Python 3.12 support + scipy 1.16 fix#462
Merged
ThomasMBury merged 2 commits intoThomasMBury:mainfrom Apr 3, 2026
Merged
Conversation
- Add Python 3.12 to CI matrix with TF gate (no 3.12 wheels for TF <2.16) - Drop Python 3.8 (EOL Oct 2024), update Actions to v4/v5 - Update dependency floors for 3.12 wheel availability (numpy>=1.26, arch>=6.2) - Separate TF-dependent tests with pytest marker - Correct TF version constraint (<=2.16 -> <2.16, excludes Keras 3) Prepared with assistance from Claude Opus 4.6.
scipy 1.16.0 rewrote signal.welch internals to use ShortTimeFFT, which performs tuple indexing (x[..., i0:i1]) that fails on pandas Series. Convert input to ndarray before passing to signal.welch. np.asarray is idempotent for ndarray input and safe across all scipy versions. Prepared with assistance from Claude Opus 4.6.
Owner
|
Hi Bruce, |
Owner
|
Just tested locally on Python 3.12 and everything looks good. I appreciate the clean and well-documented changes - thanks! |
Contributor
Author
|
You're very welcome. Happy to help.
Bruce
…On Fri, Apr 3, 2026 at 3:50 PM Thomas Bury ***@***.***> wrote:
*ThomasMBury* left a comment (ThomasMBury/ewstools#462)
<#462 (comment)>
Just tested locally on Python 3.12 and everything looks good. I appreciate
the clean and well-documented changes - thanks!
—
Reply to this email directly, view it on GitHub
<#462 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAHHSX6UCMNHAJACPQN7QQL4UA55HAVCNFSM6AAAAACWKQ6WTGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHM2DCOBVGUYDMMRZG4>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Hey Thomas — a colleague wanted to use ewstools on Python 3.12 and hit the version cap, so I dug into the dependency chain to see what it would take.
Turns out it's straightforward with a conservative approach:
pytest.importorskip— rest of the suite runs fine without TF<=2.16to<2.16(2.16.0 ships Keras 3, which is a breaking change)While testing, I also found that scipy 1.16+ breaks
signal.welch()when passed a pandas Series directly (they refactored the internals to use ShortTimeFFT). One-line fix: wrap withnp.asarray()before passing. That's the second commit.Happy to adjust anything — let me know!