2D Normalization & Colormapping for ScalerMappables #8738

patniharshit · 2017-06-09T21:12:43Z

PR Summary

This PR is in referance to GSoC project '2D colormaps'.
(Project proposal is here)

The BivariateColormap class subclasses Colormap class. After generation it is flattened to 1d look up table.
The implementation of BivariateColormap is not perfect yet and just something has been generated to show proof of concept.
The bivariate data is identified by ndim == 3 and if an instance of either Bivariate Colormap or Norm has been passed in. This data is normalized and mapped to univariate integer index so as to directly index flattened BivariateColormap.
Plotting some air temperature and surface pressure data:

Feel free to review.

@story645 @tacaswell

PR Checklist

Has Pytest style unit tests
Code is PEP 8 compliant
New features are documented, with examples if plot related
Documentation is sphinx and numpydoc compliant
Added an entry to doc/users/whats_new.rst if major new feature
Documented in doc/api/api_changes.rst if API changed in a backward-incompatible way

story645 · 2017-06-09T22:08:11Z

lib/matplotlib/colors.py

@@ -1346,6 +1346,60 @@ def __call__(self, value, clip=None):
    def inverse(self, value):
        return value

+class Norm2d:


"BivariateNorm" just so that it can follow the mpl convention of "Xnorm"?

story645 · 2017-06-09T22:10:08Z

lib/matplotlib/colors.py

+    """
+    Normalize a list of two values corresponding to two 1D normalizers
+    """
+    def __init__(self, norm_instances=None):


is there are a reason norm_instances = none and is not:

def __init__(self, norm1 = None, norm2 = None):

No reason. I think what you are proposing is more convenient from user's perspective.

story645 · 2017-06-09T22:13:09Z

lib/matplotlib/colors.py

+        norm_instances :
+            A list of length two having instances of 1D normalizers
+        """
+        if norm_instances is None:


How do you handle the use case [BoundaryNorm(), None] or [None, BoundaryNorm]?

Yes this should be handled.

This is being handled now.

anntzer · 2017-06-10T04:42:29Z

How would this relate to actual plotting functions?

patniharshit · 2017-06-10T16:40:13Z

@anntzer Currently, functions like imshow, pcolormesh, scalar mappable etc expects 1D normalizer. To extend their support for bivariate colormaps there is a need for bivariate norm also.

anntzer · 2017-06-10T21:19:13Z

I understand, but right now all I see this PR implements is some trivial way of combining to normalizers. If there's no bigger example usage, it's hard to judge whether that's sufficient, or too simple, or too complex.
For example, would it be better to just have a generic n-dimensional normalizer?

story645 · 2017-06-11T02:59:46Z

This should stay in its own PR for reviewing ease, but what about postponing the decision on whether to merge this until the whole set of 2d coloring PRs come in?

The issue with an N-color normalizer (I like it in theory/kinda agree on creating it and then sub classing to this) is that past 3 variables it's not really implementable.

anntzer · 2017-06-11T03:09:04Z

I agree with keeping the PR open for now.

Not sure why you cannot have more three dimensions, at the end we're just writing a mapping from R^n -> {colors}.

story645 · 2017-06-11T03:19:30Z

Yes, you can technically have N dimensions, but:
1 dim = colorbar
2 dim - color wheel/ color square
3 dim - color triangle
1 free dim - alph
More dims - open research problem

anntzer · 2017-06-11T03:22:53Z

Well a color triangle really is only two dimensional (the third variable is determined by the first two).
Again, as this PR does not discuss how to plot the "colorbar" at all it's hard to judge...

anntzer

Further PRs on nonscalar normalizers need to come in before we can judge whether this design makes sense.

anntzer · 2017-07-23T21:36:34Z

lib/matplotlib/axes/_axes.py


        if len(args) == 1:
            C = np.asanyarray(args[0])
+            isBivari = (isinstance(norm, mcolors.BivariateNorm) or


try to stick to pep8 (is_bivar)
also this can br written as isinstance(norm, (mcolors.BivariateNorm, mcolors.BivariateColormap))
(similarly below)

The previous change was reverted because two different variables norm and cmap are being checked for their respective instances.

anntzer · 2017-07-23T21:43:18Z

lib/matplotlib/colorbar.py

+                 xlabel='',
+                 ylabel='',
+                 ):
+        #: The axes that this colorbar lives in.


