Deploy a jspsych experiment locally
run_locally( path, experiment_folder = "experiment", data_folder = "data", port = 8000 )
| path | Path where the experiment is deployed |
|---|---|
| experiment_folder | Experiment subfolder |
| data_folder | Data subfolder |
| port | The port to use |
The purpose of the run_locally() function is to start an
R server running (using the plumber package) that will
serve the experiment from the local machine. The path,
experiment_folder and data_folder arguments specify the location
specify where the experiment should be deployed, and should be the same
that was used when calling build_experiment()) to build the
experiment originally. The port is the numeric value of the port on
which the experiment is served.
Once run_locally() has been called, a browser window should open
showing the relevant page.
There are two reasons to deploy a local experiment using run_locally()
rather than simply opening the relevant "index.html" file in the browser. The
first is for the purpose of saving data. For security reasons, browsers do not
generally permit client-side javascript (e.g., the code that runs the jsPsych
experiment) to save files to arbitrary locations. For this reason writing the
data to file is the job of the R server, not the javascript code that
is running through the browser. In other words, if the experiment is deployed
locally using the run_locally() function, then the save_locally()
function that used to record data locally will work properly. If, however, the
"index.html" file is opened without starting the R server, data will not be
saved to file.
The second reason for using run_locally() is that it opens up the possibility
that an experiment could use server-side R code at runtime. At the moment jaysire
does not have any functionality to do so, but in principle there is nothing
preventing the R server from playing a more active role when the experiment is
running, and future versions of the package may develop this functionality
further.