Great package! I tried to play around and use the deeplearning/ColorfulImageColorization/1.1.13 algorithm to colorizeblack and white images.
I uploaded the images in a new collection that I created on Algorithmia. I could access the collection directory with nlp_directory <- client$dir(path) and I could ask for a file by name in that directory with nlp_directory$file("test.jpg"). Or at least that is what I assume after the exists() method returned TRUE.
However, input_files <- nlp_directory$files() and then input_files$directoryEntries returned an empty list.
Maybe I am just missing something, I tried to find any explanation in the docs but with no luck.
p.s. I could get the filenames from my local dir at the end and I could use those in the algo call and it worked like a charm.
Thank you for your help!
Here is my full code:
install.packages("algorithmia")
library(algorithmia)
# Authenticate with your API key
apiKey <- "MY_API"
# Create the Algorithmia Client object
client <- getAlgorithmiaClient(apiKey)
# Get (and create) a directory collection
path <- "data://marton_balazs_kovacs/test"
nlp_directory <- client$dir(path)
nlp_directory$create()
# Update directory permissions (e.g. make it private)
nlp_directory$updatePermissions(ReadAcl.PRIVATE)
nlp_directory$getPermissions()$read_acl
# Get one file
## If you refer to the file with the exact name the function works
input <- nlp_directory$file("test.jpg")
## Test if the file that we asked for exists
input$exists()
# Get childs of the directory collection
## This function supposed to get all the files in my directory but it does not
input_files <- nlp_directory$files()
## This argument should get the names of the files in a list but it does not work
input_files <- input_files$directoryEntries
# Get the names of the files from a local repository
## Path to the local directory with the images
local_path <- file.path("MY_PATH")
## Get the filenames in that repository
local_filenames <- list.files(local_path)
## Append filenames to the dirpath on algorithmia and store it in a list obj
input_files <- list(image = paste(path, local_filenames, sep = "/"))
# Run the algorithm
algo <- client$algo("deeplearning/ColorfulImageColorization/1.1.13")
algo$setOptions(timeout=300) # optional
result <- algo$pipe(input_files)$result
print(result)
DataDirectory's files method returns an iterator not a list. To access each file object, you need to call hasNext and getNext in some kind of loop. Here is an example where we do that in tests
Hi!
Great package! I tried to play around and use the
deeplearning/ColorfulImageColorization/1.1.13
algorithm to colorizeblack and white images.I uploaded the images in a new collection that I created on Algorithmia. I could access the collection directory with
nlp_directory <- client$dir(path)
and I could ask for a file by name in that directory withnlp_directory$file("test.jpg")
. Or at least that is what I assume after theexists()
method returned TRUE.However,
input_files <- nlp_directory$files()
and theninput_files$directoryEntries
returned an empty list.Maybe I am just missing something, I tried to find any explanation in the docs but with no luck.
p.s. I could get the filenames from my local dir at the end and I could use those in the algo call and it worked like a charm.
Thank you for your help!
Here is my full code: