feat: add box selection to support a box like selection from given selection #4324

Cloudac7 · 2023-10-25T08:42:59Z

Description:

It introduces a new geometry selection, as described in #4323.
A new selection language is defined in format:

box dimension(s) d1_min d1_max (d2_min d2_max) (d3_min d3_max) selection

It could pick a zone from the center of geometry from selection.

Changes made in this Pull Request:

Add new geometry selection: box

PR Checklist

Tests?
Docs?
CHANGELOG updated?
Issue raised/referenced?

Developers certificate of origin

I certify that this contribution is covered by the LGPLv2.1+ license as defined in our LICENSE and adheres to the Developer Certificate of Origin.

📚 Documentation preview 📚: https://mdanalysis--4324.org.readthedocs.build/en/4324/

pep8speaks · 2023-10-25T08:43:07Z

Hello @Cloudac7! Thanks for updating this PR. We checked the lines you've touched for PEP 8 issues, and found:

In the file package/MDAnalysis/core/groups.py:

Line 3114:80: E501 line too long (98 > 79 characters)
Line 3118:74: W291 trailing whitespace
Line 3119:78: W291 trailing whitespace

In the file package/MDAnalysis/core/selection.py:

Line 560:80: E501 line too long (83 > 79 characters)
Line 565:80: E501 line too long (83 > 79 characters)
Line 570:80: E501 line too long (83 > 79 characters)
Line 605:80: E501 line too long (80 > 79 characters)

In the file testsuite/MDAnalysisTests/core/test_atomselections.py:

Line 819:80: E501 line too long (81 > 79 characters)

Comment last updated at 2024-03-25 19:34:45 UTC

github-actions

Hello there first time contributor! Welcome to the MDAnalysis community! We ask that all contributors abide by our Code of Conduct and that first time contributors introduce themselves on the developer mailing list so we can get to know you. You can learn more about participating here. Please also add yourself to package/AUTHORS as part of this PR.

github-actions · 2023-10-25T08:46:08Z

Linter Bot Results:

Hi @Cloudac7! Thanks for making this PR. We linted your code and found the following:

Some issues were found with the formatting of your code.

Code Location	Outcome
main package	⚠️ Possible failure
testsuite	⚠️ Possible failure

Please have a look at the darker-main-code and darker-test-code steps here for more details: https://github.com/MDAnalysis/mdanalysis/actions/runs/8425924832/job/23073093482

Please note: The black linter is purely informational, you can safely ignore these outcomes if there are no flake8 failures!

codecov · 2023-10-25T09:00:19Z

Codecov Report

Attention: Patch coverage is 98.00000% with 1 lines in your changes are missing coverage. Please review.

Project coverage is 93.37%. Comparing base (0582265) to head (8d7f2ce).

❗ Current head 8d7f2ce differs from pull request most recent head f72c618. Consider uploading reports for the commit f72c618 to get more accurate results

Files	Patch %	Lines
package/MDAnalysis/core/selection.py	98.00%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@             Coverage Diff             @@
##           develop    #4324      +/-   ##
===========================================
- Coverage    93.65%   93.37%   -0.28%     
===========================================
  Files          168      170       +2     
  Lines        21215    22375    +1160     
  Branches      3908     4092     +184     
===========================================
+ Hits         19869    20893    +1024     
- Misses         888      963      +75     
- Partials       458      519      +61

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

IAlibay

Apologies for the slow reviews @Cloudac7

I've added some initial comments.

@MDAnalysis/coredevs could someone please take on the responsibility of reviewing this PR?

testsuite/MDAnalysisTests/core/test_atomselections.py

IAlibay · 2023-11-05T09:39:00Z

package/doc/sphinx/source/conf.py

@@ -14,6 +14,8 @@
 import sys
 import os
 import datetime
+sys.path.insert(0, os.path.abspath('../../..'))


Could you explain why this is being added?

Sorry for forgetting remove the line.
I am developing on MacOS platform and thus the compiling of documentation with original code would raise an ModuleNotFoundError: No module named 'MDAnalysis'. My practice followed the instruction of latest documentation for developing, and no solution found. After adding the line in more upper position, sphinx-build could work.

IAlibay · 2023-11-05T09:43:36Z

package/doc/sphinx/source/documentation_pages/selections.rst

@@ -277,6 +277,23 @@ cyzone *externalRadius* *zMax* *zMin* *selection*
       relative to the COG of *selection*, instead of absolute z-values
       in the box.

+box *dimension(s)* *d1_Max* *d1_Min* (*d2_Max* *d2_Min*) (*d3_Max* *d3_Min*) *selection*


I'm very likely missing something, but I don't know if I fully understand how this differs from using prop - could you maybe explain the benefits of this new selection over using a succession of prop?

Well, prop is used to define a static zone in each axes, in my understanding. The new defined box selection do similar thing like which sphzone, cyzone or cylayer do, thus could define a zone according to another selection which could also be dynamical when updating=True.

See @Cloudac7 's description in #4323 (comment) , which looks sensible to me. If we have cyzone then we can also have box.

this would make more sense to me as min to max values rather than the opposite. E.g. box z 0 8 bilayer to select a slice of water above a bilayer and box z -8 0 to get the other half.

