It could be the third tab.

I might have been typing quite quickly. I was trying to identify iTop and iBot near the peak of Atl water.

 file <- "~/data/arctic/beaufort/2012/d201211_0048.cnv"
> ctdTag(file=file)

Listening on http://127.0.0.1:5456
Warning in min(x) : no non-missing arguments to min; returning Inf
Warning in max(x) : no non-missing arguments to max; returning -Inf
Warning in min(x) : no non-missing arguments to min; returning Inf
Warning in max(x) : no non-missing arguments to max; returning -Inf
Warning: Error in plot.window: need finite 'xlim' values
  174: plot.window
  173: localWindow
  172: plot.default
  169: renderPlot [/Library/Frameworks/R.framework/Versions/4.3-x86_64/Resources/library/ctd/shiny/ctdTag/app.R#629]
  167: func
  127: drawPlot
  113: <reactive:plotObj>
   97: drawReactive
   84: renderFunc
   83: output$plot
    2: runApp
    1: ctdTag [/Users/kelley/git/ctd/R/ctdTag.R#67]

This should be an easy fix

I need to change the meaning of the isTagged column of the files table, so the initial value can mean not-evaluated-yet. Luckily, sqlite3 doesn't have a boolean type, so it's already an integer :-)

@cameronrichrdson -- as discussed separately, perhaps this week you could have a go at

1. making a new tab, to the left of Help, called Settings
2. moving 'quit' to Settings
3. moving the arrows up to the top line, where there is likely enough space

This ought to be done in a fork, with changes in a PR. Please be sure to use the exact same indentation and spacing conventions. (Do not follow what RStudio might want you to do.)

As we link for tags and files, we should link for analyst names. That way, we could have extra info, like contact information, if desired.

I'll bump the database version number when I do this, so that ctdTag() will automatically transform old databases to this new form. That is, no user-visible changes will result.

It happens after 'returning nearest'

[0.001049]   returning nearest=276, relativeDistance=0.0097
[0.000086] [4.399178] findNearestLevel(x=0.1216139, y=277.1692, ..., view=T profile)
[0.001145]   returning nearest=276, relativeDistance=0.0055
[0.000094] [4.330659] findNearestLevel(x=0.1204382, y=277.0242, ..., view=T profile)
[0.000955]   returning nearest=276, relativeDistance=0.0010
[0.000089] [2.148689] keypress being handled since not plotting

I need to think more about whether e.g. data$ylabProfile should be used, or state$data$ylabProfile. The latter makes it reactive.

I think some of the odd non-noticing (e.g. a problem with y label, noticed in a meeting with Clark and Cameron this morning) might come down to my using data* instead of state$data*. (To Cameron -- this will make sense to you at some point.)

ctdTag() now responds to double-click, not single-click, so there is no problem of remnant brushes.

This will be an easy fix, but not my top priority right now.

I'll finish this work tomorrow, most likely. It is interlinked with #13 and #12, actually, in the sense that solutions to those will rest on the new schema.

I'm not sure how to describe the condition.

I will assign this to CaRi when I find the right username. Again, it would be a good learning experience (of vim and pull-requests).

I use unix-level commands to do this, but that requires installing some software. I'll write something to do this at the R level.

This should be an easy fix

Choices seem to be:

1. Move view to show the focus (might be hard to figure out, e.g. need to communicate the zoom scale between views)
2. Detect if the focus is offscreen, and make a notification to that effect.
3. Only permit tagging if the focus point is onscreen.

I sort of like method 3. Perhaps @cameronrichrdson has other ideas?

We want the table to be uniform, so ctdTag() can determine if the file has already been examined. But, at the moment, if I do

ctdTag(file="~/data/arctic/beaufort/2012/d201211_0047.cnv")

and later do

ctdTag(file="/Users/kelley/data/arctic/beaufort/2012/d201211_0047.cnv")

then they will appear in the database as two separate files, even though they are the same file. So, for example, if I do the first form and tag some points, but then use the second form on the same file, the points will not be tagged.

Users may use either the ~ notation or the /Users/USERNAME notation in the filesystem, with identical results, so ctdApp ought to do the same.

I think it makes sense to convert to ~ notation because that will be uniform across users, so long as they put their data files in the same locations relative to their home directories.

I will think about this for a while and then decide on what I want to do, this afternoon.

NOTE: this will be done without the user needing any action. If I do decide to go ahead with the idea, ctdTag() will translate the database when it loads it. (It already does significant changes, to maintain compatibility with two previous versions.)

The snapshot illustrates. For a TS plot, the user should not be able to alter whether the y axis is pressure or sigmaTheta. (I will assign this to CaRi when I know the right GH username to use ... we could work on this together sometime as a learning exercise.)