Content for EGU22 PyGMT short course #1
Comments
|
@weiji14 thanks for organizing all of this! I'll be on vacation next week (2-10 April) so none of those dates work for me. I can catch up on notes afterwards if it's not possible to move this to the week after (or this week still). |
|
Ah ok, that's alright with me, I'll move the meeting to the week after (11-16 April), will update the when2meet poll. Edit: done, new link at https://www.when2meet.com/?15145108-mijPx. |
Rough timing, but I am on vacation April 09 - 17 and will be unavailable the following week (until April 23). But I can share some thoughts via GitHub ahead of time and review the notes. |
|
Ok, let's just skip the meeting and do this asynchronously then 🙂. I think @andrebelem is doing fieldwork too, so we might as well just do this at our own pace. I've started a draft outline at https://hackmd.io/@pygmt/egu22pygmt_outline which you should all be able to edit 📝 Just pick any free slot and write some notes on what you'd like to teach. We can then discuss more about how to turn all of this into a coherent piece. |
|
Sorry, kind of dropped the ball on this. I can do the first tutorial with a PyGMT basics (Anatomy of a PyGMT figure) to introduce things like the layer-cake, how to read the docs, common arguments, coast, and plot. Happy to switch as well if someone else was planning on doing that one already. |
|
Cool, thanks @leouieda! @meghanrjones and @andrebelem, do you both have any preference on what you would like to include in the short course? If the individual deadline for this Friday 6 May is too tight, we can move it to next Tuesday 10 May so people can get the video done over the weekend or so. Just be aware that there is a hard deadline on 18 May when the full tutorial will need to be uploaded. |
|
I'm glad to take either of the two remaining sections mentioned in the abstract
Thanks for the flexibility @weiji14, I will make the content and video over the weekend. We should aim for some consistency between videos in how the tutorials are presented. What do you all think about having the videos be demos of Jupyter notebooks? We could share the notebooks in this repo either as just a simple collection or if there's time collated into a Jupyter book. |
Absolutely!
That's probably the most useful way to do this. Then people watching can have a copy of the code to play with. I was planning on making a brief notebook and then writing the code live in the recording (but leaving any markdown cells populated in the notebook already). I've been doing this for my courses for the past few years and it works very well. I would avoid having a notebook full of code already that we just run in the video. Whenever I see or do that, the tendency is to go waaaay too fast and then the video becomes more of a demo and less a tutorial/workshop. Even for a demo, that only works if you don't want people to actually read and understand the code as you're going through it and just focus on the output. |
|
Yep, agree with going with the live-coding/typing method, it'll make it much easier for folks to follow along. I'll try and set up a Binder link (or links) today on this repo so that there will be a JupyterLab environment for us to code on. |
|
Dear colleagues, I was to reply to @weiji next and - bump! - another emergency! I apologize for only being able to sit down now. In my head, I was thinking about the last part of Tutorial 1 (anatomy of a PyGMT figure) but based on "Mars" (more or less along the same lines as this forum post https://forum.generic-mapping-tools.org/t/how-can-i-make-a-scale-bar-for-mars/1168/5) as a way to attract people from the Planetary Science. At the time we talked I thought about doing the same for Europa (Jupiter) but I saw that the explanation of the projections could get complicated (but you can use Moon IAU2000). @leouieda, what kind of map are you thinking of making? something about bathymetry? I can try to fit something in from there. What will Tutorial 4 be? Another thing - @weiji, are we going to do everything in Binder ? I ask this because somehow I "wrecked" my conda with a setup of a jupyter server and will have to reinstall the whole machine. At the moment I'm just using Colab (I haven't tested it for the new version of pyGMT but it would be an idea to also make a tutorial with it!). Once again sorry for this inconvenience and thanks for your patience. |
|
Hi @andrebelem, glad you found this post! A non-Earth example would definitely be interesting! From the sounds of it, you and @leouieda could coordinate a bit on the first half of the short course. I.e. these 2 points in the abstract
Maybe Leo can talk a bit about projections/inset maps, and Andre you can focus on colorbars/legends? Will leave it up to you on what works best. If it helps, I made a basic PyGMT tutorial recently which you're welcome to adapt, see https://icesat-2.hackweek.io/tutorials/DataVisualization/dataviz2d.html.
