Use the new R decorator wrappers in the docs

getting-started/tutorials/season-2-scaling-out-and-up/episode05.md

@@ -53,8 +53,8 @@ metaflow("HelloAWSFlow") %>%
          r_function = start, 
          next_step = "hello") %>%
     step(step = "hello", 
-         decorator("retry", times=2),
-         decorator("batch", cpu=2, memory=2048),
+         retry(times=2),
+         batch(cpu=2, memory=2048),
          r_function = hello,  
          next_step = "end") %>%
     step(step = "end", 

metaflow-on-aws/metaflow-sandbox.md

@@ -11,7 +11,7 @@ Only a limited number of sandboxes are available. When you sign up, you are adde
 Here are some ideas that you can try with the sandbox:
 
 * [The season 2 of tutorials](../getting-started/tutorials/season-2-scaling-out-and-up/) focuses on scaling out. This is a good way to get started. Note that the Season 1 tutorials work with the Sandbox too, when executed using [the `batch` decorator](../metaflow/scaling.md).
-* You have up to 64 CPU cores at your disposal using [the `batch` decorator](../metaflow/scaling.md). Test some number crunching! You can run everything in the cloud simply by or you can mix local and remote steps by adding `decorator("batch",...)` to select steps.
+* You have up to 64 CPU cores at your disposal using [the `batch` decorator](../metaflow/scaling.md). Test some number crunching! You can run everything in the cloud simply by or you can mix local and remote steps by adding `batch(...)` to select steps.
 * Test your favorite ML libraries in the cloud using [`batch`](../metaflow/scaling.md) decorator. For instance, try a basic hyper-parameter search using [a custom parameter grid and foreach](../metaflow/basics.md#foreach). 
 * Evaluate Metaflow's [experiment tracking and versioning](../metaflow/tagging.md) using local runs and the [Client API](../metaflow/client.md) in a local notebook. In contrast to the local mode, all runs are registered globally in the Metaflow Service regardless of the directory where you run them.
 * Test how you can [`resume` tasks locally](../metaflow/debugging.md#how-to-use-the-resume-command) which were originally run remotely using [the `batch` decorator](../metaflow/scaling.md).

metaflow/failures.md

@@ -33,7 +33,7 @@ end <- function(self){
 
 metaflow("RetryFlow") %>%
     step(step="start", 
-         decorator("retry"),
+         retry(),
          r_function=start, 
          next_step="end") %>%
     step(step="end", 
@@ -115,7 +115,7 @@ you may end up withdrawing up to $4000 instead of the intended $1000. To make su
 metaflow("MoneyFlow") %>%
     ...
     step(step="withdraw", 
-         decorator("retry", times=0),
+         retry(times=0),
          r_function=withdraw_money_from_account, 
          next_step="end") %>%
     ...
@@ -195,7 +195,7 @@ metaflow("CatchFlow") %>%
          next_step = "sanity_check",
          foreach = "params") %>%
     step(step = "sanity_check", 
-         decorator("catch", var="compute_failed", print_exception=FALSE),
+         catch(var="compute_failed", print_exception=FALSE),
          r_function = sanity_check, 
          next_step = "join") %>%
     step(step = "join", 
@@ -234,8 +234,8 @@ end <- function(self) {
 
 metaflow("SuicidalFlowR") %>%
   step(
-    decorator("catch", var = "start_failed"),
-    decorator("retry"),
+    catch(var = "start_failed"),
+    retry(),
     step = "start",
     r_function = start,
     next_step = "end"

metaflow/scaling.md

@@ -39,7 +39,7 @@ end <- function(self) {
 
 metaflow("BigSumFlowR") %>%
   step(
-    decorator("resources", memory=60000, cpu=1),
+    resources(memory=60000, cpu=1),
     step = "start",
     r_function = start,
     next_step = "end"
@@ -57,7 +57,7 @@ This example creates a huge 80000x80000 random matrix, `big_matrix`. The matrix
 If you attempt to run this on your local machine, it is likely that the following will happen:
 
 ```bash
-Evaluation error: vector memory exhausted (limit reached?).
+Evaluation error: vector memory exhausted (limit reached?).
 ```
 
 This fails quickly due to a `MemoryError` on most laptops as we are unable to allocate 48GB of memory.