Add a Tidal Analysis vignette

[dankelley]

This came up in an email, and I'm making an issue so the work can be cross-referenced.

[dankelley]

@richardsc I just pushed (with new vignette) to GH, including the docs, so the webpage has a semi-blank entry for tides. I won't likely get time to put in the "univariate case" -- the tidem() case -- until next week. I don't think we will get merge conflicts, if you only edit in the "bivariate case" block and I only edit in the "univariate case" block. But even if we get conflicts, it's not like it will break code.

The best way to check the building of the vignette is to do that in RStudio. Please don't use pkgdown::build_site() because I think that inserts a date in every single html file (one per function, plus for vignettes and datasets) and your date and my date will conflict so every single file will have a merge conflict.

[richardsc]

I pushed some material, mostly based on what was on my blog post, but it's causing some crashes of the auto-build (not surprising, given the external dependencies required). Perhaps we can chat this week about how best to write/include a vignette that we don't actually want the system (e.g. CRAN) to build.

[dankelley]

Sure. Basically the answer is to enclose code in triple backticks. That won't colourize it, but I think "triple-backtick R" (that is, with no braces) might do that. For figures, they need to be included like normal markdown. We need to be tight on diagrams space so the fewer the better, and ideally use preview and save to jpg with enough compression to get say 100K or maybe 150K.

[richardsc]

Ok, so the idea is to create the output (figures etc) separately and then include it directly?

[dankelley]

Yup. Basically, don't run any code, unless the CRAN system will have all the software (which it won't ... no python, for example).

[richardsc]

👍

[dankelley]

@richardsc maybe you can post a comment sometime when you think you're done with your part of the tidal vignette. Obviously, there's no rush, but if you feel you're done, then I could start taking stuff from other vignettes and help files and put it into the univariate part. I'm holding off on that to avoid merge conflicts.