add note to view vignettes online

[ci skip]

DESCRIPTION

@@ -2,7 +2,7 @@ Package: loo
 Type: Package
 Title: Efficient Leave-One-Out Cross-Validation and WAIC for Bayesian Models
 Version: 2.3.1
-Date: 2020-07-13
+Date: 2020-07-14
 Authors@R: c(person("Aki", "Vehtari", email = "[email protected]", role = c("aut")),
              person("Jonah", "Gabry", email = "[email protected]", role = c("cre", "aut")),
              person("Mans", "Magnusson", role = c("aut")),

vignettes/children/SEE-ONLINE.txt

@@ -0,0 +1,2 @@
+**NOTE: We recommend viewing the fully rendered version of this vignette online at
+https://mc-stan.org/loo/articles/**

vignettes/children/SETTINGS-knitr.txt

@@ -1,4 +1,4 @@
-```{r, SETTINGS-knitr, include=FALSE}
+```{r SETTINGS-knitr, include=FALSE}
 stopifnot(require(knitr))
 opts_chunk$set(
   comment=NA,

vignettes/loo2-example.Rmd

@@ -15,6 +15,9 @@ params:
 ```{r, child="children/SETTINGS-knitr.txt"}
 ```
 
+```{r, child="children/SEE-ONLINE.txt", eval = if (isTRUE(exists("params"))) !params$EVAL else TRUE}
+```
+
 # Introduction
 
 This vignette demonstrates how to use the __loo__ package to carry out 
@@ -35,7 +38,7 @@ encourage readers to refer to the following papers for more details:
 In addition to the __loo__ package, we'll also be using __rstanarm__ and 
 __bayesplot__:
 
-```{r, setup, message=FALSE}
+```{r setup, message=FALSE}
 library("rstanarm")
 library("bayesplot")
 library("loo")
@@ -72,7 +75,7 @@ Because the number of days for which the roach traps were used is not the same
 for all apartments in the sample, we use the `offset` argument to specify that
 `log(exposure2)` should be added to the linear predictor.
 
-```{r}
+```{r data}
 # the 'roaches' data frame is included with the rstanarm package
 data(roaches)
 str(roaches)
@@ -86,7 +89,7 @@ roaches$roach1 <- roaches$roach1 / 100
 We'll fit a simple Poisson regression model using the `stan_glm` function from
 the __rstanarm__ package.
 
-```{r, count-roaches-mcmc, results="hide"}
+```{r count-roaches-mcmc, results="hide"}
 fit1 <-
   stan_glm(
     formula = y ~ roach1 + treatment + senior,
@@ -121,7 +124,7 @@ to the `loo` function (see, e.g. `help("loo.array", package = "loo")`). We'll
 also use the argument `save_psis = TRUE` to save some intermediate results to be
 re-used later.
 
-```{r}
+```{r loo1}
 loo1 <- loo(fit1, save_psis = TRUE)
 ```
 
@@ -130,7 +133,7 @@ some observations the leave-one-out posteriors are different enough from the
 full posterior that importance-sampling is not able to correct the difference.
 We can see more details by printing the `loo` object.
 
-```{r}
+```{r print-loo1}
 print(loo1)
 ```
 
@@ -159,7 +162,7 @@ Using the `plot` method on our `loo1` object produces a plot of the $k$ values
 with horizontal lines corresponding to the same categories as in the
 printed output above.
 
-```{r, out.width = "70%"}
+```{r plot-loo1, out.width = "70%"}
 plot(loo1)
 ```
 
@@ -181,7 +184,7 @@ the LOO-PIT values for our model (thick curve) is compared to many
 independently generated samples (each the same size as our dataset) from the
 standard uniform distribution (thin curves).
 
-```{r}
+```{r ppc_loo_pit_overlay}
 yrep <- posterior_predict(fit1)
 
 ppc_loo_pit_overlay(
@@ -202,17 +205,17 @@ regression, which is commonly used for overdispersed count data.
 Unlike the Poisson distribution, the negative binomial distribution 
 allows the conditional mean and variance of $y$ to differ.
 
-```{r, count-roaches-negbin, results="hide"}
+```{r count-roaches-negbin, results="hide"}
 fit2 <- update(fit1, family = neg_binomial_2)
 ```
 
-```{r}
+```{r loo2}
 loo2 <- loo(fit2, save_psis = TRUE, cores = 2)
 print(loo2)
 ```
 
 
-```{r}
+```{r plot-loo2}
 plot(loo2, label_points = TRUE)
 ```
 
@@ -232,7 +235,7 @@ calculations are performed exactly for that observation. The results are then
 recombined with the approximate LOO calculations already carried out for the
 observations without problematic $k$ values:
 
-```{r}
+```{r reloo}
 if (any(pareto_k_values(loo2) > 0.7)) {
   loo2 <- loo(fit2, save_psis = TRUE, k_threshold = 0.7)
 }
@@ -248,7 +251,7 @@ still some degree of model misspecification, but this is much better than the
 `p_loo` estimate for the Poisson model.
 
 For further model checking we again examine the LOO-PIT values.
-```{r}
+```{r ppc_loo_pit_overlay-negbin}
 yrep <- posterior_predict(fit2)
 ppc_loo_pit_overlay(roaches$y, yrep, lw = weights(loo2$psis_object))
 ```
@@ -263,7 +266,7 @@ the data.
 We can use the `loo_compare` function to compare our two models on 
 expected log predictive density (ELPD) for new data:
 
-```{r, count-roaches-loo}
+```{r loo_compare}
 loo_compare(loo1, loo2)
 ```
 

vignettes/loo2-large-data.Rmd

@@ -13,6 +13,9 @@ params:
 %\VignetteIndexEntry{Using Leave-one-out cross-validation for large data}
 -->
 
+```{r, child="children/SEE-ONLINE.txt", eval = if (isTRUE(exists("params"))) !params$EVAL else TRUE}
+```
+
 # Introduction
 
 This vignette demonstrates how to do leave-one-out cross-validation for large

vignettes/loo2-lfo.Rmd

@@ -26,6 +26,9 @@ knitr::opts_chunk$set(
 )
 ```
 
+```{r, child="children/SEE-ONLINE.txt", eval = if (isTRUE(exists("params"))) !params$EVAL else TRUE}
+```
+
 ## Introduction
 
 One of the most common goals of a time series analysis is to use the observed

vignettes/loo2-moment-matching.Rmd

@@ -15,6 +15,9 @@ params:
 ```{r, child="children/SETTINGS-knitr.txt"}
 ```
 
+```{r, child="children/SEE-ONLINE.txt", eval = if (isTRUE(exists("params"))) !params$EVAL else TRUE}
+```
+
 # Introduction
 
 This vignette demonstrates how to improve the Monte Carlo sampling accuracy of

vignettes/loo2-non-factorized.Rmd

@@ -23,6 +23,9 @@ knitr::opts_chunk$set(
 )
 ```
 
+```{r, child="children/SEE-ONLINE.txt", eval = if (isTRUE(exists("params"))) !params$EVAL else TRUE}
+```
+
 # Introduction
 
 When computing ELPD-based LOO-CV for a Bayesian model we need to

vignettes/loo2-weights.Rmd

@@ -15,6 +15,9 @@ params:
 ```{r, child="children/SETTINGS-knitr.txt"}
 ```
 
+```{r, child="children/SEE-ONLINE.txt", eval = if (isTRUE(exists("params"))) !params$EVAL else TRUE}
+```
+
 # Introduction
 
 This vignette demonstrates the new functionality in __loo__ v2.0.0 for 
@@ -273,11 +276,11 @@ a comparison to models without these terms.
 fit11 <- update(fit10, formula = total_tools ~ log_pop + contact_high)
 fit12 <- update(fit10, formula = total_tools ~ log_pop)
 ```
-```{r}
+```{r loo-contact_high}
 (loo11 <- loo(fit11))
 (loo12 <- loo(fit12))
 ```
-```{r}
+```{r relo-contact_high}
 loo11 <- loo(fit11, k_threshold=0.7)
 loo12 <- loo(fit12, k_threshold=0.7)
 lpd_point <- cbind(

vignettes/loo2-with-rstan.Rmd

@@ -13,6 +13,9 @@ params:
 %\VignetteIndexEntry{Writing Stan programs for use with the loo package}
 -->
 
+```{r, child="children/SEE-ONLINE.txt", eval = if (isTRUE(exists("params"))) !params$EVAL else TRUE}
+```
+
 # Introduction
 
 This vignette demonstrates how to write a Stan program that computes and stores