Unless there is a strong reason to expose them, attributes should be private (start with an underscore) by default. Otherwise, they all become part of the public API, which is very difficult to change while maintaining back-compatibility.
(Same comment throughout the PR.)

Any specific attribute you are pointing to? I tried to keep it similar to ColorbarBase class.

ax is probably fine to stay public (it is public on every Artist), but the xvalues, xboundaries, etc can probably be made private.

story645 · 2017-07-24T01:55:05Z

lib/matplotlib/axes/_axes.py

+        isNorm = isinstance(norm, (mcolors.Normalize, mcolors.BivariateNorm))
+        if norm is not None and not isNorm:
+            msg = "'norm' must be an instance of 'mcolors.Normalize' " \
+                  "or 'mcolors.BivariateNorm'"


shouldn't BivariateNorm subclass Normalize? 'specially since now it's normalizing down to a 1d space?

BivariateNorm does not inherit anything from Normalize so I thought it should not subclass it.

With the new design, BivariateNorm should havel all the same classes as Normalize...and I think subclassing it so it's registered generically as a Norm is preferable to having to treat it as a special case when it doesn't need to be.

This can be done with an ABC https://docs.python.org/3/library/abc.html#abc.ABCMeta.register which both Normalize and BivariateNorm register with.

Sure, but I don't see a reason why BivariateNorm shouldn't be subclassing Norm as it has the same architecture as any other norm...take data (scaler or vector)->map to 1D lut value ->get color, and color ->lut->scaler/vector.

If the code paths are completely different, why force the sub-class? In general, sub-classing is only worth it when you can share significant amounts of implementation details, otherwise duck-typing is better.

My point is mostly that the code paths shouldn't end up being all that different and more to the point in mpl land there's probably a ton of code that checks if things are norms and if this code works in those instances it doesn't make sense to change all that code to cover both cases when this is still fundamentally a norm. But I think I'm on the same page as you on the solution maybe being a "Norm" metaclass.

story645 · 2017-07-24T02:01:37Z

lib/matplotlib/colors.py

+        temp[0] = temp[0] * (256)
+        temp[1] = temp[1] * (256)
+        temp = temp.astype(int)
+        return temp[0] + temp[1] * 256


can you guarantee that this scheme always yields uniqueness?
(I'm confused about what's going on here and sort of have been thinking about a more general mapping between nd arrays and numbers wherein you get a sorted unique list of the nd tuples in the dataset so that the lookup/map is just the index of the value in the sorted list.

Since after normalization every variable ends up between 0 to 1. So here I am changing it to index flattened 1d lut of BivariateColormap by mapping 0-1 to 1 to 256*256. Here I am currently hardcoding 256 as number of rows and columns in lut of 2d colormap. But this should be changed somehow to use N attribute of BivariateColormap.

This is column major indexing (row * num_cols + col is c / numpy style, col * num_rows + row).

I am more concerned that the descritization is happening here rather than in the color map (which is done https://github.com/matplotlib/matplotlib/blob/master/lib/matplotlib/colors.py#L483) which prevents the size of the colormap from having to be known by the norm.

Why can't this return a Nd array? I think the most important places the norm and color map are called is https://github.com/matplotlib/matplotlib/blob/master/lib/matplotlib/cm.py#L209 and
https://github.com/matplotlib/matplotlib/blob/master/lib/matplotlib/image.py#L365 . In the first case it can be blindly passed through and in the second that resampling needs to happen over each plane.

In the case of the color bar, that already needs to be heavily special cased.

20:11 $ git grep '\.norm[ (]' | grep -v linalg lib/matplotlib/cm.py: self.norm = norm lib/matplotlib/cm.py: x = self.norm(x) lib/matplotlib/cm.py: self.norm = norm lib/matplotlib/collections.py: self.norm = other.norm lib/matplotlib/colorbar.py: y = self.norm(self._boundaries.copy()) lib/matplotlib/colorbar.py: b = self.norm(self._boundaries, clip=False).filled() lib/matplotlib/colorbar.py: xn = self.norm(x, clip=False).filled() lib/matplotlib/colorbar.py: self.norm = mappable.norm lib/matplotlib/colorbar.py: facecolor=self.cmap(self.norm(val)), lib/matplotlib/image.py: A = self.norm(A)

I am still a bit concerned that the descritization is being done here rather than in the color map. Won't this break for any colormap which is not 256 by 256?

tacaswell · 2017-07-27T21:46:07Z

lib/matplotlib/axes/_axes.py

+        isNorm = isinstance(norm, (mcolors.Normalize, mcolors.BivariateNorm))
+        if norm is not None and not isNorm:
+            msg = "'norm' must be an instance of 'mcolors.Normalize' " \
+                  "or 'mcolors.BivariateNorm'"


This can be done with an ABC https://docs.python.org/3/library/abc.html#abc.ABCMeta.register which both Normalize and BivariateNorm register with.

tacaswell · 2017-07-27T21:48:27Z

lib/matplotlib/axes/_axes.py

+        temp = np.asarray(X)
+        if temp.ndim == 3 and isinstance(norm, mcolors.BivariateNorm):
+                    temp = norm(temp)
+                    X = cmap(temp, alpha=self.get_alpha(), bytes=True)


Doing this here means you can not change anything about the color map or norm later.

Removed this.

tacaswell · 2017-07-27T21:49:48Z

lib/matplotlib/axes/_axes.py


        if len(args) == 1:
            C = np.asanyarray(args[0])
+            isBivari = isinstance(norm, (mcolors.BivariateNorm,


Why would the norm ever be a colormap?

Thanks for pointing this out. Fixed now.

tacaswell · 2017-07-27T21:50:29Z

lib/matplotlib/axes/_axes.py

+                    cmap = mcolors.BivariateColormap()
+                if norm is None:
+                    norm = mcolors.BivariateNorm()
+                C = norm(C)


Why do this up-front? defering all of this to _make_image means the colormap and norm and be updated.

tacaswell · 2017-07-27T23:44:31Z

lib/matplotlib/colorbar.py

+                                      linthresh=norm.linthresh,
+                                      base=10)
+                else:
+                    if mpl.rcParams['_internal.classic_mode']:


Don't need any of the classic shims as there is no back-compatibility to maintain.

tacaswell · 2017-07-27T23:48:38Z

lib/matplotlib/colorbar.py

+             or len(self._x) >= self.n_rasterize):
+            self.solids.set_rasterized(True)
+
+    def _ticker(self, norm):


Probably better to factor this as def _make_ticker(norm, values, boundaries, locator, formatter): as a local function in update_ticks.

You mean local/private function of class or nested function?

nested function

tacaswell · 2017-07-28T00:18:18Z

lib/matplotlib/colors.py

+        temp[0] = temp[0] * (256)
+        temp[1] = temp[1] * (256)
+        temp = temp.astype(int)
+        return temp[0] + temp[1] * 256


This is column major indexing (row * num_cols + col is c / numpy style, col * num_rows + row).

I am more concerned that the descritization is happening here rather than in the color map (which is done https://github.com/matplotlib/matplotlib/blob/master/lib/matplotlib/colors.py#L483) which prevents the size of the colormap from having to be known by the norm.

Why can't this return a Nd array? I think the most important places the norm and color map are called is https://github.com/matplotlib/matplotlib/blob/master/lib/matplotlib/cm.py#L209 and
https://github.com/matplotlib/matplotlib/blob/master/lib/matplotlib/image.py#L365 . In the first case it can be blindly passed through and in the second that resampling needs to happen over each plane.

In the case of the color bar, that already needs to be heavily special cased.

20:11 $ git grep '\.norm[ (]' | grep -v linalg lib/matplotlib/cm.py: self.norm = norm lib/matplotlib/cm.py: x = self.norm(x) lib/matplotlib/cm.py: self.norm = norm lib/matplotlib/collections.py: self.norm = other.norm lib/matplotlib/colorbar.py: y = self.norm(self._boundaries.copy()) lib/matplotlib/colorbar.py: b = self.norm(self._boundaries, clip=False).filled() lib/matplotlib/colorbar.py: xn = self.norm(x, clip=False).filled() lib/matplotlib/colorbar.py: self.norm = mappable.norm lib/matplotlib/colorbar.py: facecolor=self.cmap(self.norm(val)), lib/matplotlib/image.py: A = self.norm(A)

tacaswell · 2017-07-28T00:22:02Z

lib/matplotlib/colorbar.py

+        self.ax.set_xlabel(self._xlabel, **self._labelkw)
+        self.stale = True
+
+    def set_label(self, xlabel, ylabel, **kw):


