Re-trying config branch

[ldecicco-USGS]

@lindsaycarr could you give this a good vetting? Maybe run your AQCU stuff? In the meantime, I'll try to figure out the best way to add some tests. Here's what I'm running locally:

superFat.yaml:

lwd: 3
lines: 
  lwd: 3
abline:
  col: "grey"
  lty: 1
  lwd: 2

df <- data.frame(x = 1:10, y=1:10, z = seq(2,20,2))
loadConfig("D:/LADData/RCode/gsplot/superFat.yaml")
gsp <- gsplot() %>%
  lines(df$x, df$y, col="red", legend.name = "points") %>%
  lines(df$x, df$z, legend.name = "lines") %>%
  legend()
gsp

loadConfig()
gsp <- gsplot() %>%
  lines(df$x, df$y, col="red", legend.name = "points") %>%
  lines(df$x, df$z, legend.name = "lines") %>%
  legend()
gsp

gspFAT <- gsplot(config.file = "D:/LADData/RCode/gsplot/superFat.yaml") %>%
  lines(df$x, df$y, col="red", legend.name = "points") %>%
  lines(df$x, df$z, legend.name = "lines") %>%
  abline(v=4) %>%
  legend()
gspFAT

gsp <- gsplot() %>%
  lines(df$x, df$y, col="red", legend.name = "points") %>%
  lines(df$x, df$z, legend.name = "lines") %>%
  legend()
gsp

Then, from our last pr:

g1 <- gsplot() %>% 
  lines(1:9, seq(1,3, by=0.25), 
        legend.name="second line") %>% 
  # formatting y axes
  axis(side=2, at=1:10) 

layout(matrix(1:2, 1, 2))
g1
g1

and

par(mar=c(5,4,10,2))

g1 <- gsplot() %>% 
  points(1:10, 1:10, pch=20, legend.name="first points") %>% 
  lines(4:1, 4:1, legend.name="first line") %>% 
  points(c(3,7,4), c(9,3,6), pch=20, col="black", legend.name="second points") %>% 
  legend(location="above")
g1

[ldecicco-USGS]

(I don't have time to paste in the images, but I'll do that tomorrow too!....ideally you can imagine the first one has fat lines, the second is normal, the third fat, the fourth normal, then the last 2 are your examples)

[coveralls]

Coverage decreased (-0.02%) to 79.424% when pulling 10dd479266efd1395778595e14b3b778981330fa on ldecicco-USGS:master into 982628fdf23b5b7bf5e2d5e731956892c41a6053 on USGS-R:master.

[coveralls]

Coverage increased (+3.1%) to 82.557% when pulling 35aa49425e1a3ec32f517ee7c43d62d9050b0066 on ldecicco-USGS:master into 982628fdf23b5b7bf5e2d5e731956892c41a6053 on USGS-R:master.

[coveralls]

Coverage increased (+3.1%) to 82.557% when pulling 35aa49425e1a3ec32f517ee7c43d62d9050b0066 on ldecicco-USGS:master into 982628fdf23b5b7bf5e2d5e731956892c41a6053 on USGS-R:master.

[lindsayplatt]

If you try that last one with the fat lines config, it loses the default color ('red') and seems to change the point size:

par(mar=c(5,4,10,2))
g1 <- gsplot() %>% 
    points(1:10, 1:10, pch=20, legend.name="first points") %>% 
    lines(4:1, 4:1, legend.name="first line") %>% 
    points(c(3,7,4), c(9,3,6), pch=20, col="black", legend.name="second points") %>% 
    legend(location="above")
g1

par(mar=c(5,4,10,2))
g1 <- gsplot(config.file = "D:/LRCData/R/gsplot/superFat.yaml") %>% 
    points(1:10, 1:10, pch=20, legend.name="first points") %>% 
    lines(4:1, 4:1, legend.name="first line") %>% 
    points(c(3,7,4), c(9,3,6), pch=20, col="black", legend.name="second points") %>% 
    legend(location="above")
g1

[ldecicco-USGS]

I think that's right. The second example should lose the red color since it's not looking at the default config (which is where we are saying use red for points).

The fatter points are consistent with base R:

plot(1:10, 1:10, type="p", pch=20, lwd=3)

vs:

plot(1:10, 1:10, type="p", pch=20)

[lindsayplatt]

Ahh ok - I guess I was thinking of this as someone appending to or modifying the default config.

[ldecicco-USGS]

oy vey, I suppose that will be something someone requests next. Let's start easy and add a "append" type argument later.

[lindsayplatt]

When this ran with repgen, it passes all the tests but there is not legend appearing even though it is in the object. Might be a margin issue?

[lindsayplatt]

testing_legendissue.txt

It just seems to be when it's in an Rmarkdown that the legend can't showup. It also appears on my master branch, so this must not be associated with your changes...hunting to figure out when it did get changed.

[ldecicco-USGS]

Hmm...seems like if you include the fig.height=10, fig.width=13 in the chunk options, then it loses the legends. If you just specify the width, and let it figure out the height, I think it works.

[lindsayplatt]

library(gsplot)
par(mar=c(5,4,10,2))

g1 <- gsplot() %>% 
  points(1:10, 1:10, pch=20, legend.name="first points") %>% 
  lines(4:1, 4:1, legend.name="first line") %>% 
  points(c(3,7,4), c(9,3,6), pch=20, col="black", legend.name="second points") %>% 
  lines(1:2, c(1,4), legend.name="second line") %>% 
  legend(location="above")
g1

Resetting to #400 does not ignore mar:

Resetting to #410 does ignore mar:

@jread-usgs any ideas what it could be from? It uses par the second time to call g1 if you don't sweep the plotting region first.

[lindsayplatt]

Hmm but why would it have worked with specifying both before? We need to specify both fig.height and fig.width for repgen.

[coveralls]

Coverage increased (+3.08%) to 82.137% when pulling 565e642e2dce18ea79d0a83a97aceb3ea60f2304 on ldecicco-USGS:master into 745a3b02029b13311b6d2edd90fb2b08b1553db1 on USGS-R:master.

[ldecicco-USGS]

Just so I understand what that last comment is talking about...This current pull request does not ignore the mar command in normal R scripting. The issue is that the behavior in an Rmd file has changed? But if you go back to #400, the Rmd knitting looks right?

[lindsayplatt]

margin/par issue linked to #410. Investigating separately.