@richardjgowers I had the same thoughts but unfortunately cyzone and friends started this inverse "max min" syntax. I feel consistency across the *zone and *layer selections is more important here than purity.

(I mean I am really looking for someone to convince me we should do min max here ;-) because I know that I am going to mistype it...)

package/CHANGELOG

richardjgowers · 2023-11-07T10:17:28Z

package/MDAnalysis/core/selection.py

+        super().__init__(parser, tokens)
+        self.periodic = parser.periodic
+        self.direction = tokens.popleft()
+        if self.direction not in self.combination:


can we set xmin xmax etc here to None then optionally set them to something else in the else: branch below? I think this would simplify the __getattribute__ usage below

richardjgowers · 2023-11-07T10:18:17Z

package/MDAnalysis/core/selection.py

+                # Triclinic version
+                tribox = group.universe.trajectory.ts.triclinic_dimensions
+                vecs -= tribox[2] * np.rint(vecs[:, 2] / tribox[2][2])[:, None]
+                vecs -= tribox[1] * np.rint(vecs[:, 1] / tribox[1][1])[:, None]
+                vecs -= tribox[0] * np.rint(vecs[:, 0] / tribox[0][0])[:, None]


this isn't technically correct for some triclinic boxes with some shifts.

I think you'll want this function:

mdanalysis/package/MDAnalysis/lib/distances.py

Line 1791 in c36313e

def minimize_vectors(vectors: npt.NDArray, box: npt.NDArray) -> npt.NDArray:

But it won't handle some, but not all, box dimensions being given (i.e. semi periodic calculation)

richardjgowers · 2023-11-07T10:23:21Z

package/doc/sphinx/source/documentation_pages/selections.rst

@@ -277,6 +277,23 @@ cyzone *externalRadius* *zMax* *zMin* *selection*
       relative to the COG of *selection*, instead of absolute z-values
       in the box.

+box *dimension(s)* *d1_Max* *d1_Min* (*d2_Max* *d2_Min*) (*d3_Max* *d3_Min*) *selection*


this would make more sense to me as min to max values rather than the opposite. E.g. box z 0 8 bilayer to select a slice of water above a bilayer and box z -8 0 to get the other half.

package/MDAnalysis/core/groups.py

RMeli · 2023-11-08T23:07:33Z

package/MDAnalysis/core/selection.py

+    combination = [
+        "x",
+        "y",
+        "z",
+        "xy",
+        "xz",
+        "yz",
+        "yx",
+        "zx",
+        "zy",
+        "xyz",
+        "xzy",
+        "yxz",
+        "yzx",
+        "zxy",
+        "zyx",
+    ]


Suggested change

combination = [

"x",

"y",

"z",

"xy",

"xz",

"yz",

"yx",

"zx",

"zy",

"xyz",

"xzy",

"yxz",

"yzx",

"zxy",

"zyx",

]

combination = [''.join(s) for i in range(1, 4) for s in itertools.permutations("xyz", i)]

Just for fun. Probably less readable than what you wrote.

But the question is, do we really need both xy and yx? What's the use case of having all permutations as opposed to only combinations?

Suggested change

combination = [

"x",

"y",

"z",

"xy",

"xz",

"yz",

"yx",

"zx",

"zy",

"xyz",

"xzy",

"yxz",

"yzx",

"zxy",

"zyx",

]

combination = [''.join(s) for i in range(1, 4) for s in itertools.combinations("xyz", i)]

Nitpick: the variable is called "combination", so the name might have to change to "permutation" ("permutations"?) if that's the actual use case.

The extensive list of all permutations of x, y, and z is designed to avoid the need to calculate the O(3!) = O(1) permutations each time a selection is executed. This has been reformatted by darker to comply with PEP8 standards. It's a minor optimization that can be overlooked, though. 🤣

Regarding the latter point, my original intention was to enhance the flexibility of the selection by making both yx and xy acceptable. However, this inadvertently added extra complexity. You're correct in suggesting that it would be better to have users write in the sequence of x -> y -> z. This approach would eliminate the need for additional processing.

RMeli · 2023-11-08T23:09:49Z

package/MDAnalysis/core/selection.py

+                "The direction '{}' is not valid. Must be combination of {}"
+                "".format(self.direction, ["x", "y", "z"])


Suggested change

"The direction '{}' is not valid. Must be combination of {}"

"".format(self.direction, ["x", "y", "z"])

f"The direction '{self.direction}' is not valid. Must be combination of {axis_map}"

doc: add doc about box selection fix: reformat with darker and flake8

style: reformat as pep8

test: add error catching for box selection fix: prevent abnormal input in direction

style: reformat CHANGELOG

commit 8d7f2ce Author: Futaki Haduki <812556867@qq.com> Date: Thu Nov 9 17:23:57 2023 +0800 Update package/MDAnalysis/core/groups.py Co-authored-by: Rocco Meli <r.meli@bluemail.ch> commit cfdec5b Author: Futaki Haduki <812556867@qq.com> Date: Thu Nov 9 17:23:44 2023 +0800 Update package/MDAnalysis/core/groups.py Co-authored-by: Rocco Meli <r.meli@bluemail.ch>

orbeckst · 2023-12-15T17:52:48Z

@richardjgowers I am tentatively assigning the PR to you for shepherding it through. If it's too much then please unassign yourself and ping me. Thanks.

