h2owave R Package (rdriver) Testing

[ashrith]

Hello and thank you for signing up to test h2owave package for R. Internally, it is called as rdriver by the wave team.

To be able to successfully run the h2owave library, and example codes please follow the steps below.

1. There are 5 libraries that are needed by h2owave library

   1. jsonlite
   2. devtools
   3. httr
   4. R6
   5. stringr
      You can install these packages by conveniently cutting and pasting the following code and by choosing a mirror:
      install.packages(c("R6","devtools","jsonlite","stringr","httr"))

2. Once these libraries are installed you can go ahead and install the h2owave package.

   1. This way:
      install.packages(“path_to_package/h2owave_0.0.14.tar.gz",repos=NULL,type="source")
   2. or from the SHELL this way
      $R CMD INSTALL path_to_package/h2owave_0.0.14.tar.gz

3. Download the wave server version 0.13 from here https://github.com/h2oai/wave/releases/tag/v0.13.0

4. Untar the wave server, and cd into the directory.

5. After installing the library you will now have to start (if you haven’t already) your wave server. At your SHELL you can run $./waved -debug

6. After installing the library you can start using the example applets in two ways:

   1. Line-by-line:
      1. Unzip the examples
      2. Open any one of the seven applets
      3. copy each line of code into the R interpreter environment and run it.
      4. The functionality of each one of code in each applet is described.
   2. Source:
      1. in the R interpreter type > source(“path_to_example/hello_world.R”)

7. You will be able to see the output of each applet on your browser at http://localhost:10101/demo

8. Once you have completed running of the code successfully, Please answer the following:

   1. Was the installation of the library easy?

   2. Did you understand each of the examples clearly?

      1. Was the code readable?

   3. Did code in the examples make sense to you as a R user?

      1. Was the code idiomatic?

   4. Were you able to execute the examples easily?

      1. Was the code readable?
      2. Was the code brief?
      3. Was the code ergonomic?

   5. Any other comments you would like to provide about the package? examples? experience?

We are making changes to the package as we speak, so if something is missing please do mention it. Especially docs. We would be happy to incorporate your feedback.

Updates:

1. Added devtools in the installation list.
2. Fixed the correct quotes """ enclosing httr
3. updated the h2owave_0.0.14.tar.gz (please get the new one here below)
4. Added specific information about wave server.

h2owave_0.0.14.tar.gz
pre-examples.tar.gz

[grigory93]

any requirements on R version?

[ashrith]

@grigory93 the package is built on R version 4.0.2 (2020-06-22). Let me know if you have an older/newer version and have any compatibility issues.

[ashrith]

@grigory93 I updated the library to reflect some changes. Please use the new one provided here.

[woobe]

need to fix the final double quotation mark in install.packages(c("R6","jsonlite","stringr","httr”)), it is ” instead of "

[woobe]

I needed to install ‘devtools’ before step 2.

[ashrith]

@woobe thanks for the quotes, and the devtools installation. I assumed that everybody would have devtools obviously I was wrong.

[ashrith]

@woobe updated the guide as per your comments! thanks.

[grigory93]

just wondering why do we need devtools ?

[woobe]

There was an error message about devtools and roxygen2 missing when I tried to install wave the first time (vanilla R 4.0.4 setup in a Virtualbox). I installed devtools (which also installed roxygen2) and the problem was gone.

[woobe]

All examples work and well-commented. I think we need one or two examples of visualization with popular packages like ggplot2 / plotly.

> sessionInfo()
R version 4.0.4 (2021-02-15)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 20.04.2 LTS

Matrix products: default
BLAS: /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so.3
LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/liblapack.so.3

locale:
[1] LC_CTYPE=en_GB.UTF-8 LC_NUMERIC=C LC_TIME=en_GB.UTF-8 LC_COLLATE=en_GB.UTF-8
[5] LC_MONETARY=en_GB.UTF-8 LC_MESSAGES=en_GB.UTF-8 LC_PAPER=en_GB.UTF-8 LC_NAME=C
[9] LC_ADDRESS=C LC_TELEPHONE=C LC_MEASUREMENT=en_GB.UTF-8 LC_IDENTIFICATION=C

attached base packages:
[1] stats graphics grDevices utils datasets methods base

other attached packages:
[1] h2owave_0.0.14

loaded via a namespace (and not attached):
[1] httr_1.4.2 compiler_4.0.4 magrittr_2.0.1 R6_2.5.0 tools_4.0.4 curl_4.3 tinytex_0.30 stringi_1.5.3
[9] jsonlite_1.7.2 stringr_1.4.0 xfun_0.22

[ashrith]

@woobe thank you for the feedback. We will add more examples and you will see that soon. The first set of examples were kept simple to assess the readability, brevity, and usability of the package and the examples.

[woobe]

I just installed the latest h2owave_0.0.14.tar.gz and reran all examples using wave server v0.13. All good!

[woobe]

Was the installation of the library easy? - YES

Did you understand each of the examples clearly? - YES

Was the code readable? - YES

Did code in the examples make sense to you as a R user? - YES

Was the code idiomatic? - YES

Were you able to execute the examples easily? YES (I would uncomment library(h2owave) in the first example hello_world.R

Was the code readable? - YES

Was the code brief? - YES

Was the code ergonomic? - I would split the comments into a few lines if they exceed the 80 char count (standard R margin)

Any other comments you would like to provide about the package? examples? experience? - Let's create more examples together!

[lo5]

Thanks for the review @woobe!

[brandenkmurray]

I'm having issues getting this to work on Mac. When I run one of the apps the wave server appears to be working as I can see stuff happening in terminal. For example, while stat_small.R runs I can see the following in terminal:

2021/03/26 09:53:38 * /demo {"d":[{"k":"crypto title","v":"LTC"},{"k":"crypto value","v":"131.654218069278"}]}
2021/03/26 09:53:42 * /demo {"d":[{"k":"crypto title","v":"LTC"},{"k":"crypto value","v":"375.661471097497"}]}
2021/03/26 09:53:45 * /demo {"d":[{"k":"crypto title","v":"BTC"},{"k":"crypto value","v":"95.5014126531314"}]}
2021/03/26 09:53:48 * /demo {"d":[{"k":"crypto title","v":"BTC"},{"k":"crypto value","v":"329.656990627758"}]}
2021/03/26 09:53:51 * /demo {"d":[{"k":"crypto title","v":"ETH"},{"k":"crypto value","v":"421.359045060351"}]}
2021/03/26 09:53:54 * /demo {"d":[{"k":"crypto title","v":"BTC"},{"k":"crypto value","v":"234.171027677134"}]}
2021/03/26 09:53:57 * /demo {"d":[{"k":"crypto title","v":"LTC"},{"k":"crypto value","v":"101.646449375199"}]}
2021/03/26 09:54:00 * /demo {"d":[{"k":"crypto title","v":"ZCH"},{"k":"crypto value","v":"303.540865081595"}]}
2021/03/26 09:54:03 * /demo {"d":[{"k":"crypto title","v":"LTC"},{"k":"crypto value","v":"432.879340720363"}]}
2021/03/26 09:54:06 * /demo {"d":[{"k":"crypto title","v":"ZCH"},{"k":"crypto value","v":"338.60062143812"}]}
2021/03/26 09:54:09 * /demo {"d":[{"k":"crypto title","v":"ZCH"},{"k":"crypto value","v":"4.90247914032079"}]}
2021/03/26 09:54:12 * /demo {"d":[{"k":"crypto title","v":"ETH"},{"k":"crypto value","v":"489.844166039256"}]}

However, when I navigate to localhost:10101/demo I just get a blank page. I've tried reloading, switching from Chrome to Safari, and changing demo to demo2 in the code, but nothing has worked so far. I also originally had R 3.6.0 and then upgraded to 4.0.4 -- didn't work with either. Any guidance for where to go from here?

R version 4.0.4 (2021-02-15)
Platform: x86_64-apple-darwin17.0 (64-bit)
Running under: macOS Catalina 10.15.7

Matrix products: default
BLAS:   /System/Library/Frameworks/Accelerate.framework/Versions/A/Frameworks/vecLib.framework/Versions/A/libBLAS.dylib
LAPACK: /Library/Frameworks/R.framework/Versions/4.0/Resources/lib/libRlapack.dylib

locale:
[1] en_GB.UTF-8/en_GB.UTF-8/en_GB.UTF-8/C/en_GB.UTF-8/en_GB.UTF-8

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
[1] h2owave_0.0.14

loaded via a namespace (and not attached):
[1] httr_1.4.2     compiler_4.0.4 magrittr_2.0.1 R6_2.5.0       tools_4.0.4    curl_4.3       stringi_1.5.3  jsonlite_1.7.2 stringr_1.4.0 

[ashrith]

@brandenkmurray Did you pick up the updated package in this discussion thread and install it? If not could you do that? The package sent in the email got a recent fix. Could you also share what version were you getting this problem? Was it 0.13?
Pasting the link again here https://github.com/h2oai/wave/files/6201635/h2owave_0.0.14.tar.gz

[brandenkmurray]

Was the installation of the library easy?

Yes. My only suggestion is to make the instructions clearer. Instruction 3 is to download wave server, then Instruction 4  says to run ./waved -debug. It might not be obvious to some that they need to extract the tar file and then navigate to the resulting directory before running the ./waved -debug command. 

Did you understand each of the examples clearly?

	Yes.

Was the code readable?

            Yes.

Did code in the examples make sense to you as a R user?

      Yes.

Was the code idiomatic?

		Yes

Were you able to execute the examples easily?

     Yes.

Was the code readable?

                In `layout.R` line 24, I would add another indent so it’s clear that line is run in the inner for-loop. Everything else looks good to me.

Was the code brief?

                Yes

Was the code ergonomic?

               Yes

Any other comments you would like to provide about the package? examples? experience?

            I think an example that’s a little more complex would be nice. One that combines together everything in the current examples. Like the one we have in the image at https://h2oai.github.io/wave/

[grigory93]

1. Was the installation of the library easy?
   Yes

2. Did you understand each of the examples clearly?
   Understood them clearly, but for people not familiar with R R6 OOP system this could be not straight forward to start. This is not to criticize R6 approach but rather to expect that not everyone is immediately familiar with it.

2.1. Was the code readable?
Yes, but suggest more spacing between lines of code and more logical grouping of code segments based on structure, such as initialization/setup, layout, interaction, updating.

3. Did code in the examples make sense to you as a R user?
   Yes, but, again, for those not familiar with R6 could be a learning curve.

Was the code idiomatic?
at the detail level Yes, at the higher app level No.

Were you able to execute the examples easily?
Yes

Was the code readable?
Yes

Was the code brief?
Yes, but examples were rather simple, so far.

Was the code ergonomic?
Yes, but again need to see more functional examples.

Any other comments you would like to provide about the package? examples? experience?
Having explicitly to place the code inside while(TRUE) loop (event loop I suppose) appears to be immature for a framework. There is hardly a case when this doesn't happen, unless it's a static image. So the questions are does the loop have to be explicit and why?

[grigory93]

Hopefully, having a basic/minimal/trivial example that involves consuming and reacting to user input will add some clarity.

[grigory93]

I've never used R6 system before so I just expressed my reaction. For example, this code appears as a weird mix of legacy and new R6:

if("page_name" %in% names(page)) page$drop()

i'd say something like this would be more clear:

if (page$name() == "page_name") page$drop()

# or

if (page$get_name() == "page_name") page$drop()

[ashrith]

@grigory93, it is totally fine if R6 library is not familiar. What we are looking to find out is if this is completely alien or if it is in familiar territory. The way to look at a R6 object is like accessing the elements in a list.
For example, here we have a simple list (no R6 used here)

> x <- list()
> y <- function(...){print(...)}
> x$a <- 10
> x$b <- y
> x
$a
[1] 10

$b
function(val){print(val+1)}

> 

An implementation in R6 would be

> library(R6)
> xx <- R6::R6Class("xx"
,public=
list(a=0
,printnum = function(){print(self$a)}))

Here notice that a similar list is created with the variable and function attached.
Now, having said that, in this code

if("page_name" %in% names(page)) page$drop()
is actually idiomatic R and not specific to R6. Here we are using an if() condition in one line.
As names(page) can give us a vector, using %in% we conduct a (vector) value match.
Hope this helps, please let me know if you have any questions.

[grigory93]

@ashrith what's next for R framework? would it be possible to test/ build whole apps with it soon/yet?

[ashrith]

Hi @grigory93 yes! Definitely. The first version that we are releasing is going to be dashboards. We are following it up with whole-build apps, that have similar interactivity as the python framework. Here is the link for whats planned #15