 Okay, so hopefully you can see my presentation now. Yeah, so I think I've met most of the people here, or at least I recognise most of the names, but if we haven't actually spoken, I'm Eli, my pronouns are he, him and they, them. I'm the one who's been doing most of the technical development on Muon Galaxy, which is our work to extend Muon, extend Galaxy to the Muon Science domain. The team currently is me, Patrick Austin, Deandre Laborio and Alejandro Gonzales Beltran. We've had a few other people work on the project, Anish Jotish and Simone, who've now left the team but have contributed in some way, so I thought I'd acknowledge them here. Every disclaimer, I forgot I was giving this talk until Monday until I was reminded and I threw these slides together this afternoon so they're a bit disjointed and dense and I apologise in advance. And so my team is based at the STFC Rutherford Ableton Laboratory in the UK. It's home to the ISIS Neutron and Muon Source, which is one of only four Muon spectroscopy facilities in the world. And so beam time is quite limited and expensive as well. The team is based in the Scientific Computing Department, which develops computational tools for the experiments at RAL, including ISIS, and also for the wider academic community. So we live in this little building here or some of us live in a little building that's not even in this picture these days. So a bit of background on what Muon Spectroscopy actually is. Basically it's a way of studying the properties of a material of interest. So the material has a particular structure and you make a little target and you just fire a beam of Muons at it. And Muons are positively charged particles that are about one-tenth of the mass of a proton. So they behave a little bit like protons when you put them in a material. So you fire your Muons at your sample and the Muons get sort of implanted in the sample somewhere. And that somewhere is dependent on what the material, what elements are in the material, the structure that it has various properties. And the place where the Muon stops is known as the implantation site or the stopping site. After an average of 2.2 microseconds, the Muon will decay into a positron. The positrons then ejected from the sample and detected with a positron detector. So by detecting these positrons you can make calculations to figure out where in the sample the Muon was and things like what spin it had and stuff like that. And that all gives you information about the material properties. Now we work specifically on Muon simulations. So I mentioned the stopping site. This is a really common thing that people want to learn or find for their particular material of interest. And simulations can help to make predictions about where the stopping sites might be in a structure, make predictions about material properties based on that. And if you've already got some experimental data, simulations can enhance your analysis of it. And overall, this just allows more efficient use of ISIS beam time. On the right here is an example of crystalline silicon. This has two Muon stopping sites. One is sort of in the middle of the tetrahedral structure. And one is right in the middle of a bond between two silicon atoms. So our project is called the Muon spectroscopy computational project. We're creating a sustainable, accessible and open source ecosystem of tools for Muon science. These include by Muon Suite, Mu Spin Sim and Mu Dirac, which are all command line tools for modeling Muon interactions in various ways. And I'll skip over the details of what each of them does. But Muon Galaxy is an STFC hosted Galaxy instance that is going to include the tools above. So we've already implemented some of these. We've still got some to do. And I guess that's the focus of this talk really. And our instance is available at muongalaxy.stfc.ac.uk if you want to go and look at it. So here's an example of the kind of workflow that you might want to run if you want to find Muon stopping sites and then do extra work based on that. You start with a structure file that represents your structure and your target structure. You basically make, you can make like a hundred copies of it and in each one of those copies put a Muon in a random location in the structure. And once you've done that, you take all those copies and you relax them to find the lowest energy state. And then you run a clustering algorithm to say are there places within this structure where the Muon tends to, I want to say gravitate towards, but it's obviously not gravity, but it's a lowest energy state. So by analogy, gravitate towards. Yeah, so there's likely to be locations within your structure where the Muon tends to end up and and those are likely to be your stopping sites. Once you have that information, you can do a bunch of post processing and using some of our other command line tools. The key thing here is that we want to get this workflow into Galaxy and as much of it as we can by building tools for each of these steps. So that's what we're doing. So we've got quite a few tools now for different stages of this workflow. We're also building workflows that chain them all together. So this is that generating the random structures, relaxing or optimizing them to get the lowest energy state and then clustering them to slightly so before and with various input files to facilitate that. And we're going to be expanding these once we as we develop more more tools as well. And we're also working on a visualizer using a package called chris fizz dot chris fizz js that was developed within our group. This has been on the back burner for a while. We did a very basic integration, but didn't wire up all the interactive bits. And so that that needs some work at some point, but right now it's sort of been deep prioritized as we focus on on the tools and the workflows. And so a bit of the timeline of how this will come together. And I've included