The different signature from the ColorBar base class is a bit problematic, but not sure what to do about it.

perhaps set_labels(self, xylabel) (where xylabel=(xlabel, ylabel)); set_xlabel; set_ylabel?
let set_label raise an error
this likely would play better with traitlets

patniharshit · 2017-08-20T09:18:59Z

@anntzer can you have a look at this PR again? Some things have changed since your last review.

tacaswell · 2017-08-20T15:17:54Z

lib/matplotlib/axes/_axes.py

-        C = C.ravel()
+        # convert to one dimensional arrays if univariate
+        if isinstance(norm, mcolors.BivariateNorm):
+            C = np.asarray([C[0].ravel(), C[1].ravel()])


This may be better done as np.asarray([c.ravel() for c in C]) to be a bit future proof.

tacaswell · 2017-08-20T15:19:03Z

lib/matplotlib/axes/_axes.py

-            msg = "'norm' must be an instance of 'mcolors.Normalize'"
+
+        if norm is not None and not isinstance(norm, mcolors.Norms):
+            msg = "'norm' must be an instance of 'mcolors.Normalize' " \


We have a style preference for using () for continuation rather than \ so

msg = ('...' + '...')

The + is not needed.

anntzer · 2017-08-20T16:08:15Z

lib/matplotlib/axes/_axes.py


        if len(args) == 1:
            C = np.asanyarray(args[0])
-            numRows, numCols = C.shape
+            isBivari = (isinstance(norm, mcolors.BivariateNorm)


use_underscores_notCapsForVariables

Changed isBivari to is_bivari.

anntzer · 2017-08-20T16:09:02Z

lib/matplotlib/collections.py

@@ -730,7 +730,8 @@ def update_scalarmappable(self):
        """
        if self._A is None:
            return
-        if self._A.ndim > 1:
+        if (self._A.ndim > 1 and
+                not isinstance(self.norm, mcolors.BivariateNorm)):


should check ndim == 2 in this case?

anntzer · 2017-08-20T16:10:57Z

lib/matplotlib/colorbar.py

+        """
+        Label the axes of the colorbar
+        """
+        self._xlabel = '%s' % (xlabel, )


just self._xlabel = xlabel should be enough (unless we explicitly want to support any object via conversion to string, and even then str(xlabel) is clearer.

Just imitating how it is done in ColorbarBase.set_label.

anntzer · 2017-08-20T16:11:39Z

lib/matplotlib/colors.py

+    """
+    Abstract Base Class to group `Normalize` and `BivariateNorm`
+    """
+    __metaclass__ = ABCMeta


definitely wrong on Py3
use six
should have been caught by pycodestyle perhaps?

tacaswell · 2017-08-27T17:49:03Z

Appveyor seems to be having issues with connectivity the last day or so. It is failing on almost all PRs.

The circle / docs failures do need to be fixed though.

tacaswell · 2017-08-27T18:10:12Z

examples/images_contours_and_fields/bivariate_demo.py

+# Bivariate plotting demo
+# -----------------------
+
+air_temp = np.load(get_sample_data('air_temperature.npy'))


How big are these data files? I have a slight preference for 'generated' data for examples (to keep the repository size small).

84.2 kB each.
Couldn't come up with generated data so used these.

That isn't too bad, suspect that is smaller than the test images.

One argument in favour of this can be that the user seeing the example might relate to use case of bivariate plotting better with a real world example than with some mathematical function.

tacaswell · 2017-08-27T18:11:10Z

lib/matplotlib/axes/_axes.py

            to colors based on the `norm` (mapping scalar to scalar)
            and the `cmap` (mapping the normed scalar to a color).

-        cmap : `~matplotlib.colors.Colormap`, optional, default: None
+        cmap : `~matplotlib.colors.Colormap`,


This is the one exception to the () over \ rule so that numpydoc correctly grabs this as a single lin.

These might be the source of the docs failures?

Added \ but still failing.

