Day-2 code: make node agnostic

[CaitlinBate]

ensure all the day 2 code works for

• [ ] -act
• [ ] fact
• [ ] glatos
• [ ] migramar (can use act code as a template for edits - data is same format)

most of these are done and can be checked off

changes need to work for

• "intro to glatos
• "more features of glatos"
• basic viz and plotting
• preparing data for actel (it appears this lesson is only ACT format right now. need to add FACT, GLATOS, MIGRAMAR)

changes need to appear in:

• gh-pages episode code
• master code folder

. no time urgency here.

[jackVanish]

I'll settle in and get started on this now.

[jackVanish]

One thing while I'm working on this, I don't think gh-pages has a code folder anymore? IIRC we got rid of it so that we only had to maintain the code in one place (master).

[jackVanish]

Had to go do some Remora stuff but I'm noting down a couple of things I encountered and will need to fix here, for posterity or in case someone wants to check them out before I do.

• Episode 8's call to prepare_tag_sheet needs to be updated; the function has a new signature with new names/ordering for the arguments. Should be tags <- prepare_tag_sheet('Tag_Metadata/Proj58_Metadata_cownoseray.xlsx', header_line = 5, sheet_name = 2)
• Also in episode 8's code (in day2) there's a mutate on line 178 that isn't addressed in the lesson code and has no comments around it. It doesn't seem to run because full_dates doesn't exist at that point. Thoughts?

[jackVanish]

The above comment deals with stuff that was in the ATT section, which has since been removed. GLATOS ACT code seems to work fine but the bubble plot by arrays at the very end seems to show some arrays inland. Can someone have a look and verify that all is well?

[jackVanish]

A couple of notes regarding FACT version of the code.

1. The alternative method for getting the polygon to plot against (included below) breaks on the readRDS line: #Alternative method of getting the polygon. f <- 'http://biogeo.ucdavis.edu/data/gadm3.6/Rsp/gadm36_USA_1_sp.rds' b <- basename(f) download.file(f, b, mode="wb", method="curl") USA <- readRDS('gadm36_USA_1_sp.rds') FL <- USA[USA$NAME_1=="Florida",]

2. The FACT version lacks the bubble by array plot, which produces the wonky output above in ACT. After a brief chat with Jon I'm wondering if we want to still include bubble plot by array in the lesson. If so, let me know so that I can put it in there.

[jdpye]

my reasoning in that conversation: The by-array breakdown isn't terribly informative even when it indicates the different gates and their names, and after the cropping we've done by setting the map boundaries it might even be misleading.

The four or so lines of code you pasted in there for the polygon fetch and plot look like they've been smashed together. Assuming the third line starts with USA <- readRDS(.... and that's the one that died?

[jackVanish]

There is now a fresh Migramar lesson up that runs pretty well and seems to spit out information that makes sense. Have a look in the branch if you like https://github.com/ocean-tracking-network/otn-workshop-base/tree/day_2_node_agnostic

[jdpye]

almost got it. :D

https://github.com/ocean-tracking-network/otn-workshop-base/blob/df4ebbe21e6931e40bf65ffe228d3f5f1b4313be/code/day2/1_GlatosCode_MIGRAMAR.R#L210 still has the old lat/lon bounds from the ACT lesson

[jackVanish]

Made a standardized GLATOS code file so that gives us code for each of the four nodes. Next is to get the Actel stuff into a four-node spread, then updating all the lessons to make sure they are textually fine.

[jdpye]

The challenge of Actel is creating the groups of receivers that are 'locations', and the network diagram connecting the 'locations', for each of the regional example datasets.

[CaitlinBate]

glatos code comments:

• R script for ACT still has comments for slicing dataset to 100,000 rows (remove this)
• FACT script (lessons and gh-pages) is missing bubble_array and only has bubble_station - on purpose?
• R script for GLATOS has us creating a format variable in step1 that doesnt exist in the gh-pages version. i think it can be removed from the R scipt too
• GLATOS script (lessons and gh-pages) dont pull in a raster before doing the bubble plots - i assume these use a built in raster?
• FACT script (lessons and gh-pages) is missing bubble_array and only has bubble_station - on purpose?
• did any of the challenges (the solutions) get tested yet? theyre on gh-pages but not the master R scripts

[jackVanish]

OK, re the above notes:

1 & 3: fixed and committed 2 & 5: I can't say for sure, but I assume that they're missing on purpose. I'm not sure why, but if they're absent in both lessons and code it suggests that we chose not to include them back when we were originally knocking these together. At least, to me. 4: Yeah, GLATOS is using built-in stuff in those instances, I'm pretty sure.

6. I don't remember if I tested them. I remember updating the text to fix the coordinates, and I may have run them in the process, but I don't specifically remember if I checked them.

[CaitlinBate]

summary: all glatos lessons are node agnostic now

todo: button up actel lessons for other nodes (exists for ACT already). @jackVanish can speak to the progress

[CaitlinBate]

summary: all glatos lessons are node agnostic now

todo: button up actel lessons for other nodes (exists for ACT already). @jackVanish can speak to the progress

current status: nodes we need to support, for which we have example data: ACT, FACT, GLATOS, Migramar, OTN (new!!)

• day 1: node agnostic complete
• glatos lessons: need OTN version only
• actel intro: uses a generic dataset, this is OK
• actel data: needs support for OTN (will be very similar to act/migramar), GLATOS, FACT.
• actel custom spatial: needs support for all nodes outside ACT data I think
• basic animations needs support for FACT, GLATOS, MIGRAMAR
• pathroutr: needs support for FACT, GLATOS, MIGRAMAR
• remora: generic OTN dataset right now, which is OK for now??
• gams: generic non-node dataset which is OK for now
• yaps: generic non-node dataset which is OK for now