The DataSHIELD Interface (DSI) defines a set of S4 classes and generic methods that can be implemented for accessing a data repository supporting the DataSHIELD infrastructure: controlled R commands to be executed on the server side are garanteeing that non disclosive information is returned to client side.
Learn more about DataSHIELD.
The DSI classes are:
DSObjecta common base class for all DSI,
DSDriverdrives the creation of a connection object,
DSConnectionallows the interaction with the remote server; DataSHIELD operations such as aggregation and assignment return a result object; DataSHIELD setup status check can be performed (dataset access, configuration comparision),
DSResultwraps access to the result, which can be fetched either synchronously or asynchronously depending on the capabilities of the data repository server.
All classes are virtual: they cannot be instantiated directly and instead must be subclassed. See DSOpal for a reference implementation of DSI based on the Opal data warehouse. See also DSLite for a server-less implementation of DSI for local datasets.
These S4 classes and generic methods are meant to be used for implementing connection to a DataSHIELD-aware data repository.
In addition to these S4 classes, DSI provides functions to handle a list of remote data repository servers:
datashield.logoutwill make use of the
DSDriverparadigm to create
DSConnections to the data repositories,
datashield.assignwill perform typical DataSHIELD operations on
DSConnections, which result will be fetched through
datashield.connections_findare functions for managing the list of
DSConnectionobjects that will be discovered and used by the client-side analytic functions.
datashield.errorswill report the last R errors that may have occurred after a
datashield.workspace_rmallow to manage R images of the remote DataSHIELD sessions (to speed up data analysis sessions),
datashield.symbol_rmoffer a minimalistic management of the R symbols living in the remote DataSHIELD sessions,
datashield.table_statuslist the tables and their accessibility across a set of data repositories,
datashield.resource_statuslist the resources and their accessibility across a set of data repositories,
datashield.methodsare utility functions to explore the DataSHIELD setup across a set of data repositories.
datashield.profileslists the DataSHIELD profiles that can be selected at login time.
datashield.* functions are meant to be used by DataSHIELD packages developers and users.
Some options can be set to modify the behavior of the DSI:
datashield.envis the R environment in which the
DSConnectionobject list is to be looking for. Default value is the Global Environment:
datashield.progressis a logical to enable the visibility of the progress bars. Default value is
datashield.progress.clearis a logical to make the progress bar disappear after it has been completed. Default value is
datashield.errors.stopis a logical to alter error handling behavior: if
TRUEan error is raised when at least one server has failed, otherwise a warning message is issued. Default value is
datashield.polling.sleep.0time in seconds to wait before checking async calls completion, before ~1 second. Default value 50 milliseconds.
datashield.polling.sleep.1base time in seconds to wait before checking async calls completion, after ~1 second. Default value is 1 second.
datashield.polling.sleep.10time in seconds to wait before checking async calls completion, after ~10 seconds. Default value is x2 the base time (2 seconds).
datashield.polling.sleep.60time in seconds to wait before checking async calls completion, after ~1 minute. Default value is x10 the base time (10 seconds).
datashield.polling.sleep.600time in seconds to wait before checking async calls completion, after ~10 minutes. Default value is x60 the base time (1 minute).
datashield.polling.sleep.3600time in seconds to wait before checking async calls completion, after ~1 hour. Default value is x600 the base time (10 minutes).