this for the benefit of the astronomy community who I think is just starting to look at Galaxy. I don't know if any of them are here today, but I'll send them the recording anyway if they're not. And so we had our first meeting in with the Galaxy team in March of 2021. And after that I started to experiment set Galaxy up on a VM play with it, build some prototype tools and stuff. And I ended up giving a talk at the Galaxy Community Conference in 2021, where I basically said what we were trying to use Galaxy for and then spewed out a bunch of questions. A lot of people gave me answers to that questions and I realized how knowledgeable and supportive the Galaxy community is. And I think this was sort of the event that really sold me on the idea that Galaxy was a really good platform to use for what we were doing. So after that we started tool development in earnest and set up a development server to host those tools in a place where other members of the team could play with them. And then as I did so and as we got more ideas about how we could connect Galaxy to different bits of infrastructure, I realized that I would really benefit from learning how to do that. So I went to the Galaxy admin training at GTN Tapas, which was a really great event, very valuable, gave me lots of ideas of how we can expand me on Galaxy in the future. In May of this year, we also set up the Material Science sub-community and sub-domain in Galaxy Europe, which is still under construction and is basically just me at the moment, but it's there and we want to expand it. In August, we finally launched our production server to the public and we put our tools on the tool shed. And then in September, we presented our work at three different conferences just to try to disseminate as much as possible to people who might be interested in using it or in developing more material science tools for Galaxy. Along the way, we've done a bunch of collaboration with the wider community around Galaxy, so we added some new data types to the Galaxy codebase. I raised some accessibility issues in the new Galaxy history, which other people then also jumped onto in which I think are being worked on at the moment. We've reported various bugs and they've often been fixed really quickly, so thanks. We received lots of support as well with code review and tools, thanks especially to Björn, who I don't know how you get to the tool, the pull requests on our tools repository quicker than I do, but I appreciate it. As I mentioned, we have Materials Galaxy under construction at materials.usegalaxy.eu. The idea with this is that there are a few other people looking at using Galaxy for things like neutron science and we are keen to collaborate with them and maybe have a unified instance where we can have all those tools on there. And recently we've kicked off the Hero Science Gateway project, so we in the material science community are part of WorkPackage 5, which is focused on community engagement, adoption and onboarding. So this is going to be great for just knowledge sharing with other communities that are also new to Galaxy. What we've got upcoming, more tools and workflows, as I mentioned, we're going to develop some tutorials for the Galaxy training network and then try and connect things to other STFC infrastructure. We've got lots of stuff at STFC that we'd love to connect to, one is our HPC cluster, there are various experimental data stuff, but we haven't had those conversations yet, so we don't quite know what's possible. But the stuff is there, we just need to figure out how and if we can access it. There's also a number of material science databases that we're interested in, in connecting to and providing a sort of reference data. And some, again, some within STFC or contributed to by STFC and some completely outside it, and lots of general outreach within the material science community, including within our own facility. We've already had some chats with people who work on in separate groups, but on other tools, but who are interested in integrating them into Galaxy and now they've seen what we're doing. So that's good news. And some of the lessons learned from this whole experience, firstly, that the Galaxy community is really supportive, I found it really good to be part of. You've got so much experience to draw from and you're really open to new collaborations, which has been even very welcoming of us as a new community, which has been great. Working within the Galaxy platform has meant less overhead for us in terms of UI and more time building and improving our tools. And the reproducibility features of Galaxy actually encouraged us to make improvements to the rest of our packages in terms of sustainability. So, and wanting to use these packages in Galaxy tools meant that we had to set up PIPI and condo releases and submit things to the bio containers repository and set up continuous integration pipelines so it's been it's been a net positive for our sustainability in general as well. There have been various challenges and since this is sort of a quite developer heavy, heavy space I thought I've mentioned some of them. And one of the ones I forgot to put in is just like, when you get started it can be quite overwhelming. And there's a lot of things like biology focused and if you don't know biology it can be hard to like, and follow some of the tutorials, just because it's a bit too much. But I kind of forget that because I've been doing this for a while now. And these are things that have come up a bit more recently. First that I realized recently there's two sources of tool best practices one is in planima ones that the IUC, and both of them are missing some things. I think this, this could be improved I think beyond raised an issue on our tools repository pointing out a bunch of the things we could, we could improve on, and I think a lot of those are documented anywhere else so that would be useful to have. And I mentioned we set up a materials that use galaxy.eu subdomain. I found this process quite confusing. And it seemed to be a case of ask someone what to do, rather than I think he could do with some more documentation for this especially as he seemed to me to make pull requests against multiple different repositories to get everything set up. And in terms of administration, the galaxy admin training is fantastic. I've been super impressed by it. And but it is tough to go beyond it. And I have rarely needed to do so but there's the occasional case where I wanted to configure something very specific and I've tried to find it in places like use galaxy and playbooks and such, but it's really hard to use those as a reference partly because it's so big, partly because it's split amongst multiple things and it just makes it really hard to search, even if I'm trying to find like a particular keyword. I don't know which repository it's in so it's hard to find it. And I don't quite know how I would suggest improving that. And I wanted to mention it. And right now we're struggling with upgrading to 22.05 and it doesn't work. I'm not sure why it might be specific to our infrastructure. And, but also it seems that the way to get updated configuration is just to go do the galaxy admin training again or at least look through that page and notice any changes. It would be nice if there was a better way to do that maybe something, a page that just sort of summarizes what's changed about sort of the recommended installation, and rather than having to look through all the galaxy admin training tutorials. And finally a couple of open questions that perhaps we could discuss now. One is around how to keep up with significant galaxy decisions and developments without attending tons of meetings. Ideally, and I found that being sort of on the periphery, some things change while I'm still trying to develop against them. And I don't know that that's happened. So this happened with them the existing visualizer plugins the JavaScript got rewritten a bit or restructured a bit. I didn't realize this happened and then I came back to our visualizer and it just didn't work. And I had to dig around to figure out why. I think some of these decisions and developments are contained in things like the release notes. And, but when a release is delayed by months that becomes less practical to keep up with those. And if there's just a repository where there are loads of notes on this sort of thing. I haven't found it. If there is one please tell me where it is. And because I think sometimes it just feels like I'm very out of the loop on stuff. But I don't want to have to attend like tons of meetings to try and get more in the loop. And the second one is about how to build up our user and developer community. So there's a lot of people. I think most of them haven't heard of Galaxy at all. And unless we go really into depth in explaining what Galaxy is in terms of, it's like a platform for fair data analysis, not just a GUI. People think it's just yet another GUI that they're going to have to learn or that will die out immediately. Yeah. And so I guess that's something we can draw from more experienced communities. How do you sort of make those first steps? And yeah, thank you for listening. And you can contact me at Eli dot chapic at stfc.ac.uk, but I am going to be absent for all of November and December. So Patrick Austin, I don't know if he's here is going to be covering for me. And you can contact him at Patrick Austin at stfc.ac.uk. Thanks everyone. Take questions I see Martin's got his hand up. Awesome talk Eli. Thank you very much for sharing that. Can we go back one slide to your open questions. So I was curious. So the first bullet point here. Is this like on top of the release? Did you hear me? I think you dropped out for a second. I'm sorry. Is this question on top of the release? Yeah, so they said like I can read through the release notes, but like 22.05 was a release that took eight months. And so a lot happens in that time and like there's so many changes on the repository every day that I can't keep up with it all. And yeah, it feels like sometimes. I guess I'll ask about something and the answer will be, oh, we changed this a while ago. And maybe it's just that I'm not, I don't know. I must, if I might add to this also a great talk. Thank you for that. And regarding these JS visualizers, so our JavaScript plugins currently do not particularly the external visualization plugins. We don't have proper tests for them. So hopefully when in the next releases we have testing for these plugins, then these things won't fall through the gaps anymore. So if there are changes during the release cycle, people would see, oh, look, these tests are failing. And hopefully we can avoid that. But I totally understand your frustration there. If there are no tests for the visualizations make changes. Some are not properly adjusted to it because they're not tested. So we don't detect that. And that can cause problems. But hopefully that'll get better soon. Thanks, that sounds good. Yeah, I think it's like the modularity of Galaxy is really useful in the sense that we can just plug in a new visualizer or plug in tools that we've written. But then it means that, yeah, they're not covered by any tests because it's a new file sort of thing. And yeah, I remember like tool best practices suddenly changing as well. Suddenly from our perspective suddenly we suddenly discovered the change and had to rewrite a bunch of stuff. And like our tool tests as well. So hopefully those kinds of changes should be pretty rare. I'd be interested in following up on specifically the visualization issue after this. And figure out what what change caused issues for you because we do try to make sure that stuff remains backwards compatible like even the old Mako based visualizations should still for the most part