Matthew Alger comments

Results 22 comments of


                                            Matthew Alger

Create coastal masking `Frequently_used` notebook

We do sort of have one, here: https://docs.dea.ga.gov.au/notebooks/Scientific_workflows/DEAWaterbodies/DEAWaterbodiesSupplement/CreateAustralianCoastlineUsingITEM.html

Codify the notebook PR best practices as GitHub Actions

The action would delete empty cells, rather than rejecting the PR, right?

Speed up xr_vectorize

No, but I have postprocessing code on the `subpixel_contours` to get them in the same format, and it's not the bottleneck as far as I can tell

Add more virtual product examples (or even a bunch of virtual products)

I'd kinda like multiple YAML files - a monolithic YAML file is hard to navigate and understand, as I discovered when I tried to learn about virtual products the other...

Add more virtual product examples (or even a bunch of virtual products)

@uchchwhash I didn't save the error messages - whoops! Perhaps I can try naively making some virtual products and see what issues I hit, as a user who "gets" the...

Add more virtual product examples (or even a bunch of virtual products)

https://gist.github.com/MatthewJA/e754ab7ed6c58f582e6fe9ea57e699d5 Here's the MNDWI + WOfS YAML I pulled out of #641. A lot messier but it shows how to do this kind of index.

Error loading bom_rainfall_grids

Are you part of rr5 on the NCI?

Error loading bom_rainfall_grids

Yep, that data seems to have been deleted. Anyone know why or where it went?

Allow `dea_waterbodies.get_waterbodies` to accept multiple geohashes, and allow `dea_waterbodies.get_time_series` to accept multiple geohashes

There's some technical issues that make this a bit tricky, but we can for sure do it (possibly inefficiently)

maxent seems to be using max instead of softmax for V_soft?

Hey, thanks for the comments. It could very well be a bug (including the initialisation of gridworld) — it's been a long while since I've looked at inverse reinforcement learning...