Automatic orbital wise analysis functionality in analyze module #132

naik-aakash · 2023-08-02T19:24:44Z

This PR implements the feature request of #113

Added

Automatic orbital-wise analysis functionality in analyze module

Todo

…rieve plot data

naik-aakash · 2023-08-03T07:27:51Z

Below is the current output of condensed bonding analysis dict when oribital wise analysis is switched on. Please have a look at the output information I am providing. Suggestions are most welcome 😃

JaGeo · 2023-08-08T16:18:33Z

Closes #113

naik-aakash · 2023-08-11T13:23:45Z

Hi @JaGeo , here is the updated output skeleton that fixes the discrepancies in previous results.

NaCl

3s-3s (57%)

3px-3s (42%)

#ZnS

3s-4s (38%)

3py-4s (20%)

naik-aakash · 2023-08-18T15:24:27Z

This PR is ready for review now. I have updated the interactive plotter legends for Orbitalwise analysis.

JaGeo · 2023-08-18T15:25:33Z

Btw, could you make a PR adding the units to the plotter as well?

naik-aakash · 2023-08-18T15:31:05Z

Btw, could you make a PR adding the units to the plotter as well?

Yes sure! Thanks for pointing this out. Somwhow got overlooked so far

naik-aakash · 2023-08-25T16:40:14Z

Hi @JaGeo, I think it would be better if I introduce another parameter to limit the orbital contributions, I am currently using the same parameter limit of ICOHP cutoff i.e. 10% for limiting the orbital contributions. I found 10% cutoff can result in no orbital interactions being shown. (For example, in diamond, mp-66: no orbital-orbital interactions are detected using a 10% cutoff as all the contributions to total summed icohp are around 8%)

If we simply change the cutoff_icohp arg in the analysis, this fixes the issue. But that could affect our other results

Please let me know what you think is the best.

naik-aakash · 2023-09-07T15:58:04Z

Hi @JaGeo , it would be best to merge this after COBI coop extension PR as lot of changes needs to be updated in here (Merge conflicts will arise)

JaGeo · 2023-10-05T17:50:58Z

lobsterpy/cohp/test/test_describe.py

+                "In the Sb-F bond, the maximum contribution are from Sb(5pz)-F(2s), Sb(5py)-F(2s), "
+                "and Sb(5px)-F(2s) orbitals, contributing 39.12, 39.12, and 39.12 percent, respectively, "
+                "whereas the maximum antibonding contribution is from the Sb(5s)-F(2s) orbital, "
+                "contributing 35.799 percent.",


It says: there is no antibonding contribution but then you list an orbital contribution for it.

Yes, I also spotted this inconsistency yesterday. Trying an alternative approach for it. It happens because I did not use summed cohps of all relevant bonds , but rather looked relevant bond label to extract orbital contributions. Working on it since yesterday

I tried with alternative approach, results change when using summed cohps, but it will still happen and it makes sense as we also see it in the plots 😅

Then you need to rephrase the text ;). The sum is 0 but there are antibonding contributions from individual orbitals.

Yes, but then the text is not describing it correctly. It's not a percentage of the 0.0% antibonding interactions. Make sure this is clear from the text.

In general: what percentage are you giving there? orbital COHP/overall COHP? And then antibonding orbtia interactions/overall orbital interaction strength? Might also not be exactly clear.

JaGeo · 2023-10-05T17:52:01Z

lobsterpy/cohp/test/test_describe.py

+                "It has 4 C-C (mean ICOHP: -9.59 eV, 0.0 percent antibonding interaction below EFermi) bonds.",
+                "In the C-C bond, the maximum contribution are from C(2py)-C(2s), C(2pz)-C(2s), C(2px)-C(2s), "
+                "C(2s)-C(2py), C(2s)-C(2pz), and C(2s)-C(2px) orbitals, contributing 7.94, 7.94, 7.94, 7.94, "
+                "7.94, and 7.94 percent, respectively, whereas the maximum antibonding contribution is from "


0.0 percent antibonding contribution and then a list of the max antibonding contribution.

naik-aakash · 2023-10-06T14:53:15Z

Hi @JaGeo , I found the current version of implementation does not work very well with just using ICOHPs to find the contributions, I would need to rewrite it.

