Fix bogus column and row at edge #318
Conversation
| @@ -152,14 +154,16 @@ def test_multiple_aggregates(): | |||
|
|
|||
|
|
|||
| def test_log_axis(): | |||
| # Upper bound for scale/index of x-axis | |||
| x_max_index = 10 ** (1 / (2 / np.log10(11))) | |||
| sol = np.array([[5, 5], [5, 5]], dtype='i4') | |||
There was a problem hiding this comment.
Thanks for tracking that calculation down!
| @@ -70,7 +72,7 @@ def test_sum(): | |||
|
|
|||
|
|
|||
| def test_min(): | |||
| out = xr.DataArray(df.i64.reshape((2, 2, 5)).min(axis=2).astype('f8').T, | |||
| out = xr.DataArray(df.i64.values.reshape((2, 2, 5)).min(axis=2).astype('f8').T, | |||
| coords=coords, dims=dims) | |||
There was a problem hiding this comment.
What's the implication of this change?
There was a problem hiding this comment.
It removes the deprecation warning about reshape. It looks like it was deprecated in Pandas 0.19.0.
|
This is ready for merging. |
|
@jbcrail We've been seeing kernel crashes in the notebook since this PR was merged. I did a git bisect and tracked it down to this commit: [68d51fe] Fix off-by-one error You can reliably reproduce this error by executing this code twice in a notebook: import pandas as pd
import datashader as ds
data = pd.read_csv('crash.csv')
data['k'] = data['k'].astype('category')
cvs = ds.Canvas(plot_width=400,
plot_height=400,
x_range=(data['x'].min(), data['x'].max()),
y_range=(data['y'].min(), data['y'].max()))
cvs.line(data, 'x', 'y', ds.count_cat('k'))You can download the crash.csv here. I suspect it's because |
|
@philippjfr I think you're right. Thanks for catching this. I'll open an issue and fix it. |
This reverts to exclusive ranges both manual and auto for all glyphs (points/line) without introducing regressions of holoviz#318, holoviz#330, and holoviz#343. I refactored several tests to make xarray coordinate indices easier to read and more explicit.
I applied @jbednar's changes to make the upper bound exclusive.
I also updated the test data to reflect the new upper bound semantics. In particular, the log axis tests required a seemingly magic value,
x_max_index. This value results from scaling, translating, and indexing the original x range from(1, 11)to(1, sqrt(11)). To account for precision errors, I calculated the final value using the same formula used by Datashader.Fix #259