orbeckst · 2024-03-25T19:35:58Z

@richardjgowers could you have a look if @Cloudac7 addressed the reviewers' comments? Thanks!

github-actions bot added the Component-Core label Oct 25, 2023

Cloudac7 mentioned this pull request Oct 25, 2023

Add a geometry selection supporting selecting an zone shaped like a box #4323

Open

github-actions bot reviewed Oct 25, 2023

View reviewed changes

orbeckst linked an issue Oct 26, 2023 that may be closed by this pull request

Add a geometry selection supporting selecting an zone shaped like a box #4323

Open

IAlibay reviewed Nov 5, 2023

View reviewed changes

IAlibay added the enhancement label Nov 5, 2023

richardjgowers requested changes Nov 7, 2023

View reviewed changes

Cloudac7 force-pushed the box-selection branch from bf93731 to 036eea5 Compare November 8, 2023 03:28

RMeli reviewed Nov 8, 2023

View reviewed changes

package/MDAnalysis/core/groups.py Outdated Show resolved Hide resolved

RMeli reviewed Nov 8, 2023

View reviewed changes

package/MDAnalysis/core/groups.py Outdated Show resolved Hide resolved

RMeli reviewed Nov 8, 2023

View reviewed changes

Cloudac7 force-pushed the box-selection branch from 8d7f2ce to 0090016 Compare November 10, 2023 03:05

Cloudac7 added 8 commits November 10, 2023 11:06

feat: add box selection

630beea

doc: add doc about box selection fix: reformat with darker and flake8

test: add test for orthogonal distance

2f94545

style: reformat as pep8

test: add test in triclinic cell for box selection

01758fc

test: add error catching for box selection fix: prevent abnormal input in direction

fix: remove extra line in sphinx conf.py

953b33a

style: reformat CHANGELOG

fix: use minimize_vectors to calculate min vec

79e307f

fix: set default dmin and dmax for box selection

7b3fb0d

doc: format docstring as well as doc for selections

827397c

Cloudac7 force-pushed the box-selection branch from 0090016 to 773ef5a Compare November 10, 2023 03:06

fix: rollback unexpected change in unittest

0430726

orbeckst assigned richardjgowers Dec 15, 2023

Merge branch 'develop' into box-selection

f72c618

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat: add box selection to support a box like selection from given selection #4324

feat: add box selection to support a box like selection from given selection #4324

Cloudac7 commented Oct 25, 2023 •

edited

pep8speaks commented Oct 25, 2023 •

edited

github-actions bot left a comment

github-actions bot commented Oct 25, 2023 •

edited

codecov bot commented Oct 25, 2023 •

edited

IAlibay left a comment

IAlibay Nov 5, 2023

Cloudac7 Nov 6, 2023 •

edited

IAlibay Nov 5, 2023

Cloudac7 Nov 6, 2023

orbeckst Nov 6, 2023

richardjgowers Nov 7, 2023

orbeckst Nov 10, 2023

richardjgowers Nov 7, 2023

richardjgowers Nov 7, 2023

richardjgowers Nov 7, 2023

richardjgowers Nov 7, 2023

RMeli Nov 8, 2023

Cloudac7 Nov 9, 2023 •

edited

RMeli Nov 8, 2023

orbeckst commented Dec 15, 2023

orbeckst commented Mar 25, 2024

		"The direction '{}' is not valid. Must be combination of {}"
		"".format(self.direction, ["x", "y", "z"])

	"The direction '{}' is not valid. Must be combination of {}"
	"".format(self.direction, ["x", "y", "z"])
	f"The direction '{self.direction}' is not valid. Must be combination of {axis_map}"

feat: add box selection to support a box like selection from given selection #4324

Are you sure you want to change the base?

feat: add box selection to support a box like selection from given selection #4324

Conversation

Cloudac7 commented Oct 25, 2023 • edited

PR Checklist

Developers certificate of origin

pep8speaks commented Oct 25, 2023 • edited

Comment last updated at 2024-03-25 19:34:45 UTC

github-actions bot left a comment

Choose a reason for hiding this comment

github-actions bot commented Oct 25, 2023 • edited

Linter Bot Results:

codecov bot commented Oct 25, 2023 • edited

Codecov Report

IAlibay left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Cloudac7 Nov 6, 2023 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Cloudac7 Nov 9, 2023 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

orbeckst commented Dec 15, 2023

orbeckst commented Mar 25, 2024

Cloudac7 commented Oct 25, 2023 •

edited

pep8speaks commented Oct 25, 2023 •

edited

github-actions bot commented Oct 25, 2023 •

edited

codecov bot commented Oct 25, 2023 •

edited

Cloudac7 Nov 6, 2023 •

edited

Cloudac7 Nov 9, 2023 •

edited