It is probably fastest to build the docs locally. That also lets you get access to the full traceback.

pip install -r doc-requirements.txt cd docs python make.py html

should work.

tacaswell · 2017-08-27T18:30:13Z

This is looking good 👍

tacaswell · 2017-08-27T18:48:21Z

Can you build the docs locally?

patniharshit · 2017-08-28T19:02:41Z

@tacaswell doc build is passing now.

tacaswell · 2017-09-02T11:52:03Z

@patniharshit I broke your branch via #8966 this needs to be rebased (sorry).

I think it is ok to either just update the shim around the resampling code or to move it all into a loop (I think clearer now) to resample each plane.

patniharshit · 2017-09-03T09:12:29Z

I don't have access to my laptop for two days. Will resolve the conflicts by Wednesday.

shoyer · 2017-09-18T18:07:42Z

lib/matplotlib/axes/_axes.py

            - MxNx3 -- RGB (float or uint8)
            - MxNx4 -- RGBA (float or uint8)
+            - 2xMxN -- bivariate values to be mapped (float or int)


What happens when the input is shape (2, 3, 4)?

This sort of ambiguity should be avoided (and indicates that this API is probably not the right design choice).

jklymak

This PR came up in #14168.

I'm not sure the status of this, but in case someone comes back to it...

This looks technically well done. However, I'm not a fan of this API (see @shoyer comment as well). Its very complex from a documentation point of view. There is no way from the dosctrings that someone who didn't know what this PR did would understand what is meant by "2xMxN -- bivariate values to be mapped (float or int)". Further propagating support for two norms across all the 2-D plotting functions is going to be a maintenance nightmare. I'd far prefer this just got its own methods, i.e.: ax.bivariate_pcolormesh(x, y, Z1, Z1, norm1, norm2) etc.

I'm also very concerned about adding a whole new colorbar implementation. Maybe its unavoidable, but given that its so different, I wonder if it should also get its own method...

shoyer · 2019-05-08T16:24:28Z

I agree with putting this into a separate API. imshow is already pretty overloaded.

In shaping this API, I think it would be helping to identify clear use-cases / critical user journeys and make sure they are well supported and documented. For example, one good use case is visualizing complex-valued functions with luminosity and (perceptually uniform?) hue.

My concern is that in general this isn't a terribly effective visualization technique. That's not to say matplotlib shouldn't have this functionality, but let's try to limit its impact on the rest of the API.

jklymak · 2019-05-08T17:30:32Z

My concern is that in general this isn't a terribly effective visualization technique. That's not to say matplotlib shouldn't have this functionality, but let's try to limit its impact on the rest of the API.

My use case is shading a quantity based on gradient strength... If you'll forgive the sacrilege of linking to mathworks. The API for this was just:

shadedpcolor(x,z,U,(dU),[-1 1]/4,[-1 1]/300,0.7,cmp,0); 
axis ij; 
shadedcolorbar('v',[-1 1]/4,0.7,cmp);

https://www.mathworks.com/matlabcentral/fileexchange/14157-shaded-pseudo-color

github-actions · 2023-04-13T01:51:56Z

Since this Pull Request has not been updated in 60 days, it has been marked "inactive." This does not mean that it will be closed, though it may be moved to a "Draft" state. This helps maintainers prioritize their reviewing efforts. You can pick the PR back up anytime - please ping us if you need a review or guidance to move the PR forward! If you do not plan on continuing the work, please let us know so that we can either find someone to take the PR over, or close it.

timhoffm · 2024-12-13T06:08:40Z

This is superseeded. Tracked in #14168, and worked on in #28658, #28454, #29221.

story645 reviewed Jun 9, 2017

View reviewed changes

anntzer requested changes Jun 11, 2017

View reviewed changes

story645 changed the title ~~2D normalizer~~ [WIP] 2D normalizer Jun 11, 2017

patniharshit force-pushed the norm2d branch 2 times, most recently from 1a3d16b to 8e93d20 Compare July 7, 2017 12:51

patniharshit force-pushed the norm2d branch 2 times, most recently from 13627d3 to 2f8b5bf Compare July 23, 2017 17:22

anntzer reviewed Jul 23, 2017

View reviewed changes

story645 reviewed Jul 24, 2017

View reviewed changes

story645 changed the title ~~[WIP] 2D normalizer~~ [WIP] 2D Normalization & Colormapping for ScalerMappables Jul 24, 2017

