 12, 2019 19:48

WillAyd closed this Jul 15, 2019

TomAugspurger mentioned this pull request Sep 19, 2019

pd.Series.to_string(index=False) behavior changed to include leading spaces #28538

Closed

charlesdong1991 mentioned this pull request Nov 17, 2019

BUG: extra leading space in to_string when index=False #29670

Closed

5 tasks

charlesdong1991 mentioned this pull request Sep 2, 2020

Bug on .to_string(index=False) #24980

Closed

	- Bug in :meth:`to_string` with ``index`` set to ``False``, leading space is added. (:issue:`24980`)
	- Bug in :meth:`Series.to_string` adding a leading space when ``index=False`` (:issue:`24980`)

BUG: on .to_string(index=False) #25000

BUG: on .to_string(index=False) #25000

Conversation

charlesdong1991 commented Jan 29, 2019

pep8speaks commented Jan 29, 2019 • edited Loading

Comment last updated at 2019-06-12 19:48:17 UTC

TomAugspurger commented Jan 29, 2019

codecov bot commented Jan 29, 2019

Codecov Report

codecov bot commented Jan 29, 2019 • edited Loading

Codecov Report

charlesdong1991 commented Jan 29, 2019

TomAugspurger 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

TomAugspurger Jan 29, 2019 • edited Loading

Choose a reason for hiding this comment

charlesdong1991 commented Jan 29, 2019

TomAugspurger commented Jan 29, 2019

TomAugspurger commented Jan 29, 2019

TomAugspurger commented Jan 29, 2019

charlesdong1991 commented Jan 29, 2019 • edited Loading

gshiba commented Jan 29, 2019 • edited Loading

Old history

More recent history

Going forward

charlesdong1991 commented Jan 29, 2019

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

TomAugspurger 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

TomAugspurger commented Jan 31, 2019 • edited Loading

gshiba commented Jan 31, 2019

TomAugspurger commented Feb 1, 2019 via email

TomAugspurger commented Feb 1, 2019

jreback left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

charlesdong1991 commented Feb 11, 2019 • edited Loading

charlesdong1991 commented Feb 15, 2019 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

charlesdong1991 commented Feb 16, 2019

jreback commented Mar 20, 2019

charlesdong1991 commented Mar 24, 2019

charlesdong1991 commented Apr 12, 2019

jreback commented Jun 8, 2019

charlesdong1991 commented Jun 12, 2019

WillAyd commented Jun 12, 2019

WillAyd commented Jun 27, 2019

WillAyd commented Jul 15, 2019

pep8speaks commented Jan 29, 2019 •

edited

Loading

codecov bot commented Jan 29, 2019 •

edited

Loading

TomAugspurger Jan 29, 2019 •

edited

Loading

charlesdong1991 commented Jan 29, 2019 •

edited

Loading

gshiba commented Jan 29, 2019 •

edited

Loading

TomAugspurger commented Jan 31, 2019 •

edited

Loading

charlesdong1991 commented Feb 11, 2019 •

edited

Loading

charlesdong1991 commented Feb 15, 2019 •

edited

Loading