work, unless they were using Galaxy internals right that we can't necessarily maintain forever and that's why we've tried to establish a better interface there that that's a little more stable. Yeah, this was a little while ago but I think I remember at least what the commit that caused problems was like trying to go back and find that. Yeah, that'd be great. Thank you. Eli. So, first of all, yes, very great talk and very, very useful to us, especially, especially all the details about the frustrations and the difficulties going from release to release. This is this is helpful. So what you are saying is with regard to the changes. What you're saying is with regard to the changes to set up an installation that happened after a given release. We tend to include them all sort of grew them all together in release notes so the release notes will contain stuff that has changed in terms of here is what you need to change in your setup plus here is this new great feature etc etc. And I think that is that it would be helpful to have a designated section or page or whatever which would be specifically for admins what you need to change to make this release work right. Yeah, definitely. Yeah, because like, I think I've got like good attention to detail and I can read the Galaxy admin training tutorial for the Ansible installation and notice what's changed from before. Most of the time if it's like one line then probably not but if it's like a chunk like it was this time with a change to the unicorn. I can sort of spot those things. But I don't want to have to do that with 10 tutorials and I don't think you can expect everyone to go through and try and sort of filter out the whole thing this time. Each time, or try and read like the markdown diff or whatever. So yeah having a dedicated page that comes out with the release notes saying this is what we with is what changes we recommend for an Ansible configuration would be great. Okay, we'll try to improve that. That was very quiet, but yes, go ahead. I have two questions. One for the community so in the galaxy user interface they are still a little bit of genomics included. How big. I mean is it a big problem for you or is it okay. And the second question. You as an administrator that deploy a galaxy instance. Do you think that two releases a year would help you. Or are three releases okay. Or are you actually skipping releases. I'll answer the releases one first I think because we've only I think we set up our servers properly for the release. And then there's only been one release since that so I wouldn't I don't think I've got enough experience with like regularly upgrading for releases to comment on that. And the other question was about. Yeah, can you be more specific about like what do you mean with the galaxy you are including genomic stuff. Cool. I mean if you haven't recognized that that's maybe the I have but I was wondering if you think you have a specific area. No, I mean that there are some very tiny details like the DB key that you need to set for the data type or for data sets. And there's a reference genome I think still in the upload dialogue. I'm just wondering how big this problem is to communities that are not from life science. I mean, we have the assumption that this might be a big problem but we don't know so that's why I'm asking how much of a problem is that actually. Yeah, so I'd say the most places I've noticed it most have been in the sort of tutorial material and like the like interactive tour of galaxy. There's a few bits in that that like very, like reference very specific genomics or biology topics and it's sort of like oh that was that come from am I supposed to know what this is. And it does I guess it makes it feel like it's, it's for that and not for material science. And, and in a way that like, I feel like most of it could be made, especially in the interactive tools could be separated out from the biology stuff. I guess there are like specific integrations of things that we don't strictly need. Like in the tool developments you have like the e dam ontology, which we don't really use. So, I mean, we can just ignore that. But yeah, if there's sort of an expectation if we're contributing to things like bio containers for example or workflow hub, if there's an expectation if those are built with the expectation that you're going to be uploading biology stuff to them. And then we come in with stuff that doesn't fit into that domain. That would be confusing. I think within galaxy itself. I haven't encountered too much that's been like, Oh, this is really genomics heavy as a feeling. But we'll probably have to check that with users and get their feedback as well. Cool, that's super helpful and since you mentioned interactive tours. Do you use them a lot. Is that something that we actually should more improved. For example, we changed the one on one galaxy training with and to be not genomic focus right there now to one on ones in the training material, one that is still a bit genomics and one that is completely generic. And we can do these things for the tours for example as well. For the galaxy one on ones. It's, well I guess it's cool galaxy one on one for everyone but it's still like a biology example but yeah you're right it's it's a bit more it's a bit easier to follow. I, so we're not using interactive tools specifically at the moment, we've got the visualizer but then we just got I can't regular regular galaxy tools. I can't remember what your question is actually about on the interactive tool side of things. No, that's okay. Cool. Any more questions. I have a comment on the second question that you have here. So we were discussing the other day during the course of the European galaxy days to have something like community page so that would be like, same as you have your sub domain now for the material science have something like that but in the future just for specific communities