Office Hours (2026-03-03)

🌟 Join Us for Weekly Office Hours! 🌟

Have questions about stdlib? Need help fixing a bug, figuring out what to do next, or just looking for feedback? Join our weekly office hours to connect with project maintainers, stay updated on the latest project news, and chat with other community members. This is a great opportunity to ask questions, share ideas, and engage directly with the stdlib team.

Everyone is welcome—drop in and say hello!

Time

UTC Tue 03-Mar-2026 17:30 (05:30 PM):

Timezone	Date/Time
US / Pacific	Tue 03-Mar-2026 09:30 (09:30 AM)
US / Mountain	Tue 03-Mar-2026 10:30 (10:30 AM)
US / Central	Tue 03-Mar-2026 11:30 (11:30 AM)
US / Eastern	Tue 03-Mar-2026 12:30 (12:30 PM)
EU / Western	Tue 03-Mar-2026 17:30 (05:30 PM)
EU / Central	Tue 03-Mar-2026 18:30 (06:30 PM)
EU / Eastern	Tue 03-Mar-2026 19:30 (07:30 PM)
Moscow	Tue 03-Mar-2026 20:30 (08:30 PM)
Chennai	Tue 03-Mar-2026 23:00 (11:00 PM)
Hangzhou	Wed 04-Mar-2026 01:30 (01:30 AM)
Tokyo	Wed 04-Mar-2026 02:30 (02:30 AM)
Sydney	Wed 04-Mar-2026 04:30 (04:30 AM)

or in your local time:

https://www.timeanddate.com/worldclock/fixedtime.html?msg=stdlib%20Office+Hours+2026-03-03&iso=20260303T1730

Links

Public calendar: https://calendar.google.com/calendar/embed?src=a72677fe2820c833714b8b9a2aa87393f742bcaf0d0f6c9499eee6661795eae0%40group.calendar.google.com

Joining the meeting

Google Meet link: https://meet.google.com/yct-chmu-owx

Agenda

Have specific things to discuss? Please comment in the issue thread below to add specific items to the agenda.

While working on feat: add blas/base/cgbmv #10492, I ran into a few points I’d like to confirm.

I want to confirm whether my band matrix packing is correct for M = 4, N = 4, KL = 1, KU = 2, (no-transpose).

For Column Major

"A_mat": [
  [ 1.0, 1.0, 3.0, 3.0, 6.0, 6.0, 0.0, 0.0 ],
  [ 2.0, 2.0, 4.0, 4.0, 7.0, 7.0, 10.0, 10.0 ],
  [ 0.0, 0.0, 5.0, 5.0, 8.0, 8.0, 11.0, 11.0 ],
  [ 0.0, 0.0, 0.0, 0.0, 9.0, 9.0, 12.0, 12.0 ]
],
"A_compact": [
  [ 0.0, 0.0, 0.0, 0.0, 6.0, 6.0, 10.0, 10.0 ],
  [ 0.0, 0.0, 3.0, 3.0, 7.0, 7.0, 11.0, 11.0 ],
  [ 1.0, 1.0, 4.0, 4.0, 8.0, 8.0, 12.0, 12.0 ],
  [ 2.0, 2.0, 5.0, 5.0, 9.0, 9.0, 0.0, 0.0 ]
],
"A": [ 0.0, 0.0, 0.0, 0.0, 1.0, 1.0, 2.0, 2.0, 0.0, 0.0, 3.0, 3.0, 4.0, 4.0, 5.0, 5.0, 6.0, 6.0, 7.0, 7.0, 8.0, 8.0, 9.0, 9.0, 10.0, 10.0, 11.0, 11.0, 12.0, 12.0, 0.0, 0.0, 0.0, 0.0 ]

For Row Major

"A_mat": [
  [ 1.0, 1.0, 3.0, 3.0, 6.0, 6.0, 0.0, 0.0 ],
  [ 2.0, 2.0, 4.0, 4.0, 7.0, 7.0, 10.0, 10.0 ],
  [ 0.0, 0.0, 5.0, 5.0, 8.0, 8.0, 11.0, 11.0 ],
  [ 0.0, 0.0, 0.0, 0.0, 9.0, 9.0, 12.0, 12.0 ]
],
"A_compact": [
  [ 0.0, 0.0, 1.0, 1.0, 3.0, 3.0, 6.0, 6.0 ],
  [ 2.0, 2.0, 4.0, 4.0, 7.0, 7.0, 10.0, 10.0 ],
  [ 5.0, 5.0, 8.0, 8.0, 11.0, 11.0, 0.0, 0.0 ],
  [ 9.0, 9.0, 12.0, 12.0, 0.0, 0.0, 0.0, 0.0 ]
],
"A": [ 0.0, 0.0, 1.0, 1.0, 3.0, 3.0, 6.0, 6.0, 2.0, 2.0, 4.0, 4.0, 7.0, 7.0, 10.0, 10.0, 5.0, 5.0, 8.0, 8.0, 11.0, 11.0, 0.0, 0.0, 9.0, 9.0, 12.0, 12.0, 0.0, 0.0, 0.0, 0.0 ]

