Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fix dominick dataset bug. #2364

Merged
merged 4 commits into from
Oct 11, 2022
Merged

Fix dominick dataset bug. #2364

merged 4 commits into from
Oct 11, 2022

Conversation

haskarb
Copy link
Contributor

@haskarb haskarb commented Oct 11, 2022

Fixes: #2355

Description of changes:
changed to "W" to "W-SUN"

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

Please tag this pr with at least one of these labels to make our release process faster: BREAKING, new feature, bug fix, other change, dev setup

@jaheba
Copy link
Contributor

jaheba commented Oct 11, 2022

Thanks for the PR!

A bit of a nitpick, but the title should be in present tense and also has a . at the end.

@haskarb haskarb changed the title Fixed dominick dataset Fix dominick dataset bug. Oct 11, 2022
@haskarb
Copy link
Contributor Author

haskarb commented Oct 11, 2022

Done :)

jaheba
jaheba previously approved these changes Oct 11, 2022
@@ -238,7 +238,7 @@ def default_prediction_length_from_frequency(freq: str) -> int:
"T": 60,
"H": 48,
"D": 30,
"W": 8,
"W-SUN": 8,
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This fix doesn't specific to dominick, but be an issue with the canonical name of the weekly frequency.

Otherwise looks good to me.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes, I was trying to find out which other dataset it affects, but there aren't many dataset. I think it comes from a specific condition arising out of the Dominick dataset. I marked it as Dominick, because yesterday, I encountered issue because of it.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Should W-SUN be added along with W, instead of replacing?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes. I think that would be better. :)

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Updated.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

No, it should not. W-SUN is the canonical frequency

@haskarb haskarb marked this pull request as ready for review October 11, 2022 13:07
Copy link
Contributor

@jaheba jaheba left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Let's remove W again

@jaheba jaheba merged commit a5a5bb1 into awslabs:dev Oct 11, 2022
@haskarb haskarb deleted the updatedominick branch October 11, 2022 14:34
@lostella lostella added bug fix (one of pr required labels) pending v0.10.x backport This contains a fix to be backported to the v0.10.x branch pending v0.11.x backport This contains a fix to be backported to the v0.11.x branch labels Oct 11, 2022
lostella pushed a commit to lostella/gluonts that referenced this pull request Oct 28, 2022
@lostella lostella mentioned this pull request Oct 28, 2022
lostella pushed a commit to lostella/gluonts that referenced this pull request Oct 28, 2022
@lostella lostella mentioned this pull request Oct 28, 2022
lostella added a commit that referenced this pull request Oct 28, 2022
* Fix numerical bug in `BinnedUniforms` (#2344)

prob.cumsum() - prob is a numerically unsafe way to shift prob.cumsum() to the right. The numerical error will create small gaps or overlaps in the bins defined by lower and upper bounds. If a quantile falls in one of these overlap or gap the one hot bin indicator will have 0 or 2 hot bins and generate a shape mismatch error. The fix consists of doing a shift by copy.

* Fix dominick dataset bug. (#2364)

* Proposed fix to zero seed bug. (#2379)

Co-authored-by: Sigrid Passano Hellan <[email protected]>

Co-authored-by: Marc van Oudheusden <[email protected]>
Co-authored-by: Bhaskar Dhariyal <[email protected]>
Co-authored-by: Sigrid Passano Hellan <[email protected]>
Co-authored-by: Sigrid Passano Hellan <[email protected]>
lostella added a commit that referenced this pull request Oct 28, 2022
* Fix dominick dataset bug. (#2364)

* Remove strange quoting marks from docstrings (#2368)

* Change 'confidence interval' to 'prediction interval' (#2373)

* Proposed fix to zero seed bug. (#2379)

Co-authored-by: Sigrid Passano Hellan <[email protected]>

Co-authored-by: Bhaskar Dhariyal <[email protected]>
Co-authored-by: Nils Kiele <[email protected]>
Co-authored-by: Sigrid Passano Hellan <[email protected]>
Co-authored-by: Sigrid Passano Hellan <[email protected]>
@lostella lostella removed pending v0.10.x backport This contains a fix to be backported to the v0.10.x branch pending v0.11.x backport This contains a fix to be backported to the v0.11.x branch labels Nov 17, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
bug fix (one of pr required labels)
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Unable to load Dominick dataset
3 participants