You can if you want, just note that files in Binder aren't permanent, so you'll need to save and download a copy of your notebook frequently. Basically anything that works, and using Colab is fine too.
So maybe @meghanrjones and me can work on these two last points. The geopandas <-> PyGMT integration is something we haven't really showcased before so I'd be excited to see that, but we could also show a geoprocessing pipeline like gridding or such, using some of the functions in https://www.pygmt.org/v0.6.1/api/index.html#data-processing. Do you have any ideas Max? |
Your lidar use-case seems excellent for this point, but let me know if you want tutorial 4 to be more application focused as well.
Good point! I'll take on tutorial 4 and will highlight the geopandas <-> pygmt integration. I may also include short examples for other pydata structures depending on timing. I'll share a notebook tomorrow with some content and will wait until Monday or Tuesday to record the video in case there's any suggestions for revisions. |
Cool, I'll work on a LiDAR to DEM tutorial then. What sort of other applications were you thinking of? Btw, if anyone needs any other Python library (e.g. rioxarray), feel free to open a PR to add it to the environment.yml file. |
|
That sounds good to me. Sorry to hear about all the trouble your going through @andrebelem. I finally got some time today to wrap this up. My plan was to combine these two tutorials:
That's all that will fit in 20 min and even then I'll be stretching it a bit. So insets, subfigures, and more advanced labelling will be left out. |
|
Hey everyone, got a section kind of ready at #7. This section stops at creating a map with shorelines and points coloured by a variable. We include colorbar but don't mess with positioning. I think this would flow nicely into @meghanrjones's part on pydata integration. Then that could flow into #5 by @weiji14? The only things missing would be insets and subplots, which have a bit of a complicated syntax to be honest. I don't think they would fit well in either of these. But I'm also fine with leaving them out of this tutorial and pointing people to the docs to learn more. Given @andrebelem's understandable current difficulties, I think these 3 x 20min sections would already be a great tutorial. Then if the situation for Andre gets better, we can include one more 20 min section at the end about planetary data + putting insets, scale bars, and other annotations. This way we take the pressure off of @andrebelem and are still able to meet EGU's deadline. What do you think? |
I really want to have a tutorial for either inset or subplot, but yeah, it can take a lot of time to explain, and not sure too where this would fit while looking at the draft tutorials we have currently.
Would check in with @andrebelem first to see what he prefers, but yep, I'm ok with having just 3 x 20min (or we could each get an extra 5-10mins). We can also just have a planetary data Jupyter notebook on the website (without a recording), and that would be an equally great contribution as well. Note that the Wednesday 18 May 2022 deadline is just for the video recording. The Jupyter Book can be modified afterwards, and there's a few things I'd definitely like to tidy up on there (e.g. make a proper welcome page, some basic installation instructions, etc). |
Depending on Andre's preference, we could add a shorter snippet on insets and subplots in between #7 and #8. Basically just combine https://www.pygmt.org/latest/gallery/embellishments/inset.html and https://www.pygmt.org/latest/gallery/embellishments/colorbars_multiple.html#sphx-glr-gallery-embellishments-colorbars-multiple-py. I think the amount of content in the three sections is pretty good as-is. Let me know if you want to me put that combination together. |
|
Thank you @leouieda. I appreciate your consideration. A basic question. How are you starting? any previous presentation about the course proposal? Do you already have a video ready that I can watch? Another thing - I re-tested the installation of the newest version of PyGMT/GMT on google colab and it's a lot faster than last year. I believe that Google has been improving the VMs. With that, I also thought about adapting the notebooks that @leouieda mentioned above into a colab version, interacting with data in gdrive, since with COLAB this makes it much easier. What do you think of the idea? This week was difficult but I think Saturday I will have enough time for everything (I know it's on the limit, but for those who like basketball, 2 seconds is still in the game). |
|
Dear all, Basically my script is a comparison between the topography of Hawaii (without the ocean, of course!) and Olympus Mons, same projection and plot conditions. Obviously it's going to be pretty easy to see that Mars is something like 3X larger in proportion. The Hawaii part is still missing, but it's the simplest and I can do it in 5 sec. Tomorrow I'll be stuck in the hospital all morning but I'll dedicate the afternoon to that. As we talked - if we have time, it will be nice. If not, that would be nice too. I'm doing everything in COLAB but I'm going to create a binder version of this notebook. A question - in the video, we started from scratch and not on top of a finished notebook, correct? to demonstrate the line of thought in creating the map (that's what I understood).
