How to install reco R package ?

[saraswatmks]

Hello

I am unable to install this package. Also, I cannot find any documentation about how to install.
I am getting the following error:

install.packages("reco")
Installing package into ‘/home/manish/R/x86_64-pc-linux-gnu-library/3.4’
(as ‘lib’ is unspecified)
Warning in install.packages :
package ‘reco’ is not available (for R version 3.4.1)

Please advise.

Best,
Manish

[dselivanov]

Hi. It is not on cran yet. You can use devtools::install_github("dselivanov/reco")

[saraswatmks]

This works. Thanks.

[mayankgoel4]

Hi..I tried installing the package using devtools::install_github("dselivanov/reco") and then installed rtools.
Finally used library(reco) to load the package, However still model=ALS$new(rank = 8) is not working and even ?ALS is saying no documentation for function ALS. Any thoughts?

[dselivanov]

duplicate #14 #11 - see WRMF class

[mayankgoel4]

thanks. WRMF function works instead of ALS..

[PalinMemoi]

devtools::install_github("dselivanov/reco")
Downloading GitHub repo dselivanov/reco@master
from URL https://api.github.com/repos/dselivanov/reco/zipball/master
trying URL 'https://cran.rstudio.com/bin/windows/Rtools/Rtools34.exe'
Content type 'application/x-msdownload' length 108085090 bytes (103.1 MB)
downloaded 103.1 MB

WARNING: Rtools is required to build R packages, but is not currently installed.

Please download and install Rtools 3.4 from http://cran.r-project.org/bin/windows/Rtools/.
Installation failed: Could not find build tools necessary to build rsparse

library(reco)
Error in library(reco) : there is no package called ‘reco’

[dselivanov]

@PalinMemoi please install Rtools from CRAN as suggested in error message

[PalinMemoi]

Plz, help me...I installed Rbuildtools. however, it doesn't work. I tried to install reco again by devtools.... It shows message as below
Downloading GitHub repo dselivanov/reco@master
from URL https://api.github.com/repos/dselivanov/reco/zipball/master
Installing rsparse
Installing 1 package: Rcpp
Installing package into ‘C:/Users/anik/Documents/R/win-library/3.4’
(as ‘lib’ is unspecified)
trying URL 'https://cran.rstudio.com/bin/windows/contrib/3.4/Rcpp_0.12.19.zip'
Content type 'application/zip' length 4433824 bytes (4.2 MB)
downloaded 4.2 MB

package ‘Rcpp’ successfully unpacked and MD5 sums checked
Warning: cannot remove prior installation of package ‘Rcpp’

The downloaded binary packages are in
C:\Users\anik\AppData\Local\Temp\RtmpEVybHU\downloaded_packages
"C:/PROGRA1/R/R-341.4/bin/x64/R" --no-site-file --no-environ --no-save --no-restore --quiet CMD INSTALL
"C:/Users/anik/AppData/Local/Temp/RtmpEVybHU/devtools9a82643348d/dselivanov-rsparse-b465c0d"
--library="C:/Users/anik/Documents/R/win-library/3.4" --install-tests

ERROR: dependency 'Rcpp' is not available for package 'rsparse'

• removing 'C:/Users/anik/Documents/R/win-library/3.4/rsparse'
  In R CMD INSTALL
  Installation failed: Command failed (1)

[zakariaBensmida]

Hi Dmitry,
i used a 3.3.3 r version with 3.4 rtolls versions and i runed this following code:

install.packages("devtools",type="win.binary")
library(devtools)
devtools::install_github("dselivanov/reco")

(without "win.binary" I was not able to install devtools that"s why i have used it )
so after running i got rsparse installed but reco is not . i have repeated the same process but with 3.5 version for both R and rtools ,it did not worked at all (nothing installed)(ERROR: compilation failed for package 'rsparse)..Since a month and I m trying to solve this problem but I could not install recp package till now.

Could you please help me.

[dselivanov]

@zakariaBensmida it is on CRAN now and you can use just install.packages('rsparse')

[zakariaBensmida]

Hi Dmitry
and then? how to install reco?

[dselivanov]

There is no more reco. It was renamed/refactored to rsparse. чт, 25 апр. 2019 г., 11:55 zakariaBensmida <notifications@github.com>:

…

Hi Dmitry and then? how to install reco? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#10 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABHC5XN5X7LQAHRH365GIHDPSFPVZANCNFSM4EFF3FZQ> .

[zakariaBensmida]

Oh my god !!! so you mean i must change some code in your tutorial also? i mean the artist/music recommender system example?

[dselivanov]

Yes, but I think it will be not too many changes. чт, 25 апр. 2019 г., 12:00 zakariaBensmida <notifications@github.com>:

…

Oh my god !!! so you mean i must change some code in your tutorial also? i mean the artist/music recommender system example? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#10 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABHC5XNKST433MSERPAX223PSFQIPANCNFSM4EFF3FZQ> .

[zakariaBensmida]

Thank youuuuuuuuuuu veryyyyyyy muchhhhhhhhhhhh

[zakariaBensmida]

Hi Dmitry,
when i run this:
model$add_scorer(x = X_cv_history_conf, y = X_cv_future, name = "ndcg-10", metric = "ndcg@10")

model$add_scorer(x = X_cv_history_conf, y = X_cv_future, name = "map-10", metric = "map@10")

i got this error :
Error: attempt to apply non-function

[dselivanov]

This was removed from pkg API as I found it redundant. Please consult with the pkg documentation - `?WRMF`. You can examine model performance after model fitting - see `?metrics`.

…

On Thu, Apr 25, 2019 at 12:36 PM zakariaBensmida ***@***.***> wrote: Hi Dmitry, when i run this: model$add_scorer(x = X_cv_history_conf, y = X_cv_future, name = "ndcg-10", metric = ***@***.***") model$add_scorer(x = X_cv_history_conf, y = X_cv_future, name = "map-10", metric = ***@***.***") i got this error : Error: attempt to apply non-function — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#10 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABHC5XI2GBW4ULNBBWBX5HTPSFUPNANCNFSM4EFF3FZQ> .

-- Regards Dmitriy Selivanov

[zakariaBensmida]

Hi dmitry,

when i use the below code i cant find any predictiosn so i can use this code"ap_k(predictions, actual)"
set.seed(1)
library(data.table)
raw_data<-fread("J:/MK-OM-AC/Projekte/Ranking/Daten/usersha1-artmbid-artname-plays.tsv",showProgress=T)

setnames(raw_data, c("user_id", "artist_id", "artist_name", "number_plays"))
head(raw_data)
user_encoding = raw_data[, .(uid = .GRP), keyby = user_id]
print(user_encoding)
item_encoding = raw_data[, .(iid = .GRP, artist_name = artist_name[[1]]), keyby = artist_id]
print(item_encoding)
dt = user_encoding[raw_data, .(artist_id, uid, number_plays), on = .(user_id = user_id)]
dt = item_encoding[dt, .(iid, uid, number_plays), on = .(artist_id = artist_id)]
rm(raw_data)
dt
library(Matrix)
X = sparseMatrix(i = dt$uid, j = dt$iid, x = dt$number_plays,
dimnames = list(user_encoding$user_id, item_encoding$artist_name))
N_CV = 1000L
cv_uid = sample(nrow(user_encoding), N_CV)
X_train = X[-cv_uid, ]
dim(X_train)
X_cv = X[cv_uid, ]
dim(X_cv)
rm(X)
temp = as(X_cv, "TsparseMatrix")
temp = data.table(i = temp@i, j = temp@j, x = temp@x)
temp[, history := sample(c(TRUE, FALSE), .N, replace = TRUE, prob = c(0.5, 0.5)), by = i]
head(temp)
X_cv_history = temp[history == TRUE]
X_cv_future = temp[history == FALSE]
rm(temp)
X_cv_history = sparseMatrix( i = X_cv_history$i, j = X_cv_history$j, x = X_cv_history$x,
dims = dim(X_cv), dimnames = dimnames(X_cv), index1 = FALSE)
X_cv_future = sparseMatrix( i = X_cv_future$i, j = X_cv_future$j, x = X_cv_future$x,
dims = dim(X_cv), dimnames = dimnames(X_cv), index1 = FALSE)

install.packages('rsparse')
library(rsparse)
model = WRMF$new(rank = 8)
make_confidence = function(x, alpha) {
x_confidence = x
stopifnot(inherits(x, "sparseMatrix"))
x_confidence@x = 1 + alpha * x@x
x_confidence
}
alpha = 0.1
X_train_conf = make_confidence(X_train, alpha)
X_cv_history_conf = make_confidence(X_cv_history, alpha)

RhpcBLASctl::blas_set_num_threads(1)
user_embeddings = model$fit_transform(X_train_conf, n_iter = 10L, n_threads = 8)

library(ggplot2)
trace = attr(user_embeddings, "trace")
g = ggplot(trace) + geom_line(aes(x = iter, y = value, col = scorer))
plotly::ggplotly(g, width = 9, height = NULL)

futile.logger::flog.threshold(futile.logger::ERROR)

RhpcBLASctl::blas_set_num_threads(1)

trace = NULL
alpha = c(0.01, 1)
rank = c(16, 8)
n_iter_max = 10L
n_threads = 8L
grid = expand.grid(alpha = alpha, rank = rank)
convergence_tol = 0.01
for(k in seq_len(nrow(grid))) {
alpha = grid$alpha[[k]]
rank = grid$rank[[k]]

futile.logger::flog.info("alpha = %.3f, rank = %d", alpha, rank)

model = WRMF$new(rank = rank)

X_train_conf = make_confidence(X_train, alpha)
X_cv_history_conf = make_confidence(X_cv_history, alpha)

fit model

user_embeddings = model$fit_transform(X_train_conf, n_iter = n_iter_max,
convergence_tol = convergence_tol, n_threads = n_threads)

store strace

grid_trace = attr(user_embeddings, "trace")

grid_trace$param_set = sprintf("alpha=%.3f; rank=%d", alpha, rank)
trace = c(trace, list(grid_trace))
}
trace = rbindlist(trace)
g = ggplot(trace) +
geom_line(aes(x = iter, y = value, col = param_set)) +
facet_wrap( ~ scorer, scales = "free") +
theme(legend.position="bottom")

plotly::ggplotly(g, width = 9, height = NULL)

[zakariaBensmida]

i tried this one but it does not work
actual = as(X_train_conf, "RsparseMatrix")
identical(rsparse::ap_k(user_embeddings[,7], actual), 1)
Error in rsparse::ap_k(user_embeddings[, 7], actual) :
is.matrix(predictions) is not TRUE

[dselivanov]

Try user_embeddings[, 7, drop = F] чт, 25 апр. 2019 г., 13:33 zakariaBensmida <notifications@github.com>:

…

i tried this one but it does not work actual = as(X_train_conf, "RsparseMatrix") identical(rsparse::ap_k(user_embeddings[,7], actual), 1) Error in rsparse::ap_k(user_embeddings[, 7], actual) : is.matrix(predictions) is not TRUE — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#10 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ABHC5XPASXNR2UXA2POXPLTPSF3HJANCNFSM4EFF3FZQ> .

[zakariaBensmida]

sorry but i have soome idiot questions .is the below the predictions or only the seventh column is the predictions?
head(user_embeddings)
[,1] [,2] [,3] [,4] [,5] [,6]
00000c289a1829a808ac09c00daf10bc3c4e223b -4.4366019 -6.4496755 7.759140 4.0778027 3.021041 -2.5092545
00001411dc427966b17297bf4d69e7e193135d89 -3.8179604 -6.3121820 3.802888 -9.0638672 3.691029 1.5814032
00004d2ac9316e22dc007ab2243d6fcb239e707d -4.0583470 4.2391097 1.512497 -1.2811394 2.197221 -0.3218515
000063d3fe1cf2ba248b9e3c3f0334845a27a6bf -0.3175396 -0.2926188 2.726539 0.4600829 1.136645 -1.6827423
00007a47085b9aab8af55f52ec8846ac479ac4fe 0.8436100 2.6911790 2.080224 -3.6851317 2.073310 1.0996730
0000c176103e538d5c9828e695fed4f7ae42dd01 1.0942813 2.5679246 4.667832 -4.2642418 1.951552 2.3230622
[,7] [,8]
00000c289a1829a808ac09c00daf10bc3c4e223b 5.1683390 1.4608108
00001411dc427966b17297bf4d69e7e193135d89 5.2887858 -3.4221277
00004d2ac9316e22dc007ab2243d6fcb239e707d 3.8647705 -2.4137396
000063d3fe1cf2ba248b9e3c3f0334845a27a6bf 0.3753359 0.5390940
00007a47085b9aab8af55f52ec8846ac479ac4fe 5.4227957 -0.9695473
0000c176103e538d5c9828e695fed4f7ae42dd01 5.4995095 0.5390066