Repository navigation
Update showcase notebook - #345
hannahbaumann wants to merge 13 commits into
Conversation
|
Check out this pull request on See visual diffs & provide feedback on Jupyter Notebooks. Powered by ReviewNB |
|
|
|
View / edit / reply to this conversation on ReviewNB hannahbaumann commented on 2026-10-23T11:43:28Z I added all this background from the openfe demo and the ccp bio sim tutorial, please let me know if you think this is too much for the showcase! jthorton commented on 2026-10-26T15:07:42Z While I do like it maybe we are better of pointing the protocol page in the user guide so we only have a single source to keep up-to-date with how the protocols work? And here assume some knowledge of the calculations and really show off the API and the interchangeable components stuff? IAlibay commented on 2026-10-27T09:08:22Z Since it's the main thing folks usually look at, I'm actuallly pro having "more information". I understand the need for reducing work when doing updates, but it's also the showcase so it's ok to have this be the main place where we have to be vigilant in updating things. |
|
View / edit / reply to this conversation on ReviewNB hannahbaumann commented on 2026-10-23T11:43:28Z I thought it would be easier to just create all of them, and put the detailed ChemicalSystem creation and transformation creation into a cookbook? jthorton commented on 2026-10-26T15:42:50Z Yeah I think this looks good we could link to the components user guide as well https://docs.openfree.energy/en/stable/guide/setup/chemical_systems_and_thermodynamic_cycles.html hannahbaumann commented on 2026-10-28T09:34:47Z Added the link in the info box below. |
|
View / edit / reply to this conversation on ReviewNB hannahbaumann commented on 2026-10-23T11:43:29Z I think the full explanation on how to run things in Python should probably be in the RBFE Python API tutorial, but I wasn't sure. jthorton commented on 2026-10-26T15:47:39Z Yeah that makes sense validate and create are probably better suited to this showcase, maybe we should introduce validate here as well with some examples of things it can catch. We should probably show validating a protein component as well? IAlibay commented on 2026-10-27T12:25:42Z Dry run is a bit of a unique thing we've implemented for hannahbaumann commented on 2026-10-28T09:49:30Z Changed this to showcase validate instead. |
|
View / edit / reply to this conversation on ReviewNB hannahbaumann commented on 2026-10-23T11:43:30Z Also here I thought maybe the analysis Python API part should be in the RBFE python API tutorial? IAlibay commented on 2026-10-27T12:27:51Z I think it's ok to have some redundancy - this notebook tends to be "the thing folks look at", so some small bit about how to analyze results is ok. |
|
View / edit / reply to this conversation on ReviewNB hannahbaumann commented on 2026-10-23T11:50:15Z Will need to update this link when the PR is merged (also other links below are still WIP. IAlibay commented on 2026-10-27T09:13:13Z Here and elsewhere, how about using info boxes when linking to other areas - e.g. for the "openfe does not prepare structures". That would make it easier for folks to easily recognise when we're signposting. hannahbaumann commented on 2026-10-28T08:43:08Z Put this into a box, also the protocol tutorials in the intro. IAlibay commented on 2026-10-30T11:35:51Z Thanks - I noticed reviewnb doesn't show boxes, so apologies if I asked for some in places where there were already! |
|
View / edit / reply to this conversation on ReviewNB atravitz commented on 2026-10-26T12:52:42Z I think this is great, and also makes me that we should maybe be noisier about using duecredit when citing openfe. jthorton commented on 2026-10-26T15:12:15Z Lets pull this into the user guide for the protocol and link to it, having a single source to keep up-to-date would be easier but I agree this is all great info which is missing from the guide. The credit to perses and OpenMM tools should probably go into both though for maximum publicity?
