mantaJob.launch: The interface from which compute Jobs are launched.
In joyent/mantaRSDK: Joyent Manta Storage Service R Software Development Kit

Description Usage Arguments Details Value See Also Examples

Submits R format Manta Job specification, runs Job, sends inputs if specified, closes inputs, polls Job status, returns Job status JSON.

1 2	mantaJob.launch(inputs, job, batchsize = 500, watch = TRUE, sleep = 15, watchtimeout = 600, silent = FALSE, verbose = FALSE)

`inputs`	vector of character optional. List of `inputs` as a vector of character, each containing valid paths to Manta objects as the intended job input files. You may use output from `mantaFind` or `mantaLs.paths` here. If you have no inputs, your initial Job task must be a `mantaReduce` step.
`job`	required. The R job structure as created with `mantaJob.setup` and Map and Reduce job tasks as defined therein by one or more `mantaMap` and/or `mantaReduce` steps. More information and parameters are explained in the help for these three functions.
`batchsize`	numeric. Maximum number of input object paths to upload in one batch to the running job. This function sends `inputs` in batches until they are all sent. Default is `500`.
`watch`	logical. When `TRUE` calls `mantaJob.done` in polling mode, after job is initiated, sleeping for for `sleep` seconds up to the duration of the `watchtimeout` value in seconds. This causes the function to wait until the job is done to return, or until timed out. Timeout does not imply job success or failure.
`sleep`	integer. Number of seconds to wait between status requests in polling mode when `watch` is `TRUE`. Default is 15 seconds.
`watchtimeout`	integer. Number of seconds after which polling ends. Passed to `mantaJob.done` when `watch` is set to `TRUE`. Default is 10 minutes (600 seconds).. If `watchtimeout` is exceeded, it means the job is still running or queued on Manta. `mantaJob.done(poll = TRUE)` or `mantaJob.status` can be called afterward for more monitoring.
`silent`	logical. Supress console messages, does not affect verbose setting.
`verbose`	logical optional. Passed to `RCURL` to reveal `HTTP` communication.

Job is created by mantaJob.setup and tasks as defined therein by mantaMap, and/or mantaReduce functions. Note that Manta tasks are UNIX shell commands, not native R commands.

Returns a Manta status JSON structure. The Manta Job identifier is the "id": field - like this "70c30bab-873b-66da-ebc8-ced12bd35ac4". This value is the jobid parameter to be used used by other mantaJob functions for inputs, status, errors and outputs as Job lookup key. This key can also be used by Node.js Manta command-line mjob commands.

Other mantaJobs: mantaJob.cancel; mantaJob.done; mantaJob.errors.stderr; mantaJob.errors; mantaJob.failures; mantaJob.inputs; mantaJob.outputs.cat; mantaJob.outputs; mantaJob.setup; mantaJob.status; mantaJobs.running; mantaJobs.tail; mantaJobs; mantaMap; mantaReduce

## Not run: 
## Example - Map/Reduce Unix Word Count

## Part 1.
## Job to download all of Shakespeare's plays to your account
plays <-
c("1kinghenryiv.txt", "1kinghenryvi.txt", "2kinghenryiv.txt",
"2kinghenryvi.txt", "3kinghenryvi.txt", "allswellthatendswell.txt",
"antonyandcleopatra.txt", "asyoulikeit.txt", "comedyoferrors.txt",
"coriolanus.txt", "cymbeline.txt", "hamlet.txt", "juliuscaesar.txt",
"kinghenryv.txt", "kinghenryviii.txt", "kingjohn.txt", "kinglear.txt",
"kingrichardii.txt", "kingrichardiii.txt", "loverscomplaint.txt",
"loveslabourslost.txt", "macbeth.txt", "measureforemeasure.txt",
"merchantofvenice.txt", "merrywivesofwindsor.txt",
"midsummersnightsdream.txt",
"muchadoaboutnothing.txt", "othello.txt", "periclesprinceoftyre.txt",
"rapeoflucrece.txt", "romeoandjuliet.txt", "sonnets.txt",
"tamingoftheshrew.txt",
"tempest.txt", "timonofathens.txt", "titusandronicus.txt",
"troilusandcressida.txt",
"twelfthnight.txt", "twogentlemenofverona.txt", "various.txt",
"venusandadonis.txt", "winterstale.txt")

file <- file("plays_list.txt", "wb")
# Important: This forces Windows to use /n instead of /r/n on write()
write(plays, file)
close(file)
rm(file)

mantaSetwd.stor()
mantaPut("plays_list.txt")

inputlist <- mantaLs.paths(grepfor = "plays_list.txt")

mantaMkdir("shakespeare")
mantaSetwd("shakespeare")
fileslocation <-
"https://us-east.manta.joyent.com/cwvhogue/public/shakespeare/"
destination <- mantaGetwd()
mantaJob.setup("Get Plays",
 mantaMap(paste("xargs -I {} sh -c 'curl -ksL ",
                 fileslocation,
                "{} | mput ",
                destination,
                "/{}'",
                sep=""))) -> moveplays

## Launch the first job to download the plays:
mantaJob.launch(inputlist, moveplays)

## See if they arrived.
mantaLs()
mantaLs.n()
mantaLs.du()

## Copy all the plays to your local drive?
# mantaGet(mantaFind())

## Speedread all of Shakespeare?
# mantaCat(mantaFind())

## Part 2.
## Map/Reduce Count all the words with wc and awk

inputs <- mantaFind()
job <- mantaJob.setup(
            name = "Word Count",
            mantaMap("wc"),
            mantaReduce("awk '{ l += $1; w += $2; c += $3 } END { print l, w, c }'")
        )

mantaJob.launch(inputs, job) -> status
## Getting Job Results:
## These functions find the last Job run if no jobid provided.
# mantaJob.status()  ## check to see if job is complete, as JSON information
# mantaJob.done()    ## returns logical job done (TRUE/FALSE)
# mantaJob.inputs()  ## returns list of inputs
mantaJob.outputs() ## retrieve list of paths to Manta output objects
# mantaJob.errors()  ## retrieve JSON formatted job error information
mantaJob.outputs.cat()   ## Print job output (text files) to console
# mantaJob.errors.stderr() ## Print any stderr messages to console

## End(Not run)