@meghanrjones yes, I'll put a snippet for the whole planet showing the region and my notebook will be more or less like the @leouieda idea adapted to "Making your first Mars figure". Best wishes |
|
Just a few updates ... |
Sounds like a good example, I think having a non-Earth example will be a small win already! If you've got a draft ready, feel free to start a Pull Request or upload it somewhere online and we can try to help fix the bugs you spotted or suggest workarounds.
Cool, we can also add an 'Open in Colab' button in addition to the current Binder button following https://jupyterbook.org/en/stable/interactive/launchbuttons.html#add-a-launch-on-google-colab-button. I thought Colab would be troublesome as mentioned in GenericMappingTools/try-gmt#17, but it sounds like you have some code to get it working somehow? Oh, and feel free to make the recording in Google Colab if it's easier for you, doesn't have to be in standard Jupyter/Binder.
Yes, you can keep a few code comments, but try to type out the code and talk about it as you go along. As an example, have a look those we've posted at #7 (comment), #8 (comment) and #5 (comment) if you have time to get a rough idea on the style of the tutorial. |
|
@weiji14 sorry for the long delay. Take a look at https://github.com/andrebelem/MarsMaps. I think everything is working fine in the last final tests. I'm still wondering if the |
Yep, just tested it locally on my computer and it seems to run fine. Not to pressure you, but I really really need the video by tomorrow (17 May) since 18 May is the hard deadline (and I still need to combine all of the videos together). To keep the video under 20 minutes definitely have the PyGMT install done beforehand, and start from the
We can add |
|
Sure ! I have already the whole thing in the Filmora. Just let me know where to cut (may be by the very end where I talked a little more about the inset and options of the 3D map). |
|
@leouieda @weiji14 @meghanrjones I don't know... I didn't just want to talk too fast because actually the most important thing is that the user understands what they are doing and why the choices of options. My students always ask me to slow down when I'm showing pygmt or python because they want to do it together. I'm cutting out some unimportant parts and it's already reducing well. However, I remember seeing other EGU courses in the past and they were a little more flexible in terms of timing. Perhaps, one option is to leave the 3x20 videos you made already there on the EGU and move this one to pygmt's youtube as a "bonus track" for Mars fans. What do you think ? |
|
Given the upload deadline, I think the simplest solution would be to cut the video at 19:19 (right before going over the grdview component) and record a short 40 second snippet to be placed in between this video and Wei Ji's that describes the contents of the remainder of the video and directs viewers to the GMT YouTube channel for the extended version. |
|
Thanks @andrebelem, I think you've actually explained the concepts at a good pace, it's definitely oriented at a beginner level which is the audience we're trying to target for this short course. Unfortunately 20 minutes is very short, and so we need to be quite selective of what we show (and yes that can be hard).