Then I would just have a small details list of the protocol here covering key concepts. IAlibay commented on 2026-10-27T09:11:56Z I don't think I agree with just pulling all this into the user guide and removing it from here. As mentioned above, this quite often the first (and only) place folks look at - I'm pro being very noisy and detailed rather than assuming folks will read the userguide first. In fact, we usually tell folks "if you do anything, look at the showcase". |
|
For the analysis parts its probably tied to the simulation health check page we want make which applies to most protocols. We should advertise some of the analysis in the showcase and link to the protocols for more info and then to the health check page. The API tutorials for each protocol should have examples on accessing the data used in the analysis, though. |
|
While I do like it maybe we are better of pointing the protocol page in the user guide so we only have a single source to keep up-to-date with how the protocols work? And here assume some knowledge of the calculations and really show off the API and the interchangeable components stuff? View entire conversation on ReviewNB |
|
Lets pull this into the user guide for the protocol and link to it, having a single source to keep up-to-date would be easier but I agree this is all great info which is missing from the guide. The credit to perses and OpenMM tools should probably go into both though for maximum publicity?
Then I would just have a small details list of the protocol here covering key concepts. View entire conversation on ReviewNB |
|
Yeah I think this looks good we could link to the components user guide as well https://docs.openfree.energy/en/stable/guide/setup/chemical_systems_and_thermodynamic_cycles.html View entire conversation on ReviewNB |
|
Yeah that makes sense validate and create are probably better suited to this showcase, maybe we should introduce validate here as well with some examples of things it can catch. We should probably show validating a protein component as well? View entire conversation on ReviewNB |
|
View / edit / reply to this conversation on ReviewNB jthorton commented on 2026-10-26T15:58:33Z
Can we put the available protocols back into the grouped list we had? I think it's clearer to see the individual links. IAlibay commented on 2026-10-27T09:05:58Z +1, I would also add back in the link to the tutorials. hannahbaumann commented on 2026-10-27T13:28:52Z Changed this back into the list with the links. |
|
View / edit / reply to this conversation on ReviewNB jthorton commented on 2026-10-26T15:58:34Z points 10-12 are not in the list. hannahbaumann commented on 2026-10-28T09:55:08Z I think this may only be a reviewnb issue. IAlibay commented on 2026-10-30T11:32:12Z Looks like github renders this list fine - I also noticed there were other things in reviewnb that didn't show up, like the blue boxes D: ! |
|
View / edit / reply to this conversation on ReviewNB jthorton commented on 2026-10-26T15:58:35Z I like the figure but maybe it needs updating having alchemiscale is probably confusing at this point, it should have quickrun and task-based execution instead.
The headings in the list should be the same as the figure as well to make mapping them easy for the reader so 1. Plan -> 1. Setup hannahbaumann commented on 2026-10-28T08:37:22Z I started drafting a new workflow diagram here: https://docs.google.com/presentation/d/16afpfvu_UD9x56UleVTRk39hYKlg-g1CFF-pCi_SNVE/edit?slide=id.p#slide=id.p Please let me know what you think! IAlibay commented on 2026-10-30T11:35:01Z Happy to go as-is if needed, i.e. take my comment as nits.
I'm only really 50/50 on this, so would be keen to hear what Josh & Alyssa think. I think maybe I'd remove the API/CLI complication and just show "what are you doing here". Or maybe even more conceptual things like "what am I doing in each stage". atravitz commented on 2026-10-30T15:03:06Z I think a modified figure of the one above (I made it originally, and have an updated version) is good for explaining our software stack, and might be good outside the showcase.
I like hannah's workflow diagram proposal and am adding comments/suggestions over on that doc. |
|
View / edit / reply to this conversation on ReviewNB jthorton commented on 2026-10-26T15:58:36Z How does it look with some binding site side chains shown, you can't really see any interactions from that 3D view IAlibay commented on 2026-10-27T09:13:58Z If this is from the ccpbiosim tutorial - adding the side chains was a real mess / too hard to see anything. hannahbaumann commented on 2026-10-28T08:49:04Z I tried this out below with the binding site residues, what do you think is better? IAlibay commented on 2026-10-30T11:36:33Z I'm happpy with either. |
|
View / edit / reply to this conversation on ReviewNB jthorton commented on 2026-10-26T15:58:37Z I think it would be good to keep some of the interop examples with OpenFF. hannahbaumann commented on 2026-10-28T08:54:08Z Added this back in! |
|
View / edit / reply to this conversation on ReviewNB jthorton commented on 2026-10-26T15:58:38Z This is great! Maybe as the reference for the 2.3.0 AshGC paper, this might be some users first exposure to it. hannahbaumann commented on 2026-10-28T09:10:42Z Added the citation. |
|
View / edit / reply to this conversation on ReviewNB jthorton commented on 2026-10-26T15:58:39Z This is the default. -> This is the CLI default as we have no API default for the mapping functions? IAlibay commented on 2026-10-27T09:18:44Z Maybe also add the phrase "in our hands this performs the best, especially when you have well aligned ligands"? hannahbaumann commented on 2026-10-28T09:14:30Z Modified this accordingly. |
|
View / edit / reply to this conversation on ReviewNB jthorton commented on 2026-10-26T15:58:40Z Is this just the notebook issue but for me all the ids are on a single atom? hannahbaumann commented on 2026-10-28T09:15:24Z In the jupyter notebook this looks normal, not sure if maybe this is just a reviewnb problem? |
|
Modified this accordingly. View entire conversation on ReviewNB |
|
In the jupyter notebook this looks normal, not sure if maybe this is just a reviewnb problem? View entire conversation on ReviewNB |
|
Changed this! View entire conversation on ReviewNB |
|
Added something. View entire conversation on ReviewNB |
|
Added this back in. View entire conversation on ReviewNB |
|
@IAlibay Did you mean start switching to json or graphml? View entire conversation on ReviewNB |
|
added! View entire conversation on ReviewNB |
|
Added the link in the info box below. View entire conversation on ReviewNB |
|
Changed this to showcase validate instead. View entire conversation on ReviewNB |
|
Added this. View entire conversation on ReviewNB |
|
View / edit / reply to this conversation on ReviewNB hannahbaumann commented on 2026-10-28T09:51:21Z Replace this with @jthorton new figure! |
|
I think this may only be a reviewnb issue. View entire conversation on ReviewNB |
|
Looks like github renders this list fine - I also noticed there were other things in reviewnb that didn't show up, like the blue boxes D: ! View entire conversation on ReviewNB |
|
Happy to go as-is if needed, i.e. take my comment as nits.
I'm only really 50/50 on this, so would be keen to hear what Josh & Alyssa think. I think maybe I'd remove the API/CLI complication and just show "what are you doing here". Or maybe even more conceptual things like "what am I doing in each stage". View entire conversation on ReviewNB |
|
Thanks - I noticed reviewnb doesn't show boxes, so apologies if I asked for some in places where there were already! View entire conversation on ReviewNB |
|
I'm happpy with either. View entire conversation on ReviewNB |
|
Sorry I meant we should start switching to View entire conversation on ReviewNB |
|
View / edit / reply to this conversation on ReviewNB IAlibay commented on 2026-10-30T12:29:08Z [nit] It might be good to highlight trajectory analysis and debugging separately as "things you would do normally but we don't do here". |
IAlibay
left a comment
IAlibay
left a comment
There was a problem hiding this comment.
I just had the one thing left (alongside maybe going to to_json for ligand networks).
I think as-is, this represents a good improvement on the, so I'm going to approve with the idea that if you merged today that'd be ok for me.
|
I think a modified figure of the one above (I made it originally, and have an updated version) is good for explaining our software stack, and might be good outside the showcase.
I like hannah's workflow diagram proposal and am adding comments/suggestions over on that doc. View entire conversation on ReviewNB |
|
I agree, but we should support both. I want to resolve OpenFreeEnergy/openfe#2323 (which is straightforward) then we can switch this.
View entire conversation on ReviewNB |
I added links to different places in our docs in this notebook, however some of those pages are still being written, so we'll need to update the links and make sure they all work later.
There are several open issues here related to the showcase that I think we should decide on here, implement if still desired and close if not: