linear_plateau

[austinwpearce]

• default to tidy = TRUE
• return tibble instead of dataframe (dplyr has tibble and as_tibble so no new dependencies)
• if target = NULL, defaults to join point (promote the use of join point)
• section # CSTV for plateau or for target mistakenly checked if target is greater than CSTV, but it should be greater than plateau. Now it reports that targets greater than plateau will result in CSTV at join point.
• simplified code repetition for readibility concerning the coef(lp_model)[[3]] by saving join point to jp variable instead
• join point is now referred to as jp throughout the internal code - equation now shows CSTV/JP value instead of text "CSTV"
• lp_line() shouldn't use seq() because the resulting line will not break cleanly
• STVt should be reported right after target in the table
• Use lowerCL and upperCL instead of LL_cstv and UL_cstv
• replace "Wald Conf. Interval" with "Wald, 95%"
• Always calculate and report Wald CI for CSTV (join point parameter) now that we have a STVt to find a soil value at the target. The STVt doesn't represent CSTV
• cleaned up plotting code (line wraps, vlines with c(lower, upper) instead of two vlines)
• move LP line up so it is underneath vlines
• change size to linewidth arg in the line geoms
• propose alternative bootstrap function
• I THINK WE ARE NOT KEEPING THIS CHANGE, FOR NOW AT LEAST
• • by using modelr::bootstrap, which is already a dependency for rsquare, we avoid multiple extra lines of code*
• likewise, because the default of linear_plateau is now a tibble() we don't need as.data.frame or tidy = TRUE
• removed extra irrelevant columns like Wald CI and equation for neater tibble

test-linear_plateau

• do we need to specify defaults in the test? like plot = FALSE, tidy = TRUE

• cleaned up alignment some

[adriancorrendo]

Good to go and update the rest of the functions with the same line!

Thanks!