~javiljoen/lttb-numpy

Remove builds.sr.ht script
Update CI script to stay on debian/buster

debian/stable is now bullseye, which doesn't provide Python 2 packages.
Update links for issues/PRs in Contributing docs
Update changelog
Add docstrings; check with pydocstyle
Add flake8 plugin to check for naming convention
Add build script task to render & upload readme

Sourcehut only displays `README` or `README.md` by default.
Rst-formatted docs need to be compiled to HTML manually:
https://man.sr.ht/git.sr.ht/#setting-a-custom-readme
Replace Travis CI with builds.sr.ht script
Update latest supported version to Python 3.9
Move "Contributing" docs to separate file
Update project links (remove GitHub references)
b13cb1ae — Jack Viljoen 2 years ago v0.3.0
Update changelog and setup.py
f7ed36ea — Jack Viljoen 2 years ago
Add more usage & contribution info to readme
24de93fe — Jack Viljoen 2 years ago
Add check for NaN values in input data

This check has also been added to the list of default validators.
This change is not backwards-compatible,
but the previous behaviour was that `downsample()` would preferentially
select the NaN values for inclusion in the output
(presumably because it considers these points to be maximally distant
from the anchors in the neighbouring bins),
so users with missing data would either have gotten suboptimal results
(a higher proportion of missing data in the output array)
or they might have implemented their own up-front checks.
So in the unlikely case where this breaks someone's workflow,
this now requires them to make an explicit decision whether to
throw an error, exclude missing values up-front,
or accept output in which nulls are overrepresented.
6d279912 — Jack Viljoen 2 years ago
Make validation of input data configurable

By changing the `validators` argument of `lttb.downsample()` from the
default list of validation functions, the data can be checked against
stricter or looser requirements.
For example, if the data is known to always satisfy the requirement for
the x-values to be strictly monotonic, some CPU cycles can be saved by
removing `x_is_strictly_increasing()` from the list.

This change is backwards-compatible.

It would be better to make the `validators` argument keyword-only, but
that would prevent this library from working on Python 2.7.
cfc86f5b — Jack Viljoen 2 years ago
Fix `make lint` command and add `isort` check
3543a0ce — Jack Viljoen 3 years ago v0.2.2
Update changelog
aa507d4a — Jack Viljoen 3 years ago
Merge pull request #6 from jkrueger1/master

Fixes setup.py non-ASCII character issue.
545e9e11 — Jens Krüger 3 years ago
Fixes setup.py non-ASCII character issue.
058057a2 — Jack Viljoen 3 years ago
Add coverage commands to tox & Makefile
Next