tacaswell reviewed Jul 27, 2017

View reviewed changes

tacaswell reviewed Jul 28, 2017

View reviewed changes

patniharshit force-pushed the norm2d branch from ca3cf70 to a5cd938 Compare August 5, 2017 13:49

patniharshit force-pushed the norm2d branch from b43e6c7 to 23b1ffa Compare August 15, 2017 19:08

story645 changed the title ~~[WIP] 2D Normalization & Colormapping for ScalerMappables~~ 2D Normalization & Colormapping for ScalerMappables Aug 20, 2017

tacaswell reviewed Aug 20, 2017

View reviewed changes

anntzer reviewed Aug 20, 2017

View reviewed changes

tacaswell reviewed Aug 27, 2017

View reviewed changes

tacaswell added this to the 2.2 (next next feature release) milestone Aug 27, 2017

patniharshit added 2 commits August 28, 2017 00:49

Add line continuation in docstrings

75a289f

Debug doc build fail

7803974

patniharshit force-pushed the norm2d branch from 33cd323 to 84ff764 Compare August 28, 2017 18:14

Add missing symbols in docs

e40c87b

patniharshit force-pushed the norm2d branch from 84ff764 to e40c87b Compare August 28, 2017 18:39

shoyer reviewed Sep 18, 2017

View reviewed changes

anntzer mentioned this pull request Jan 10, 2018

Clip RGB data to valid range for imshow #10220

Merged

5 tasks

QuLogic mentioned this pull request May 8, 2019

Feature request: Bivariate colormapping #14168

Open

jklymak requested changes May 8, 2019

View reviewed changes

tacaswell mentioned this pull request Sep 10, 2019

Support pixel-by-pixel alpha in imshow. #14889

Merged

6 tasks

tacaswell marked this pull request as draft August 17, 2020 19:22

tacaswell mentioned this pull request Oct 26, 2020

colormap generalization #18809

Closed

tacaswell mentioned this pull request Aug 29, 2022

imshow doesn't normalize the color range in RGB images #9391

Closed

github-actions bot added the status: needs rebase label Oct 18, 2022

github-actions bot added the status: inactive Marked by the “Stale” Github Action label Apr 13, 2023

story645 mentioned this pull request Aug 24, 2023

[ENH]: 2D colormaps #26590

Closed

timhoffm closed this Dec 13, 2024

2D Normalization & Colormapping for ScalerMappables #8738

2D Normalization & Colormapping for ScalerMappables #8738

Conversation

patniharshit commented Jun 9, 2017 • edited Loading

PR Summary

PR Checklist

Choose a reason for hiding this comment

Choose a reason for hiding this comment

patniharshit Jun 9, 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

anntzer commented Jun 10, 2017

patniharshit commented Jun 10, 2017

anntzer commented Jun 10, 2017

story645 commented Jun 11, 2017 • edited Loading

anntzer commented Jun 11, 2017

story645 commented Jun 11, 2017 • edited Loading

anntzer commented Jun 11, 2017

anntzer left a comment

Choose a reason for hiding this comment

anntzer Jul 23, 2017 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

patniharshit Jul 29, 2017 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

patniharshit Jul 27, 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

story645 Jul 28, 2017 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

patniharshit Jul 27, 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

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

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

patniharshit commented Aug 20, 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

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

anntzer Aug 20, 2017 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

tacaswell commented Aug 27, 2017

Choose a reason for hiding this comment

patniharshit Aug 27, 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

patniharshit commented Jun 9, 2017 •

edited

Loading

patniharshit Jun 9, 2017 •

edited

Loading

story645 commented Jun 11, 2017 •

edited

Loading

story645 commented Jun 11, 2017 •

edited

Loading

anntzer Jul 23, 2017 •

edited

Loading

patniharshit Jul 29, 2017 •

edited

Loading

patniharshit Jul 27, 2017 •

edited

Loading

story645 Jul 28, 2017 •

edited

Loading

patniharshit Jul 27, 2017 •

edited

Loading

anntzer Aug 20, 2017 •

edited

Loading

patniharshit Aug 27, 2017 •

edited

Loading

jklymak commented May 8, 2019 •

edited

Loading