add probing code

[moneyball]

Can we add Mutiny's probing code? @benthecarman @TonyGiorgio

Or Matt's? @TheBlueMatt

Would like to add the simple prober code so that other LDK users can borrow from it to run their own (there should be at least one prober for each LSP used being run). I know of at least 2 projects that could use this right now.

[AnthonyRonning]

Our probing code has gotten significantly more complex since it basically has turned into a analytics platform for us to test algorithms and visualize results. On top of that it requires a web server to serve the results. Perhaps it's something to add to ldk-sample or spin off into a different project? There's nothing that needed to be added to rust-lightning to get it to work.

[moneyball]

Agree it doesn't need to be in the rust-lightning repo per se even though I filed the issue here. Probably a new repo in the LDK GitHub org with a forked version of ldk-sample. Or, an option added to ldk-sample?

[tnull]

Agree it doesn't need to be in the rust-lightning repo per se even though I filed the issue here. Probably a new repo in the LDK GitHub org with a forked version of ldk-sample. Or, an option added to ldk-sample?

I think the sample is really just a sample and shouldn't start to include production code. I think just adding some really naive probing code to an already-existing LDK node is trivial and should be in the order of ~10 lines that any user should easily be able to write. However, if we want to offer a more sophisticated background prober, it would probably warrant creating a new bare-bones crate based on LDK or LDK Node that includes a few more complex algorithms.

So either way we probably don't want to just add something to the sample real quick.

[moneyball]

Ok, let's not add this trivial 10 lines of code to ldk-sample or rust-lightning. Let's put it somewhere else in the LDK GitHub org. If it is so simple, this will be very easy for someone to do. I'll add as a good first issue.

[AnthonyRonning]

Would it be okay to take LDK-sample as a starting off point and add a "probe" command that does the work? Our probing code is basically LDK sample with a lot of additional code, SQL database, authenticated API endpoints, analytics, etc.

But most of our complexity wouldn't be needed to do the same probing as us so perhaps I can make a minimized starting off point. It started off simple but a few things we learned doing background probing that we can contribute to.

[TheBlueMatt]

If we want to include something in the lightningdevkit repo, we should probably do something smart, like actually picking target paths that we have less information on rather than just trying the same path over and over again, but the naive approach at https://git.bitcoin.ninja/index.cgi?p=ldk-sample;a=commit;h=f47b6f50b85ef3770173cc8e4bcddd3662cb5d5e (just pick a random target and a random amount once a minute) seems to learn pretty well in spite of its naivete.

[007harshmahajan]

Well i had a unsucessful but enlightning experience with probing, now I have depth with general probing and rust, can i give it a shot ?

[TheBlueMatt]

Sure, if you want. I think, to start, we'd like to have a single function which, given a target node to probe, picks a path. As an initial go, we probably want to "just" call our router, but rather have a separate mode on our ProbabilisticScorer. In this mode, rather than scoring based on probability, we should score based on probability (the existing metric), but also avoid any channels we've "recently" probed (with some heuristic for recent that falls off with time-since-probe).

Concretely, we'll probably want to just have a wrapper struct that wraps ProbabilisticScorer and implements ScoreLookUp itself.

[Geoelycom]

is this still being worked on by you @007harshmahajan

[007harshmahajan]

is this still being worked on by you @007harshmahajan

not now

[Geoelycom]

Can I tackle it instead??

…

On Mon, 5 May 2025 at 10:58, Harsh Mahajan ***@***.***> wrote: *007harshmahajan* left a comment (lightningdevkit/rust-lightning#2720) <#2720 (comment)> is this still being worked on by you @007harshmahajan <https://github.com/007harshmahajan> not now — Reply to this email directly, view it on GitHub <#2720 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AM4GCI3S5UPEPEYK3NHM6ZT244Y45AVCNFSM6AAAAAB4NKRXMCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDQNJQGQ4TSNRSGI> . You are receiving this because you commented.Message ID: ***@***.***>

[007harshmahajan]

Can I tackle it instead??

On Mon, 5 May 2025 at 10:58, Harsh Mahajan @.***> wrote:

007harshmahajan left a comment (#2720)

#2720 (comment)

is this still being worked on by you @007harshmahajan

https://github.com/007harshmahajan

not now

—

Reply to this email directly, view it on GitHub

#2720 (comment),

or unsubscribe

https://github.com/notifications/unsubscribe-auth/AM4GCI3S5UPEPEYK3NHM6ZT244Y45AVCNFSM6AAAAAB4NKRXMCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDQNJQGQ4TSNRSGI

.

You are receiving this because you commented.Message ID:

@.***>

yes go ahead with it 🙌

[Geoelycom]

Ok

…

On Mon, 5 May 2025 at 11:31, Harsh Mahajan ***@***.***> wrote: *007harshmahajan* left a comment (lightningdevkit/rust-lightning#2720) <#2720 (comment)> Can I tackle it instead?? On Mon, 5 May 2025 at 10:58, Harsh Mahajan *@*.***> wrote: *007harshmahajan* left a comment (#2720 <#2720>) #2720 (comment) <#2720 (comment)> is this still being worked on by you @007harshmahajan <https://github.com/007harshmahajan> https://github.com/007harshmahajan not now — Reply to this email directly, view it on GitHub #2720 (comment) <#2720 (comment)> , or unsubscribe https://github.com/notifications/unsubscribe-auth/AM4GCI3S5UPEPEYK3NHM6ZT244Y45AVCNFSM6AAAAAB4NKRXMCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDQNJQGQ4TSNRSGI . You are receiving this because you commented.Message ID: *@*.***> yes go ahead with it 🙌 — Reply to this email directly, view it on GitHub <#2720 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AM4GCI6F5DWYRIRGX3CPCBD2444WJAVCNFSM6AAAAAB4NKRXMCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDQNJQGU3TAOBYG4> . You are receiving this because you commented.Message ID: ***@***.***>