I Think to find most relevant orbitals, I would need to first compute area under intergrals from COHPs and then compute the percentage of contribution (This will take longer to run especially for larger structures or that have many orbital pairs)

If you have any suggestions, please let me know 😃

JaGeo · 2023-10-06T15:23:54Z

Maybe, you could add this as an additional optional flag and not make it part of the standard orbital-wise analysis? I think people might just want to plot the most important orbital interactions and not much more.

lobsterpy/cohp/test/test_describe.py

naik-aakash · 2023-10-11T07:08:34Z

Hi @JaGeo, I made the code changes to include space between the sentences. So this PR could be merged. Let me know if I should add more comments for the implementation.

JaGeo

Hi @naik_aakash, this looks great. I left some further comments to avoid potenial code repetition.

Next steps should be documenting this functionality very well - both within our repository and for our publication. I would like to wait with a release until we have done this. I think users might have a lot of questions without it.

Another idea that I had: should we move to pytest completely? I think there are automatic tools to do so.

lobsterpy/plotting/__init__.py

JaGeo · 2023-10-11T11:10:25Z

lobsterpy/plotting/__init__.py

+                if label_with_count + suffix not in self._cohps:
+                    self._cohps[label_with_count + suffix] = {}
+                plot_data_orb = analyse.get_site_orbital_resolved_labels()
+                drop_down_key = label_with_count + suffix


this looks similar below.

this looks similar below.

Now with comments added, it probably makes clearer how it differ

lobsterpy/plotting/__init__.py

naik-aakash · 2023-10-11T11:37:44Z

Hi @naik_aakash, this looks great. I left some further comments to avoid potenial code repetition.

Next steps should be documenting this functionality very well - both within our repository and for our publication. I would like to wait with a release until we have done this. I think users might have a lot of questions without it.

Another idea that I had: should we move to pytest completely? I think there are automatic tools to do so.

Yes, I wiIl document this soon for repo and publication. And I am not familiar with the tools, but should be easy I guess to find it out. Will check that

JaGeo · 2023-10-23T17:30:02Z

I will merge this now. However, we need a clear description of how this orbital-wise automatic analysis works in our documentation, the tutorials and the paper. We should showcase the results for several examples of different complexity.

naik-aakash · 2023-10-24T04:32:49Z

I will merge this now. However, we need a clear description of how this orbital-wise automatic analysis works in our documentation, the tutorials and the paper. We should showcase the results for several examples of different complexity.