I noticed in feat: add blas/base/sgbmv #5928 and feat: add blas/base/dgbmv #6121 that KL is not used in the same way as described in the intel/onemkl band storage reference for column-major and row-major layouts.
The current feat: add blas/base/sgbmv #5928 and feat: add blas/base/dgbmv #6121 tests only use KL = 1, KU = 1, and square matrices. Would it make sense to add tests with different KL and KU values and rectangular matrices to improve coverage?
While going through the complex BLAS routines, I noticed that there don’t appear to be explicit test cases covering the conjugate-transpose case. Although the implementation seems to handle this case, I couldn’t find dedicated tests that directly validate its correctness. Would you recommend adding specific test cases for the conjugate-transpose scenario?
how to properly generate a random complex banded matrix (e.g., for cgbmv) for examples/index.js. I saw examples for real banded matrices and complex dense matrices, but I’m unsure what the recommended approach is for complex banded storage. Should I use filled2dBy with Complex64 or construct the compact band array manually?

@DivitJain26

@DivitJain26 Have you already explored using our banded storage utilities for generating nested matrices: https://github.com/stdlib-js/stdlib/tree/develop/lib/node_modules/%40stdlib/array/base/banded/to-compact?

@DivitJain26

@DivitJain26 Re: 4. Yes, you need to add cases for the conjugate-transpose. That doesn't exist in the other packages because they are real-valued.

Re: 3. Sure. You'll likely want to follow the same JSON fixture based approach.

Re: 2. How does it differ? Can you show the difference in code and generated results? And does it differ than our banded/to-compact utility?

Re: 5. Yes, you can use with filled2by with Complex64. Likely no need to construct manually.

I would like to discuss about:

Scope of a GSoC Proposal to add Batch ML algorithms (@stdlib/ml/*). I am planning 1-2 algorithms per week as it can get quite huge (atleast that is what I understood in my initial trial of implementing dkmeansld). I would like to also discuss if it would be too less work from what is expected.
Possibility of implementing @stdlib/ml/dkmeans before writing proposal so that I could plan the scope of the proposal accordingly. (Getting this package merged will give me an exact idea on the expected API Design for rest of the packages)
What all implementations to be given more priority? Should I consider only the clustering algorithms for the proposal or should I consider some algorithms from each subsection (eg: clustering, regression, classification, dimensionality reduction)?

I would like to discuss. :

About integrating stdlib inside jstat, I have some initial idea which I want to discuss and want to know whether it aligns with the scope of project or not.

@DivitJain26 Re: 4. Yes, you need to add cases for the conjugate-transpose. That doesn't exist in the other packages because they are real-valued.

Re: 3. Sure. You'll likely want to follow the same JSON fixture based approach.

Re: 2. How does it differ? Can you show the difference in code and generated results? And does it differ than our banded/to-compact utility?

Re: 5. Yes, you can use with filled2by with Complex64. Likely no need to construct manually.

For Point 2.
Please see the changes in the second branch in
feat: add blas/base/cgbmv #10492 commit (5d4e1c8f73af76e93f643bcd5d8888cbbef97832)

Before

	for ( i1 = 0; i1 < ylen; i1++ ) {
		retmp = 0.0;
		imtmp = 0.0;
		oa2 = oa + ( i1 * sa1 );
		i0start = max( 0, i1 - KU );
		i0end = min( xlen, i1 + KL + 1 );
		for ( i0 = i0start; i0 < i0end; i0++ ) {
			kup1 = i0 - i1;
			ia = oa2 + ( ( kup1 + KU ) * sa0 );
			if ( i0 < xlen && ia >= 0 ) {
				ix = ox + ( i0 * sx );
				rea = viewA[ ia ];
				ima = sign * viewA[ ia + 1 ];
				rex = viewX[ ix ];
				imx = viewX[ ix + 1 ];
				retmp = f32( retmp + f32( ( rea * rex ) - ( ima * imx ) ) );
				imtmp = f32( imtmp + f32( ( rea * imx ) + ( ima * rex ) ) );
			}
		}
		iy = oy + ( i1 * sy );
		muladd( realpha, imalpha, retmp, imtmp, viewY[ iy ], viewY[ iy + 1 ], viewY, 1, iy ); // eslint-disable-line max-len
	}
	return y;

Now

	for ( i1 = 0; i1 < ylen; i1++ ) {
		retmp = 0.0;
		imtmp = 0.0;
		oa2 = oa + ( i1 * sa1 );
		i0start = max( 0, i1 - KL );                           // updated
		i0end = min( xlen, i1 + KU + 1 );                // updated
		for ( i0 = i0start; i0 < i0end; i0++ ) {
			kup1 = i0 - i1;
			ia = oa2 + ( ( kup1 + KL ) * sa0 );       // updated
			if ( i0 < xlen && ia >= 0 ) {
				ix = ox + ( i0 * sx );
				rea = viewA[ ia ];
				ima = sign * viewA[ ia + 1 ];
				rex = viewX[ ix ];
				imx = viewX[ ix + 1 ];
				retmp = f32( retmp + f32( ( rea * rex ) - ( ima * imx ) ) );
				imtmp = f32( imtmp + f32( ( rea * imx ) + ( ima * rex ) ) );
			}
		}
		iy = oy + ( i1 * sy );
		muladd( realpha, imalpha, retmp, imtmp, viewY[ iy ], viewY[ iy + 1 ], viewY, 1, iy ); // eslint-disable-line max-len
	}
	return y;

here the inner loop starts at max( 0, i1 - KL ), end at min( xlen, i1 + KU + 1 ) and ia = oa + ( i1 * sa1 ) ( ( i0 - i1 + KL ) * sa0 )
which is overlooked when KU = KL but it becomes critical when KU > KL., the previous bounds result in incorrect element access.

In the same PR the test cgbmv/test/fixtures/row_major_xpyp.json only passes after the mentioned changes. The test fails if the previous bounds and index access technique is used because here KU > KL (eg. KU = 2, KL = 1).

feat: add blas/base/sgbmv #5928 and feat: add blas/base/dgbmv #6121 uses the old pattern.

Review on my PR implementing dgetc2 LAPACK routine. I have made changes that were suggested for the other PRs I have made so I hope it should be in a good state.

I would like to ask about xorshift32 PRNG
Since it is single internal state :
why a separate seedLength property is exposed? whether this is mainly for API consistency with multi-state generators(MINSTD,MT)

@DivitJain26 Have you already explored using our banded storage utilities for generating nested matrices: https://github.com/stdlib-js/stdlib/tree/develop/lib/node_modules/%40stdlib/array/base/banded/to-compact?

Should I use this as a reference for making the fixtures?
if so then what cahges should i make in base implementation of feat: add blas/base/cgbmv #10492
It is working for my current packing pattern.

Review on PR #8187 - related to implementation of math/base/special/sici
Review on PR #10657 - related to implementation of math/base/special/expm1f

Do we have I1 Modified Bessel function of first kind in stdlib?

I have questions regarding implementation dnancumin.
Need clarification on the C doctest GSoC issue. Does eslint doctesting applies to README.md files too? I took a look at the eslint/doctest, can you guide me to the code where README files examples are also tested?
- For C code inside README, I mainly found two kinds of code snippet, one is annotation on the expected output and one is example.c. Can we execute them in a child process to check the output? Is this feasible?

I am working on my proposal notes for math/base/special, and I've mentioned some questions inside, can you please look into it?
https://befitting-tuberose-75b.notion.site/Base-Special-Proposal-Notes-301f09ba3c548095a4c4fd2a76ca67c3

Review on PR stdlib-js/stdlib#10620 - related to implementation of stats/base/dists/anglit/cdf
Review on PR stdlib-js/stdlib#10596 related to implementation of stats/base/dists/anglit/entropy
Review on PR stdlib-js/stdlib#9983 - related to implementation of stats/base/dists/bradford/logpdf

I also wanted to ask about the documentations to take reference from along with this https://github.com/scipy/scipy/blob/main/scipy/stats/_continuous_distns.py

I noticed that stats/strided/snanstdevpn and stats/strided/snanvariancepn are missing and depend on each other. They also require blas/ext/base/snannsumpw, which has an open PR this but marked as “needs discussion.”

How would you like me to proceed? If these get merged, we can also move forward with implementing the other similar stdev packages.

Sorry I shall not be able to speak but will be present in the meet, but if i could kindly have this clarified.

I was working on my proposal for the GSheet Integration repo and as per the last meeting i think my main goals will include the following:

Makefile targets to make development and adding APIs easier
Linting
Git Hooks
Adding all the missing APIs and making it up to date

supporting goals include:

developing the Sidebar for gsheets
contribution guidelines
technical tutorials

I shall be working on these, just need a nod of head so I can move ahead drafting it.

should i also include fused operations? if so any pointers would be helpful

thanks!

Thanks everyone!

Navigation Menu

Repository navigation

Office Hours (2026-03-03) #86

🌟 Join Us for Weekly Office Hours! 🌟

Time

Links

Joining the meeting

Agenda

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Office Hours (2026-03-03) #86

Description

🌟 Join Us for Weekly Office Hours! 🌟

Time

Links

Joining the meeting

Agenda

Activity

DivitJain26 commented on Mar 3, 2026

kgryte commented on Mar 3, 2026

kgryte commented on Mar 3, 2026

nakul-krishnakumar commented on Mar 3, 2026

shubham220420 commented on Mar 3, 2026

DivitJain26 commented on Mar 3, 2026

prajjwalbajpai commented on Mar 3, 2026

AyushiJain18270 commented on Mar 3, 2026

DivitJain26 commented on Mar 3, 2026

navyansh007 commented on Mar 3, 2026

manit2004 commented on Mar 3, 2026

Witty-31-06 commented on Mar 3, 2026

nirmaljb commented on Mar 3, 2026

bhargava-d16 commented on Mar 3, 2026

iampratik13 commented on Mar 3, 2026

GlyphicGuy commented on Mar 3, 2026

kgryte commented on Mar 9, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions