Like the DataLoader of das2, but provides minimal dataset discovery metadata
loading the data is slow, so load the data asynchronously (on a separate thread). This should return true if getDataSet will take more than 100 milliseconds (interactive time). Note this is currently ignored, and may be indefinitely ignored.
cookie jar of capabilities, see org.autoplot.datasource.capability. Each capability can be queried, and either an object implementing the capability or null is returned. Example capabilities include:
| TimeSeriesBrowse | which allows the user to request data from a different interval. |
| Caching | which should be queried before the loader is called again. |
retrieve the dataset. This allowed to be sub-interactive or batch time scale, and will block until the dataset is produced. This may return null when no data is available, or NoDataInIntervalException can be thrown. If the user cancelled the operation, then java.io.InterrupedIOExcaption or better yet org.das2.CancelledOperationException should be called. These will simply display "cancelled" (or similar) on the status bar.
Return arbitrary metadata for the dataset. This is a map of String to Objects, and to form a tree structure, property name may map to another Map<String,Object>. Note the order of the properties may be controlled by using LinkedHashMap for the implementation. Even though this takes a monitor, it will be called after getDataSet, and the monitor may be safely ignored. This should return new HashMap() if no metadata is found.
return a MetadataModel that scrapes the Metadata tree returned to provide a set of properties identified in QDataSet.
discovery properties for the dataset. These should follow the QDataSet conventions, such as TITLE, LABEL, etc, and should mirror the structure of the dataset. Note getMetadataModel().getProperties( getMetaData() ) should return the same thing.
return the fully-qualified URI of this data source, including the "vap+