Allow to use a relative score threshold, which can detect identical points passed to FPS #129
Conversation
This file contains 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
rosecers
reviewed
May 13, 2022
The issue with thresholding is that it works using absolute values (as far as I understand); and here I would like to use relative values (i.e. stop when the score goes below a fraction of the first selection). |
|
One quick comment: I wouldn't throw an error if the same points are given
as input, only if one starts selecting identical points.
Eliminating duplicates might be one of the reasons to use FPS in the first
place.
Michele
…On Fri, 13 May 2022 at 17:25, Guillaume Fraux ***@***.***> wrote:
***@***.**** commented on this pull request.
------------------------------
In examples/PlotLFRE.ipynb
<#129 (comment)>
:
> @@ -140,7 +140,7 @@
"ax34.set_title(r\"$X^-$ LFRE(3-body, 4-body)\")\n",
"ax43.set_title(r\"$X^-$ LFRE(4-body, 3-body)\")\n",
"\n",
- "cbar = fig.colorbar(pcm, ax=(ax34, ax43), label=\"LFRE\", location=\"bottom\")\n",
+ "cbar = fig.colorbar(pcm, ax=[ax34, ax43], label=\"LFRE\", location=\"bottom\")\n",
recent versions of matplotlib explicitly checks if this is a list when
plotting
—
Reply to this email directly, view it on GitHub
<#129 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AAIREZ5JTS6GIVSOGAPZD4DVJZX5ZANCNFSM5U6RWGNA>
.
You are receiving this because you are subscribed to this thread.Message
ID: ***@***.***>
|
The PR title was not very clear, but that's exactly what the code is doing: trying to detect identical point being selected & raising an error in this case. |
Luthaf
force-pushed
the
fps-selection-identical-points
branch
from
5780576 to
4afb616
Compare
Luthaf
force-pushed
the
fps-selection-identical-points
branch
from
4afb616 to
7b9b775
Compare
Luthaf
changed the title
rosecers
reviewed
May 16, 2022
There was a problem hiding this comment.
In general, I am good with this bar a few thoughts.
- I think it should be
score_threshold_type, notscore_threshold_kind. Kind is a bit of an atypical word to use in this context. - I changed the docstring for clarity (this would need to be propagated throughout).
Luthaf
force-pushed
the
fps-selection-identical-points
branch
from
7b9b775 to
bff8d2c
Compare
|
@rosecers is this good to go? |
rosecers
reviewed
May 23, 2022
This allows to stop selection early when identical points are fed to FPS or another selector by using something like score_threshold=1e-12 and score_threshold_type="relative".
Luthaf
force-pushed
the
fps-selection-identical-points
branch
from
bff8d2c to
6038ace
Compare
rosecers
approved these changes
May 25, 2022
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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.
Fixes #128
I'm not sure if this is the best approach to do so, but I think this is a case we want to detect automatically by default.
Score thresholding seems like it would be useful here as well, but since it works using absolute values, we can not enable it by default without knowing about the magnitude of features.
I could also modify score thresholding to work with relative scores, and use that instead.