 We have a panel called Python in Space, Curiosity Sparks. So our moderator will be Praveen Patan. So we'll be taking care of this panel session. So what do you, Praveen? Thank you, Galen. Hello, everyone. Very warm. Good afternoon to you all. And I must wish a warm good morning to our panelists. They are here from different time zones. And together, they're going to take us to space with Python. I'm supposed to be the moderator of this session. But I would love to consider myself to be a student of this amazing interactive class. Python, the language we love, is being used at various levels in the final frontier. From deep space exploration, trajectory design, navigation, flying robotic spacecrafts, to imaging black holes, and to analyze the wonders and mysteries of space. Thank you all panelists for being here for this panel session. I'm hoping this to be a really interactive fun, thrilling, and engaging one. Our panelists have assured me responsive answers because they want to answer as many of audience questions as possible. It may not be a cohesive discussion today, but we'll get to hear about the space adventures of the panelists. We have put together an amazing panel with these shining stars from Python Galaxy, five passionate scientists. First, we have Arthur Scholes. Arthur is a spacecraft operations engineer working with European Space Agency. He founded the Libra Cube Initiative for promotion of the idea of open source to space exploration. Our second panelist is Catherine Scott. Catherine is currently working with Open Robotics. And prior to that, she has worked as image analytics team lead at Planet Labs. She was co-founder of Tempo Automation and Scythe Machine. And next, we have an astrophysicist at heart, Thomas Albin. He's a space scientist, solar system researcher, and a Python developer. He now works in automation industry. Then we have Kazunori Akiyama and Andrew Shell. Both are astrophysicists and Python developers. They are the members of the team EHT. EHT is event horizon telescope. Thanks a million everyone for joining us. Let's make it interactive and I request audience to get your questions flowing in the chat. First few minutes, let's get to know a little more about our panelists. So I request each of you, if you could introduce yourself, what are you doing now? What were you doing then about your work, your passion? You may please share your screen if you wish to present a few slides. So we have three to four minutes each. We will begin with Arthur. Arthur, please. Thank you, Praveen. I'm going to immediately share my screen. Okay. I tried to summarize this in four minutes. First of all, thanks for giving me this opportunity to have a talk here and be part of this panel. I would like to, for this panel on space, Python for space exploration, I would like to give you an overview about the LibreCube initiative, which we have founded a while ago. With the goal to develop an open-source ecosystem for space and Earth exploration. So the vision behind LibreCube is that, well, I realized that a lot of systems share commonalities, be it satellites or drones or rovers. They all have, for example, a power system. They have a communication system. They have onboard computers. And if we could modularize those subsystems and make them compatible and have defined interfaces, then we could have a kind of a reference architecture or ecosystem where we can pick the components that we need to build our mission. It's a bit like Lego Mindstorm, where you can build different kinds of robots and rovers just by using Lego bricks. The difference to Lego, however, is that our systems are designed for operation in space. So we use space engineering best practices like redundancy, fail-safe, and disoperational concepts. And we use the materials that can survive in vacuum and high temperatures. So I'm giving you now a couple of examples for projects we have worked on, maybe to wet your appetite and maybe you'd be interested to join those kind of projects. And for the last two years, we have mostly focused on software projects. So we do hardware and software. But because of the pandemic, we stay at home. So it's very easy to develop software, but it's not easy to develop hardware remotely. So we were focusing a lot on software for the ground segment. So one example is this SLE User Protocol, which is a protocol defined by CCSTS. It's a big organization that standardizes protocols for spacecrafts. And what we did is we implemented, we did a Python implementation of this SLE protocol. So basically SLE is what you use. So the ground station communicates with the satellite, receives the data, and then the ground station forwards this data to the mission control system. And this data then, this stream is encapsulated in SLE frames. So we wrote a library that actually you can use. You can download now and then you can, just with this few lines of code, you can basically connect to, let's say, another station or ether station or JAXA station. And they all use this SLE protocol. Of course, you need to know the credentials and you need to be in the network. But yeah, that's quite interesting. Another thing we did is a module, a Python module for predicting the links. So when you receive a satellite, the signal might be, the strength of the signal might change over time, maybe because the satellite is moving and you might be moving and the satellite might be change orientation. So you can basically in this module you can set up your scenario with all the details and you can make them time dependent and then you can plot and calculate your link budget and you can see if you actually can receive the satellite data. Okay, this one I skipped maybe. And then there's yet another project. It's basically a viewer for solar system and satellites in space. So there are a couple of them already that even open source that you can download. But this one is a bit special because we wanted, in the end, I mean, this is using JavaScript on the web browser, but I didn't like to code JavaScript. So I'm using this Brighton Python module. It's called Brighton. That basically is a wrapper for JavaScript. So this is using the JavaScript 3GS library, but I'm not coding. I coded everything in Python and so that's really a powerful, the Python ecosystem. Then here's another example that's more hardware oriented. So we developed a protocol. Well, we implemented a protocol. So Canvas, maybe you know it, is the bus in your car that lets all the systems communicate with each other. This has been adopted for space use. They have changed it a bit, but they want to use it in space and they're actually using it already. And we have made a implementation and even implemented this on MicroPyton and we're having two boards that communicate using this Canvas protocol for space. Yet another protocol I'd like to present here is for the file delivery. You all know FTP. So this one is basically FTP for space. It's an open standard. It's available since many years, but it's not yet been adopted really for space use because the space industry is really conservative. There's load to try out new things. I mean the traditional space industry at least. So we have implemented this CFTP protocol. You can see on the left side would be the mission control system and on the right side you have the software running on your rover and then you can send files back and forth to your devices. I would also like to mention here a bit of advertisement that there's this open source cubes that workshop where we basically talk about this open source aspect in general, hardware and software and also there's a lot of focus on Python. So this was the last years and that just concludes my short presentation here and introduction. Thank you, Arthur. Great to know about these things. Over to you, Catherine. Hi. Yeah, so I'm Catherine Scott. I've been a Python developer for probably over a decade now. I guess it all got started. I did a lot of computer vision and robotics work back in college and grad school and it sort of turned into doing a lot of start-up stuff. Some of it, like I was a defense contractor for a while, some of it was kind of tangential to space work and then I think my first real introduction to space was working at Planet Labs. I wasn't really working on the satellite. If you're not familiar, so Planet Labs has a constellation of cheese. I don't even know where they're at at this point. A couple hundred CubeSats that do earth imaging. Most of it's at three meters, but there's some higher resolution stuff. So there I mainly worked on the analytics team and also the rectification team. So that's, you know, everyone thinks, oh, well, I'm going to write Python. It's going to run on the satellite or rover, whatever. It actually, you know, most of the big chunk of the software is actually on the ground doing all the processing once you get the data down from the ground. So most of the time I was working with images that had just been pulled off the satellites and brought to earth. So worked at Planet for a few years and then now I actually work for an organization called Open Robotics. You may also know us as Open Source Robotics Foundation. So we basically build two big open source projects. One is ROS or Robot Operating System. The other one is Ignition Gazebo, which is a simulator that you can sort of plug into your physics simulator, plug in a physics simulation. You can build these environments and actually test things without having to physically have them, which has become even more useful as robots become more complex, more difficult to work with. And I'm sure anyone working with satellites also knows the value of a simulator, right? You can't always, you can't, you know, simulate conditions on the ground that exist in space, but you can't do it in software, so it's really handy. So ROS makes, I guess I should back up a little bit. ROS is basically this glue code, for lack of a better word, that connects all these disparate parts that you would need to, say, build a robot. And so it makes pretty extensive use of Python. ROS actually allows you to intersperse different languages, primarily C++ and Python, but also some C, a lot of ROS lately. And lately, I don't personally work on a lot of the space stuff, but there's been more and more interest in using ROS for space applications. We have a couple contracts with NASA. I can't actually speak on behalf of NASA, but I can kind of tell you, you know, what's publicly available about those projects. So the one I think we work on a bit is Viper, which is the volatiles, forget the acronym, but volatiles, it's basically looking at water and volatile chemicals on the polar region of the moon. I think they're going to announce the actual location of the mission sometime next week, but you'll have to go talk to NASA about why they're using ROS and why they're testing it. I think they're using it mainly to, like I said, work with data and work with systems on the ground. Lately, I've seen also just a lot of other work, because it's open source, people just kind of pick it up and use it for what they want. So I think I've seen ROS used, I believe it was also used in the Astro B program, which are small, well, they're not CubeSats, they belong inside of a space station and they're designed to move around and help astronauts do things. I think JPL also had a little toy ROS simulation that you could use, which was pretty cool. But not exactly working on space stuff myself right now, but I help with ROS, I help basically build the ROS community, which since it's robotics, it's becoming more and more interesting for planetary exploration and space exploration. So that's what I'm working on now. Great. Okay, thank you so much, Katri. Over to you, Thomas. Thank you very much for the invitation, guys. It's very nice to be here. So my name is Thomas, and I worked on, yeah, I studied astrophysics and I was part of two major missions or worked on two major missions. The one was the Rosetta Fiele mission. You may know the mission from ESA that landed on the comet. So there I conducted some calibration experiments with some dust detector in the laboratory, later on also, yeah, and making the analysis of the actual data from the comet surface. So this was a pretty nice mission back then. Afterwards, I joined, for my PhD studies, the Cassini team. So Cassini Huygens was the probe that revolved around Saturn. And there was part of the cosmic dust analyzer. So again, dust. But it's pretty interesting. So, yeah, it was a very great time where I analyzed the data. I was with machine learning algorithms and so on. So I combined miscellaneous topics with the data we got from the instrument. And also, yeah, my very first contact was back then in 2012, where I had an internship at the European Space Agency working on near-Earth objects. So simulating where are these objects and with what kind of telescope. Yeah, I can observe them best. So as you get, you get it that, yeah, I like the space topics, but I always focus on the very small stuff like asteroids, comets, and dust, and meteoros as well. So this was my academic journey. Two years ago, I then left academia because I said, well, I would like to extend, broaden my horizon a little bit and I joined the automotive industry. But in parallel, I'm still an astrophysicist at heart. So the free time, I'm working on some projects on some data analytics topics. So now I do not have, let's say, the academic pressure of publishing papers or so, but I can really focus more on developing actual products. So really libraries or so I'm working on, honestly speaking, this was now a big summer break now for me. So I was now sitting in front of the computer. So I enjoyed the summer. So this PyCon India is now the perfect starting point for autumn and winter to rejoin my programming sessions. And beside the programming, also, yeah, I'm writing some tutorials. It's called Space Science with Python on Medium. And I'm also now planning for winter and autumn to start some tutorial sessions also on YouTube because I think that citizen scientists, citizen science ship is getting bigger and bigger and more important. But it's very important, I think, to channelize all the knowledge because if you take a look at Reddit or so, people have crazy capabilities. So showing crazy 3D animations with Python whatsoever. And then they say, please don't be so harsh with me. I'm only 16. I was like, geez, this is amazing skills. So I think there are a lot of people out there with great, great skills. And I think they can really contribute a lot to the space community, I think, especially generating databases or infrastructure, creating actual products and so on. Because for me personally, I think there is a lack in the scientific community because they focus always on paper, paper, paper. But if you could focus or join the citizen science community to create actual products, more libraries and so on, this would benefit everyone, I think. And so this is my passion, you know, just doing some machine learning stuff with asteroid data and so on, just for me, myself. And on the other hand, creating tutorials and also starting some tutorials on YouTube for the space topic. And yeah, I can share some links later in the general chat if you're interested in. Yeah, that's basically me. Thank you very much. Thank you, Thomas. That was so encouraging and inspiring. One should go and check his tutorials. They are really amazing. My 10-year-old is waiting for a star in Orion constellation to explode. And she wants to see a real black hole. She has millions of questions, but I don't have convincing answers. And today she is excited to attend this session because we have two guys with us, members of EST team, the team who got us the first-ever image of a black hole, black hole at the center of Galaxy M87. So, Kazoo and Andrew, please. Sure. Let me share my screen to show some slide. Can anyone see the screen? Please go ahead. Okay, thanks for having us in Picon, India. I'm Kazak Yama from MIT. I'm at East Coast of the United States. So, Andrew and I are part of the international groups of scientists providing these astronomy pictures. So, I believe most of you saw these somewhere in the last couple of years. So, this is the first-ever picture of a black hole, one of the most exotic and mysterious objects in our universe. So, here you see a link of light lens at the edge of black hole due to its long, long gravity. So, you can see this link is actually illuminating in the center of dark shadow here. So, that is created by the event horizon of this black hole. You know, where no escape of light or any information is not allowed. So, recently we have updated an image. Maybe you already saw it. So, now you see a many spiral-like lines of the link. You know, these are visualizing the polarization of the light. So, this highlights something also invisible. So, these are these final arm-like patterns of polarization reveal the structure of magnetic field at the edge of the event horizon of the black hole. So, the black hole is one of the highest density objects in the sky and therefore extremely compact. So, the angle that I have this link is just only one over 100 million in degrees. So, this is like an orange on the moon seen from the ground. So, if you look at the sky, seeing the moon and if you can see an orange on its surface that is the size of this link. And so, to obtain such a definitely finest view of the universe we really need a little bit of a planet-size telescope. So, our picture is obtained at radio band or more precisely microwave band. So, in radio telescopes are like such kind of public antennas and if we have an R-size telescope, this big, big dish, I mean, of the meter will capture the light from the black hole and then correct them into the focus here. But, of course, in the reality we cannot make such a big dish. So, what we are doing is instead of having a literary planet-size telescope just put an array of telescopes at the distinct geographic side across this planet and combining signals, you know, which actually each telescope is receiving and then computationally form such an R-size telescope. So, we focus it in the computer. So, this is how we made these images. So, to make it happen, we formed an array of radio telescopes operated at the microwaves wavelengths named Event-Prize and Telescope. You know, you can see the network of EHT array which indeed has a planet-size and to operate such an international planet-size array, we, of course, need an international team of scientists. So, this is the Event-Prize and Telescope collaboration. Now, a group more than 300 amazing scientists, I mean, and Andrea and I are just only a part of this amazing, massive collaboration. So, you know, this conference is a pipeline, you know. So, how are pythons are involved in this collaboration? So, I'd like to switch myself to Andrea for that. Great. Thank you, Kazoo. Maybe you can go to the next slide. Yeah. So, Kazoo talked a little bit about the physical Event-Prize and Telescope, you know, the physical telescopes and the people who operate them. But a huge part of our work is computational. As I said, we're taking these signals from different telescopes all around the world and trying to combine them to sort of simulate what we might see if we were able to build a telescope besides the Earth. And a huge part of this computational work goes on in Python. This is just a small scattering of the types of Python libraries that we use. And we've developed a whole bunch of different Python libraries and software tools of our own for this project relying on this great community of people and existing work. It has really been helpful in allowing us to flexibly develop very powerful and easy things to analyze these data. Next slide. So, one thing to emphasize about the ESG data is that we're sort of taking a huge amount of data or recording tons and tons, petabytes of signals at each telescope and then sort of trying to reduce that into just an image, which is only a few pixels. So, the resolution of the telescope besides the Earth, we're just barely seeing this black hole, which is really, really small as seen from Earth. So, we're taking these data that are petabytes in size and reducing them to kilobytes in this final image, trying to extract out from all of these noise and different corrupting effects that we get from the atmosphere, from electronics, trying to extract out this one small image on the sky. Next slide. So, there are tons of different things that happen at these different stages. Just one stage of the problem is what we call the imaging stage, which is something that Kazma and I have really worked at on a lot. So, this is where, you know, because we don't have a telescope that filled up the entire Earth, we only have a few different points that we're sampling around the surface of the Earth. We have very sparse measurements, so that means that there are lots of different solutions for an image that could fit the data. There are an infinite number formally of images that could explain the data that we receive, and so we have to use algorithms to try to sift between these different images and find the image that best explains the data and that also satisfies all of our assumptions and physical knowledge of the system. Next slide. So, for the EHD, we've developed several different software tools that do this step. Just two I want to highlight are ones that I developed and Kazma developed. So, on the left here is the EHD imaging Python software package that I developed for this task of imaging of taking the EHD data, producing an image of a black hole, and on the right is a snapshot of the Smiley software package that Kazma developed that does the similar task with the same task. They have many things in common, but also we're developed independently. So, that's one thing that we really like to do doing the EHD is approach different topics from sort of different angles with different software. Python makes developing all the software really flexible. We can learn from each other while also maintaining things, keeping things somewhat independent so that we can test each other and make sure that we're not making mistakes, we don't have bugs, we don't have assumptions that don't hold up. Next slide. So, yeah, this is just sort of highlighting that idea again. One of the really critical things that allowed us to know that we were on the right track is when we were able to reproduce this image, this result from sort of three different independent methods. And this is something we always try to do is try to reproduce our work and use different methods to arrive at the same results to establish confidence. So, yeah, it's a lightning introduction to the work that we do with the EHD. Yeah, such an amazing work and that was so inspiring. I guess we have spent a little more time in this than planned, but there was much needed to know about what panelists are doing and the passion and the work. Now, let's quickly go to the questions now and I expect like responses and there are many questions coming from audience as well. My first question is for Arthur. You have been doing great work through LibreCube. So, please tell us like how one can contribute to LibreCube and where to find the information quickly. Well, the information is on the website. That would be your starting point, LibreCube.org. And it's open for everyone to join. So really everyone. We also have a channel on Matrix so where we can chat. Mostly the question comes up how can I join what projects can I contribute to? But actually I'm quite good in making projects and laying out like what I expect from projects. But this overall organization I'm still in the learning progress and most fascinating is for me how the Linux kernel, how they organize themselves and if you have an idea how to build up this kind of organization and please also get in contact with me. Okay, I would love to. But I guess if I work for LibreCube I won't get paid for. So my question is like why would someone want to work on LibreCube projects for free? Yeah, it's actually well, most of the work is not paid. That's true, it's voluntary. It's like and it's really a passion. People that join they really have an interest in space and space exploration. What benefit you get is if you have joined our projects and you put something on your CV like you have implemented a CFTP protocol. I think this makes an impression if you apply at NASA or any other space organization or space industry that might be an asset. And the other point is that we have all the scholarships from time to time like we just completed a Google summer of code. We had two students. So if you join and you're already in this community then it's easier to then to find something for you too. Great. I guess Catherine wants to add something to it. Yeah, I was going to say trouble from like working at Planet and working at Opposite Robotics Foundation or Oppo Robotics. Even though it's open source and maybe you're not always getting paid directly to contribute to open source there are lots and lots of benefits to working on open source and it will most likely help you find work. What comes to mind is the guys who who wrote GDAL which is this rectification software that I think everyone uses to do earth imaging or a good number of people use and the core devs there I don't think ever have to worry about finding a job to work on GDAL. Same with a lot of the ROS developers as well. Yeah, true. Okay, now I'm coming to you Thomas. Tell us what is your astrophysical background and how did Python help you to fulfill your academic goals? Yeah, so good question. Well, I just mentioned at the very beginning that I worked at ESA as an internship or like I think six and a half months or so back in 2012 and this was my very first contact with Python. I have to admit that I studied physics so a lot of lab work and so on and then my supervisor said here take this book I think this programming language is amazing maybe you want to try it out and I said well okay let's try it out and this was my first Python contact and it really helped me yeah conducting some numerical experiments. Later for the Rosetta mission also for Cassini Python was like the tool to really create data reduction pipelines, feature engineering and so on also for plotting routines also creating web applications and so on so I just wanted to get the job done right so this was like the best the best programming language to choose and this really helped to generate the data. I mean at the end of the day it was important to generate results and instead of starting with some low level programming language, Python was like the best choice and also with all the amazing libraries out there like scikit-learn, Keras TensorFlow and so on you don't have to reinvent the wheel right and I can simply start with applying all those methods from the beginning on of course you have to know what to do exactly right so what's the specific question and so on so you just cannot just play around but the starting hurdle is quite low you can just get the data if you have the data you can start true like I am from a non-computing background and like when I had my first encounter with Python you almost be knowing that comic from XKCD like it says import gravity and you start flying Python gives you power after wants to come in here yeah I agree totally that Python is the language that doesn't get into your way so you have your idea that you want or your problem that you want to solve and you can just write a code and it's so readable I mean we all I guess we have worked with C, C++ maybe assembler or even Java I always find it hard to remember the syntax for those languages but Python is almost just like writing English and has a rich ecosystem and what I wanted to add also is that for embedded programming there always comes the comment that yeah for embedded you need C or you know really low level but there's this MicroPython project that really lets you program your small microcontroller and maybe eventually you might port this then to C but at least you can already get started quickly and try out for measurements it works very well and this is really what amazes me about Python this rich ecosystem and you can use Python for almost everything great and that semicolon doesn't trouble you anywhere yeah so I have a quick question for Thomas what are your current space related projects and what is coming next yeah thank you for the question so currently I'm developing a small library called solar why because I think it also came up in the questions in the in the in the chat that where to get all the data right and that's very difficult there are certain databases also from NASA and ESA and so on where you can get the data also calibrated data because uncalibrated data is really like hell but there's no really like a consolidated product or website where just summarizes everything there is the international astronomical union where you can also find the minor planet center for asteroids there is the international meteor conference and organization where you can meet your data and so on and so forth you just have to stick around and find around and it's easier if you come from this background then you know where are the data but as a citizen scientist you just say yeah well I have no idea so with the solar why I just try to at least cover the minor object topic a little bit so starting with asteroids starting with meteor and at some point also maybe with the Cassini dust data that I would like to provide in a calibrated form and upload them somewhere in the cloud and beside that of course also the tutorial stuff right because you can learn programming you learn programming quite easily but if you check reddit for example where people ask yeah what can I do next the question is always what can I do and they don't know what they don't have any idea for any project so of course there are this books like I think your boring stuff and so on but really working on actual scientific project is very difficult for let's say somebody who has no academic background and I think with tutorial sessions with really getting in contact with scientists you can really contribute something and even if it's let's say yeah not writing a paper but reducing the data or introducing machine learning algorithms to support scientists this would really help and yeah I hope to build up a small community around this idea a little bit so so let's see great okay so let's move on to Catherine now like much of your work involves programming in python and you have worked on projects like using satellite imagery or for some environmental issues like to make an impact on the world and now you are into space robotics so what do you think the next 20 years of space exploration will look like well I think there's a lot of you know definitely going to see a proliferation of more and more and cheaper satellites and the ability to put them up has become amazingly cheap I think there's also I mean I think everyone here can speak more than what I could say there's going to be more earth imaging as well or imaging from earth that's going to increase over the next 20 years the question really remains is what do you do with that data once you get it you know and I see it over and over and over again is that this data comes down and then it's like well what kind of questions can we solve with this how do we actually get it to get it to people in a useful way right like the things that come to mind especially regarding like climate change and forest fires and stuff is you can get some really high quality very timely data now but getting it to say a firefighter or a farmer or somebody who can use it is still not a solve problem there's just this big disconnecters people in front of laptops that are sitting there actually looking at all this stuff but then getting it to the person who can actually use it and getting it into a way that you know they can digest it is really difficult and that's where I think Python is going to come in a lot so on the same note I have another quick question for you like how do you think space exploration has changed in the past 10 years like it's gotten incredibly cheap well I wouldn't say incredibly cheap but it's gotten much cheaper and it's gotten much easier to put things up into space and to get a hold of you know to prototype really and to see at least what I've seen is people as things have gotten cheaper people have been willing to take more risks right before it was like this thing has to work it has to be radiation hard and it has to be you know we have to put it in a thermal chamber and really put it through the space and now it's like well you know it's only going to cost as much as we can be a little bit more you know gives them a shot if it works it works if it doesn't it doesn't you probably see a lot of this yeah that's true but I don't really like this trend I must say because it's creating a lot of space degrees and that's what I actually that's why I have that's one of the reasons why I started LibreCube is to really because we have a lot of universities developing CubeSats I'm sure many of the people here in the session they are also somehow involved with the CubeSats maybe because all over the world they're building CubeSats but they're not sharing information so they're repeating the same mistakes all over again it's mostly power related or communication I've launched two CubeSats and one of them the second one was actually a big failure and we did not even communicate with it we're not able so this drove me that we need to go open source we need to have platforms that work and then we can put some experimentation on top of this but the basic thing must work so that we can always make sure where's the thing and maybe even maneuver it so that we avoid having yeah avoid having space debris but overly it's true we have a rapid kind of more rapid prototyping for space nowadays as it was in maybe 10 20 years ago yeah so we are running out of time so let me quickly go to the next question Kazoo is a developer of Smiley like a Python interface library for imaging with Event Horizon Telescope and Andrew is also a primary developer for another library a ETH imaging primarily made for PhD data so my question to both of you is why did you have to develop a new software for making black hole images for the HD and how is it different from software normally used in the field yeah yeah so that's a that's a great question so actually we are using the techniques of the very long baseline to make a partial off-size telescope it's still not new actually it was like coming from the 1970s but what is new is we are observing at extremely short wavelengths as a radio so that is something very different from the conventional one and in general if we observe at the short wavelengths then we will highly affect the atmosphere and actually that actually crafted a lot of telescope signals and the imaging problem is more and more challenging so one of the very big difficulties is actually we need to handle this kind of the anomalous spirit effect built in the imaging and also another thing is we have a planet-sized dish but we only have to make that picture of first images of black hole we only use telescope at just a five distinct geographic side across the globe so we cannot up the entire earth as a virtual mirror so we basically have a very sparse measurement of the solar informant from the sky so these actually challenges may actually pose a very big limitations to the standard techniques used in very long baseline interference previously and we need to actually develop some smart algorithm to solve actually in far what are the likely image on the sky so that's why we need to develop first of all new algorithm and also the software that implement the new algorithm so that's actually the primary motivation of many forks including myself and Andrew to develop the new software packages and even not only for imaging every single part of the data process in the EHT actually need a technical so under the show that the data are huge like a pitabyte and at the end is the kilobyte and so we really need to handle the framework to handle such huge big data as well so the EHT basically faces many challenges in every single step of the signal chain and actually that will create development of new software yeah one thing I want to emphasize too is that we are so using sort of traditional software and traditional techniques alongside some of the newer stuff that we're developing we're in house with Python with the EHT and testing these methods against each other and there are some things that the newer software does better there are lots of things that we can learn from the sort of more traditional software packages that have been used for decades at this point and yeah it's really by testing these methods against each other we established a lot of confidence in our final result one thing that's great from my perspective and about the new software that we develop in Python primarily is that you know it's much easier for everyone to contribute to and to develop and to take ideas that we come up with for one project and apply them to another to make changes on the fly and some of these older packages which have a lot of legacy and are very stable and that's very good because they're useful for a lot of different problems but you know for when we're trying to do things quickly and to make progress fast and just improve as much of our data as we possibly can on that kind of flexible development is really helpful great okay so a quick short answer from both of you like we all Pythonistas would like to know why Python for these libraries yeah I think it's basically what I just said that you know Python is super easy to experiment and I think it was said earlier to take you know work open source work that's been done by others and plug it in and play with new ideas and tools and so we really rely on all that development that's done in tons of open source packages and that's just really helped us experiment quickly and put together new ideas rapidly and with a lot of success okay yeah so yeah just one single point yeah and another thing is it's already suggested in this session you know DWT is one of the key so I would say I would just make sure that astronomers are not good programmers you know we are not professional astronomers right and but still we need to work with people of group more than 300 people and so kind of like easy to write and also the clear to maintain the code it's just crucial I mean for this kind of international problem that is I think what really Python is the real solution for the EHD okay so I have few questions left with me now but I think we have only 8 to 10 minutes left and therefore if there are any interesting questions from audience let's take them first Bhavin can you just help us or yeah in the earth size telescope slide there is a label light from the black hole how and what does it mean if light can't escape black hole due to infinite gravity then how can we capture light from it yeah that's a really good point that sometimes astronomers are sloppy about these things exactly what we say so basically when we say light from a black hole it's light that's emitted just outside the black hole so this black hole is surrounded by plasma by superheated gas that's billions of degrees and that is emitting radio light that we see it's bent by the black hole so we see the effects of the black hole's gravity in bending the light but it has to be right that we don't see the actual light from the black hole itself because once it gets on to the actual black hole it's it's invisible okay any other interesting question from audience yeah this is for EHD team again can you help us with learning suggestions while setting up proof of concepts from such initiatives with respect to computer engineering how did you take this one sure I'm not sure actually I clearly understand these questions but wait I thought we need is that actually like asking about the imaging techniques or even the question is not clear to me okay yeah we'll take it in the chat later on can we have the question actually one thing I can say is that so basically we maybe actually face the challenges like imaging or other data processing you know this is usually something actually not done in the field so we usually actually look for what's happening in the other field so for myself actually I started to take a look at the imaging techniques actually coming from the medical imaging I look at the techniques actually what kind of techniques are used in the medical imaging or you know other field which have a very similar mathematical structure that was actually motivated as to how to develop imaging techniques so I don't say but often actually it's very nice to just look at you know how smart people in the other field actually feel and solve a very similar kind of problem that is very helpful I'm not sure this answer to your question but the next question is not exactly directly related to python would like to know what is astrophysicist community take about UAP especially with the release of videos from pentagon okay who would like to take this question seems like no one me not because my focus is really on kubesat python for space so that's maybe more use the community science community question yeah let's keep the only thing I can say is I've seen a lot of satellite imagery and I've never seen a UAP so and what I can say is that this is the answer I expected if this kind of question comes nobody likes to answer it I don't know sorry what is UAP is it like a UFO identified astro something phenomenal similar to UFO okay so we have just three to four minutes remaining yeah let's take one last question and then I have one more final question for every one of you and 10 second one answer how much improvements accessibility in the area of cloud computing has helped data processing for astro data and is it still done on specially built on prem servers maybe I can say something so I think that I can not speak enough for the astrophysicists with their who have really terabyte of data I think the other guys can answer it but especially spacecraft probes so most people think that spacecraft probes they generate maybe terabyte or petabyte of data but that's actually not true because one link is very very small so you compress the data you have to really select it and take for example the Cassini spacecraft it was flying around for what like 20 years or so and a total a generated binary data of around I think 400 gigabyte over 20 years and then you take a small fraction of it which was the instrument I was working on you unzip it you feed it with extra data you make data engineering every fancy glancy thing you can imagine and at the end of the day 20 years of measurement we had like 250 gigabyte database and you can imagine we do not we do not need a cloud platform we had an on-prem system because it was cheaper for us in this case and but I can imagine of course that if you have petabyte of data you need of course some kind of data lake or something right I mean I just had one my experience has been the opposite but it's lots and lots of data the cloud is great you know I tend to suggest people never do anything locally because the likelihood of you having to go and you know the whole Silicon Valley things scaling stuff I've now that's you know there's a difference between fundamental science and building products but I used to work at a microscopy company it was the same thing it's like well when the individual scientist does well it's great on your local machine but you know the next day somebody's going to come and say well you figure out how to find this disease now we want to do it for everyone it's just easier to say well we're going to run that all in the club Babin do we have any okay Kazu wants to add something yeah so this is actually a very nice question and this is also a little bit of a topic for the EHT so just to show the EHT use both actually on-premise servers and also a cloud computing so for instance when we actually form the virtual R-sized telescope that is a kind of well-defined procedure and you know we just need to learn you know for a huge amount of time so that is actually using in the on-premise server we have a super computer at the OSAPG for instance to combine all the signals from each telescope to form a virtual telescope when the other hand actually for instance imaging or data calibration part or data modeling part it's more like you know depending on data set and what is your goal so for instance these Oracle images presented from our collaboration is actually processed at the Google cloud so we need a kind of just instantaneous shot of the huge computation and also this actually should be part by actually many people you know across the globe I mean in the cooperation so the cloud computing is much easier actually to learn this kind of stuff so I would say we are still actually have both benefit from the on-premise server and also you know more flexible framework of the cloud computing so yeah these are most of them are very important for the EHD okay so I'll put quick questions for all of you like a combined question like both satellites from ESA and NASA till today they run software written in C or ADA like so the question is will python ever be used in space in real sense what's the future okay so I would like to hear Katherine's opinion on that actually because you have worked with I mean I'm supporting now ESA European Space Agency and I'm involved in this let's say more traditional style of satellites the huge satellites you know the flight to Jupiter and they want to have everything testable so that should be C or other code but I imagine that commercial and startups they're more agile in developing and they're more you know take more risks and I guess they would go for the language that can be implemented and give the fastest results I would imagine that python is used a lot maybe for planet labs but I don't know maybe you can share something I didn't do a ton of onboard work but I wouldn't be surprised if it isn't already you know the onboard work that I did do I can't talk a lot about but it was more in C but you know at the end of the day all the satellites are just Linux boxes so who knows what scripts running behind you know some sort of cron job that runs every once in a while to do something's probably python script so there you go okay I would like to hear your thoughts on this from Catherine and Thomas how do you get career in space exploration or how citizen scientists or pythonistas can contribute my general recommendation is just to start doing cool stuff I actually talked to a student a couple weeks ago and she was really interested in doing space stuff and the one thing is it's like well you know try apply like but you may not get a job there's a lot of people interested in not a lot of slots like maybe you can come out it from a different angle like I'm sure everyone here who's building satellite or whatever there are all these different bits of hardware you don't build everything from scratch right you buy this image or you buy this you know battery by you know all these different parts and if you're really interested maybe start with some of the you know the vendors that actually built the parts that go into these big systems and that way you can get it's a little bit of a back door into the industry I think that is one way of approaching it and what I can say about the maybe citizen scientists so the question is the second how how can citizen scientists or pythonistas contribute and generate scientific output depends a little bit on the task so it's very difficult I would say for someone who is no academic affiliation to really write scientific papers because sometimes I mean correct if I'm wrong but sometimes this publishing industry can be quite conservative and I think that big contributions can be done with really working on astronomy astronomical libraries or products or solutions that generate generate data feature engineer the data and so on and as always you have to get in contact with people right so if you think hey I'm quite confident with my machine learning skills or you may be a free Lancer you want to support something just approach the guys it's also the same with the students right you like the students that come maybe in the second semesters and say hey I would like to program something please help me so you think oh this guy has motivation and it's maybe something really something about your commitment to your self presentation how you would like to contribute things Arthur please I think the fastest way really if you want to get a job in space industry is to start a space company you could tell the open source products that we develop at libel group for example okay great Kazu Andrew you want to add something a quick 10 seconds respond so I so it's not for the EHT but actually the in astronomy actually many citizens actually join in in the you know many many ASCAP or the astronomy work for instance like helping identify the object that is actually one of the primary area that citizens really help actually there are many ways to join actually we often need people's eye to see the sky Andrew do you have any addition no I think that's right I think you know just by contributing to the overall planning ecosystem too you can contribute to different space projects and especially the HD project in unexpected ways I think the amount of dependencies that we have we rely on a bunch of things that people contribute to even if they don't necessarily directly to our software libraries for example which can require some specific domain knowledge if you're working on some of these larger projects you're definitely making an impact in the HD maybe one last okay sorry maybe my last comment maybe there are actually amateur communities out there right there are communities for asteroid observations there are communities even for people observing transits of exoplanets amateurs for meteor observations for atmospheric phenomena you just have to find these communities and these communities are huge they have a lot of knowledge and I think being part of such a community is really more beneficial just being you know solo fighter trying to do something so I really recommend searching for the proper community and they will get a lot of help yeah that was amazing I wish this could be continued for like another hour but we are running out of time so that was amazing you guys were awesome and thank you for being us taking us on this journey you have definitely inspired our young pythonistas thank you so much and to the audience I'm sure this session has ignited your passion to dive into the world of python learn persistently discover and create so that you too can one day help us all understand a little more about world beyond the earth so we are ending this session here