Thanks, I will do this 😃

added orbital wise analysis functionality in analyze module

2cca572

naik-aakash added the enhancement New feature or request label Aug 2, 2023

naik-aakash self-assigned this Aug 2, 2023

naik-aakash added 3 commits August 2, 2023 21:28

add missing docstring

855f4aa

update analyze orbital wise analysis data

cf5fedf

removed redundant code block, use new method of analyze module to ret…

dc68d7f

…rieve plot data

naik-aakash and others added 13 commits August 11, 2023 15:25

fix orbitalwise analysis bug

04ff0e5

Merge branch 'JaGeo:main' into orbital_resolved_analysis

4a86355

cleanup code

52d62c2

update Plaincohpplotter> orbital resolved plots

41d0534

more cleanup and update describe and plotting

dab0fa1

add missing doc string

6682a5c

fix pylint

faaced6

clean up interactive plotter and improve plot legends

4969976

update describe module text to show relevant orbital info

ed7f672

update test workflow to show missing lines

7e358cb

fix describe module bug

a3e5591

fix describe module

4f21094

update describe text, add tests and test data

03762a4

naik-aakash and others added 3 commits August 25, 2023 18:58

fix erroneous hardcarded cutoff limit for orb resolved analysis

61122f1

add missing blank space in text description

daf2939

Merge branch 'main' into orbital_resolved_analysis

9f1953a

naik-aakash added 3 commits October 3, 2023 21:22

fix interactive plotting bug

7c4eebe

improve orbitalwise implementation for max anti bonding contri

f470b41

update tests and test files

b1560ea

JaGeo reviewed Oct 5, 2023

View reviewed changes

naik-aakash and others added 4 commits October 6, 2023 16:13

change orb resolved analysis to use summed cohps

2354af5

run black

e4060f7

Merge branch 'main' into orbital_resolved_analysis

b2c3e87

fix keyword for new tests

fa79a9c

naik-aakash added 3 commits October 8, 2023 12:52

implemented robust orbital wise analysis

2aef248

make text description clearer for orbitwalwise case

c293cce

update tests and test files

41c6088

JaGeo reviewed Oct 10, 2023

View reviewed changes

lobsterpy/cohp/test/test_describe.py Outdated Show resolved Hide resolved

naik-aakash and others added 4 commits October 10, 2023 21:21

Merge branch 'JaGeo:main' into orbital_resolved_analysis

050d730

add missing space in text description and update corresponding test

100a39a

fix flake8 linting error

f9f3f3b

use numpy testing for array equality

c0b6625

JaGeo requested changes Oct 11, 2023

View reviewed changes

naik-aakash and others added 2 commits October 13, 2023 14:37

add more comments, fix typo : address comments

018aaed

Merge branch 'JaGeo:main' into orbital_resolved_analysis

cfded26

JaGeo approved these changes Oct 23, 2023

View reviewed changes

JaGeo merged commit 8e79b1c into JaGeo:main Oct 23, 2023
24 checks passed

JaGeo mentioned this pull request Oct 24, 2023

Orbital-resolved automatic analysis #113

Closed

naik-aakash deleted the orbital_resolved_analysis branch June 20, 2024 06:54

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Automatic orbital wise analysis functionality in analyze module #132

Automatic orbital wise analysis functionality in analyze module #132

naik-aakash commented Aug 2, 2023 •

edited

Loading

naik-aakash commented Aug 3, 2023

JaGeo commented Aug 8, 2023

naik-aakash commented Aug 11, 2023 •

edited

Loading

naik-aakash commented Aug 18, 2023

JaGeo commented Aug 18, 2023

naik-aakash commented Aug 18, 2023 •

edited

Loading

naik-aakash commented Aug 25, 2023 •

edited

Loading

naik-aakash commented Sep 7, 2023

JaGeo Oct 5, 2023

naik-aakash Oct 6, 2023

naik-aakash Oct 6, 2023 •

edited

Loading

JaGeo Oct 6, 2023

naik-aakash Oct 6, 2023

JaGeo Oct 6, 2023

JaGeo Oct 6, 2023

JaGeo Oct 5, 2023

naik-aakash Oct 6, 2023

naik-aakash commented Oct 6, 2023

JaGeo commented Oct 6, 2023

naik-aakash commented Oct 11, 2023

JaGeo left a comment

JaGeo Oct 11, 2023

naik-aakash Oct 13, 2023

naik-aakash commented Oct 11, 2023

JaGeo commented Oct 23, 2023

naik-aakash commented Oct 24, 2023

Automatic orbital wise analysis functionality in analyze module #132

Automatic orbital wise analysis functionality in analyze module #132

Conversation

naik-aakash commented Aug 2, 2023 • edited Loading

Added

Todo

naik-aakash commented Aug 3, 2023

JaGeo commented Aug 8, 2023

naik-aakash commented Aug 11, 2023 • edited Loading

NaCl

3s-3s (57%)

3px-3s (42%)

3s-4s (38%)

3py-4s (20%)

naik-aakash commented Aug 18, 2023

JaGeo commented Aug 18, 2023

naik-aakash commented Aug 18, 2023 • edited Loading

naik-aakash commented Aug 25, 2023 • edited Loading

naik-aakash commented Sep 7, 2023

Choose a reason for hiding this comment

Choose a reason for hiding this comment

naik-aakash Oct 6, 2023 • edited Loading

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

naik-aakash commented Oct 6, 2023

JaGeo commented Oct 6, 2023

naik-aakash commented Oct 11, 2023

JaGeo left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

naik-aakash commented Oct 11, 2023

JaGeo commented Oct 23, 2023

naik-aakash commented Oct 24, 2023

naik-aakash commented Aug 2, 2023 •

edited

Loading

naik-aakash commented Aug 11, 2023 •

edited

Loading

naik-aakash commented Aug 18, 2023 •

edited

Loading

naik-aakash commented Aug 25, 2023 •

edited

Loading

naik-aakash Oct 6, 2023 •

edited

Loading