as applications of galaxy. This is something that could be helpful for you to reach out to people. Yeah I think so. And again like that there's like the galaxy get started page which is nice to have but is a bit overwhelming and quickly goes quickly sort of fears you into biology space again so I think having something sort of community focused. Yeah, we're, we're looking at building tutorials with like some simple examples from material science space using our tools and BM and more like general material science and truth might also be be worthwhile. Especially if we start expanding beyond just me on science to things like neutron science and other things as well. I guess that the question is for me it's like the opposite is the question for you how would you like to be reached out or how would you reach out how to reach people like you. Yeah, so there's a few like very. Like niche specific conferences like the one that's the three conferences we present that sort of run the gamut of places we want to advertise. One is like a very specific me on science conference. There was like a general research software engineering conference, and one was focused on and like software for like neutron and neon and x-ray science experiments so that last that last one was when we really targeted as like these are people that would want to develop the galaxy if they were interested in it or administer their own instance. And, but then the. So that was yeah more of a developer focus conference and be I guess we have to go to like the really specific domain science conferences and try to get people on board with this thing. Yeah, like me or science conference or like a neutron science conference but there's a lot of those around and I'm not the one who knows which ones are good to attend Leander is the one that knows that. Maybe just to add a sign here and thanks Eli for for the presentation but I mean we we are based on a national laboratory where are there are lots of people working on and so so we have started this slow process of talking to people locally. So I think in addition to the conferences which is very important and Eli has been presenting in different places, as well as others in in the group. Yeah, I think it is important to get this one to one conversations with people working in the domain what now we have also a challenge ahead so now that we have the tools into galaxy we need to get users on the million community to to try them out and see how it works for them. Well, as I said talking to we are talking to people in slightly different domains to promote adoption. So, I think I mean actually we discussed in the in the project kickoff meeting and I think this could be the kind of thing it would be great to have more domain agnostic starting guys and this is what they I think and or you had come up with his name on the propeller so we are quite keen on on starting something up so I was wondering again if what's the procedure here so if we if you create a page within the training material that's that's perhaps Eli can stop putting some thoughts into and then of course it will be part of the project as well and it could be open of course to the whole community I imagine I think the starter key the propeller thing that you think it should be updated like I mentioned there is too much biology oriented, you can just change it in the hub I think it could be fine you know. Okay I see something in the chat from Donald. Yeah I haven't been looking at the chat so sorry. Sorry you might have addressed exactly what I was talking about while I was trying to Google for the Cornell folks. I was just linking up the there were some people at GCC. They were also interested in material science I was just making sure you guys were connected, if possible. Yeah we're aware of the Oak Ridge folks and the chess folks or the folks at Cornell. Yeah I think we want to schedule a meeting and then haven't managed it yet. Questions or suggestions or comments for Eli. And also Alejandra that is here. Just one comment thank you so much for bringing up the accessibility issues. Hopefully we've sorted out most of them and we'll you know keep working on it. Yeah, and thanks for I've seen like, I haven't been keeping track of all the stuff you've been doing with like ax linting and checks and stuff but it's really good to see that being like expanded say thanks. So we have this fairly loosely defined working groups system, so we have working groups responsible for various pieces, you know, you I back and tools outreach training and so on and so forth. Maybe one of your people who are actively know this can think of showing me one of them that's a good way of sort of knowing what's going on. Yeah, I'm in, I'm in a couple of the channels. Sorry, I'm in a couple I'm in like the tools in the, I mean I follow the tools channel I occasionally depend to the UI UX channel but only when I have something to ask about. But I mainly treat them as like a place where I can go to ask questions rather than like a group that I should be coming to the meetings of it feels again because I feel like I'm on the periphery. I don't know if I have much to contribute in terms of discussion and decision making. I kind of want to be informed of the outcomes but I don't necessarily want to be there to contribute to everything. I don't think we consider you very free for us. The fact that people from domains like yours use this is this is fantastic so we actually want to help as much as possible. So you're not a very free. Right. Anything else. Any comments. Then thank you very much Eli, I think that was a great presentation. I hope it was very helpful for everybody to join. Yeah, there will be another call I think it's by Dan on in a couple of weeks if I remember well on the 27th on the release process. So please join us well at the same time, we will provide a new link. Thank you all for joining and thank you again Eli.