I really don't want to leave you out of the official video. At the very least, if you could get me a video cut of 10-20 min with just 1 cool plot of Mars, I can put that into our EGU22 submission, and then we can have a longer 'director's' cut for you on Youtube 😉 Personally I'd jump straight to the last map you made using
Or yes, we can do this too. Are you cutting the video right now @andrebelem? Or doing a re-recording? Just don't want to take up too much of your time anymore. |
|
I can do a new record just jumping to the section of 3D map of Olympus Mons and inset (for sure), but I agree with @meghanrjones idea - just cut in 20 min and put a snippet. I have an appointment in 3 min (+/- 15 min) and I'll be back recording a very fast-last-minute-20min-video just after. But plan A is just 20min cut in this we have already. |
|
Ok, you'll have until tonight to make another recording. Or just let me know otherwise, and I can make a cut of your current video at 19:19 as Max suggested. I'm not a video editing pro by any means, just stitching the videos together with Openshot and have a few fade-out-fade-in transitions 🙂 |
@andrebelem, if we cut the current video and include a snippet, I think it would flow better for you to summarize the extended version. But @weiji14's suggestion for making focusing on grdview and inset also sounds good to me. |
|
Got it - 19'58". Just converting and in a few minutes will be available here. |
|
@weiji14 here is the video https://drive.google.com/file/d/14MVz6lgB7lLxpESs7FJqoPcvwG6SwlZC/view?usp=sharing |
|
Perfect, thank you so much Andre! Glad that you found the video making relaxing, I can only imagine what drama you must gone through to get to this in the end. I'll start putting the video together and get the final cut uploaded by tomorrow. One more thing (but no rush), if you could upload your final jupyter notebook to the repo at https://github.com/GenericMappingTools/egu22pygmt/upload/main/book (via a Pull Request) before EGU22 next week, we can get that rendered on the website so people can look at the material easily. |
|
@andrebelem the video looks great! Thank you for all the effort you put into this during what sounds like a very difficult time. I hope things get better for you soon and that we can all meet up at some point. |
|
Looking at the video and book format at #10, would it be better to upload each section separately to our YouTube? We can then easily embed or link to the videos in each tutorial separately within their respective pages. We can bind things together with a playlist on YouTube (linked in the front page of the tutorial). That would make it a bit more approachable since it's a few 20 min videos instead of one 1.5h video. |
|
Hi guys, I just made the pull request as requested by @weiji14, with a short and an extended version of the notebooks. I definitely need to take a course on how to better use github hahaha 🤖 I thank everyone for understanding. In fact, all the little difficulties with timing, recording, notebooks, zoom that doesn't work, people knocking on the door here, were really "relaxing". Really. I'm fine, and the problem in the family is a terminal stage where all we can do is wait and give comfort. Impermanence is something I learned a while ago, and doing what I like (like building the notebook and maps, making the video and interacting with you guys) makes all the difference to mental health. I just have to thank you, and say "I want more"! What is the next step ? I'm ready. Forgot to reply: @weiji, that first block of the extended version of the Mars notebook with colab works great. It basically updates python and installs GMT from scratch in the Colab VM. It takes about 5-8 minutes, but after that the entire notebook can be run easily. I didn't add the colab button on the notebook because I didn't see a need, since pretty much all other notebooks are in the binder. However, I can make a COMBO version with all notebooks specifically for colab if you want. The advantage I see over Binder comes from a local experience, where students here massively use google tools and are more used to colab click-and-go than Binder. |
|
Great job with the videos and tutorials everyone - they all look great! @weiji14, feel free to ping if there's anything in particular that you want help with for wrapping it all together. |
|
@weiji14 I made a small modification to the extended script that allows you to "skip" the block relative to the colab (by default). In case you need it, it's on my github. |
|
Ok, finally got the intro done and the final cut is being exported now using OpenShot (:octocat: Open Source all the way :sunglasses:). Fingers crossed :crossed_fingers: that this fits under 3GB, otherwise I'm gonna freak out.
Yes, definitely need some helping hands! @meghanrjones, could you please help review @andrebelem's PR at #15, and include the "skip" block from https://github.com/andrebelem/MarsMaps into that PR? I'd like to get #15 finalized before merging #10. Doesn't have to be done today but best to do it sometime this week before EGU22 starts. |
will do! Do you plan to add it to the toctree in #10 or should I add a quick commit to #15? |
Just do a quick commit on #15 to add it to the toctree. I'll resolve any merge conflicts in #10 later. P.S. You can add nested sections following https://jupyterbook.org/en/stable/structure/sections-headers.html#how-headers-and-sections-map-onto-to-book-structure, may be helpful to put Andre's two tutorials under one section. |
|
Hello guys, |
|
Just heard back from the organizers and there should be a "Live Session" button at https://meetingorganizer.copernicus.org/EGU22/session/43186 15min before the short course starts at 13:10UTC/15:10CEST. I'll send out a tweet using the GMT twitter account now-ish, the short course should start in about half an hour from now. |
|
Thanks !
If I understand correctly, only those who are subscribed to the EGU will be
able to enter this live, even the conveners (I believe that access will be
via copernicus login). This weekend I was even quite frustrated with it!
