Releases: LSSTDESC/rail_sklearn
v1.0.2
minor fixes to use list instead of tuple for parameters and computing number of chunks. Added an "only_colors" config parameter that enables you to run KNearestNeighbor using only the galaxy colors and without the reference band magnitude, which will give different results, particularly when the training set magnitude distribution is dramatically different than the test set magnitude distribution.
What's Changed
- Switch to using list instead of tuple for parameters, fix stage names by @eacharles in #19
- add 'only_colors' option to KNearNeigh #20 by @sschmidt23 in #21
- The total number of chunks was not computed properly by @joselotl in #23
Full Changelog: v1.0.1...v1.0.2
v1.0.1: Merge pull request #18 from LSSTDESC/ceci2
v1.0.0
RAIL v1.0 stamp
Full Changelog: v0.0.7...v1.0.0
v0.0.7
v0.0.6
Josue tracked down a bug where NZDir was failing in Jupyter if the input data was larger than the default chunk_size of 10,000 galaxies. This should no longer be an issue.
What's Changed
New Contributors
Full Changelog: v0.0.5...v0.0.6
v0.0.5: new copier template
v0.0.4: Random forest classifier
What's Changed
- Add random forest classifier for tomographer by @hangqianjun in #1
New Contributors
- @hangqianjun made their first contribution in #1
Full Changelog: v0.0.3...v0.0.4
v0.0.3
What's Changed
- add TEENY to distances by @sschmidt23 in #5
- add test for knn NaNs not present when same data in test and train by @sschmidt23 in #6
- missed one of two things I needed to change to fix the nan issue by @sschmidt23 in #7
- move open_model from init to run by @eacharles in #9
New Contributors
- @sschmidt23 made their first contribution in #5
- @eacharles made their first contribution in #9
Full Changelog: v0.0.2...v0.0.3
API consistency improvements part 1
Implementing more consistent names throughout src/rail/estimation
initial release
Full Changelog: https://github.com/LSSTDESC/rail_sklearn/commits/v0.0.1