EUR 210 for me to sign up !! Guys, this is practically 3x what the
Brazilian government is giving as COVID aid and a significant amount of
money here.
One thing I can do is stay on standby (discord ? for example or google meet
?) for anyone who wants to clear up any doubts. What do you think ?
Em ter., 24 de mai. de 2022 às 09:43, Wei Ji ***@***.***>
escreveu:
… Just heard back from the organizers and there should be a "Live Session"
button at https://meetingorganizer.copernicus.org/EGU22/session/43186
15min before the short course starts at 13:10UTC/15:10CEST (see
https://www.timeanddate.com/worldclock/fixedtime.html?msg=EGU22+SC5.2%3A+Crafting+beautiful+maps+with+PyGMT&iso=20220524T1510&p1=259&ah=1&am=30
).
I'll send out a tweet using the GMT twitter account now-ish, the short
course should start in about half an hour from now.
—
Reply to this email directly, view it on GitHub
<#1 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AK42FD72GK4B2Z72DOCNFYTVLTFHNANCNFSM5R6OKIKQ>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
--
Prof. Andre L. Belem (Dr. rer. nat.)
H2O - Grupo de Pesquisa em Hidrodinâmica, Hidráulica e Oceanografia
ORCID: https://orcid.org/0000-0002-8865-6180
ResearchGate: https://www.researchgate.net/profile/Andre_Belem
Linkedin: https://www.linkedin.com/in/andre-l-belem e
https://www.linkedin.com/showcase/agronautas
Departamento de Engenharia Agrícola e Meio Ambiente - Escola de Engenharia
- Universidade Federal Fluminense
Rua Passo da Pátria, 156 bloco E sl 300
Campus da Praia Vermelha - São Domingos
24210-240 Niterói,RJ / Brasil
|
|
I understand your frustration. Like, I'm registered and can see the video, but can't enter the live session for some reason...
Edit: if I understand correctly, it's only the hybrid sessions that have an associated Zoom link. Our pre-recorded session won't actually have a 'LIve Session' associated with it even though there's a button... There's a way to do a virtual pop-up networking event at https://webforms.copernicus.org/EGU22/pop-up-networking. I suppose we can organize one? |
|
Ok, have posted to the following channels:
Feel free to reshare those. I'll close this issue and we can open separate ones to track the Youtube video upload and pop-up-networking stuff. |
The figure is different. How to find the updated codes ? |
|
Hello @ditafaith, There has been a significant update to the PlanetaryMaps repository. Yes, you can absolutely apply the tool to any region of Mars! However, keep in mind that the results will depend on the resolution of the MOLA data. Cheers |
|
Thanks @andrebelem. Interesting to dive in more deeper. I am looking for 3D map using geotiff or common DEM data. |
and others
Continuing on from GenericMappingTools/website#131 (comment), we're going to have a 10 min intro + 4x20min sessions (total 90 min) for the PyGMT short course on 24 May 2022. This issue is for discussing what sort of material we'd like to teach, and sort out who is doing what (i.e. logistics).
👉 Draft outline: https://hackmd.io/@pygmt/egu22pygmt_outline
Important dates:
Note that the short course content can be pre-recorded ahead of time. I'm suggesting that we set an individual deadline on
Friday 6 May 2022Tuesday 10 May 2022 to have the presentation material (in mp4 format) ready for upload (though let me know if this doesn't work). This should give us time to review each other's work, make corrections, and merge the presentation together into the final 1.5 hour presentation (max 2-3GB) by the hard deadline on Wednesday 18 May 2022. Note that the EGU upload 'portal' is open from 9–18 May 2022 according to https://egu22.eu/guidelines/short-course_guidelines.html.Also, @meghanrjones suggested at GenericMappingTools/website#131 (comment) that we can have a think about having some sort of interactive component (e.g. gather.town with the participants).
@leouieda, @meghanrjones, @andrebelem.
I've set up a poll at https://www.when2meet.com/?15145108-mijPx so that we can have a 1 hour Zoom meeting to discuss this in more detail.(Edit: cancelling the meeting since hard to get everyone together, see #1 (comment)) But feel free to jot down any ideas you have below!The text was updated successfully, but these errors were encountered: