Plot Data Pipeline

From ParaQ Wiki
Jump to navigationJump to search

Plot Data

Figure 1: Sample Histogram pipeline demonstrating tabular and bar chart views of the same data, with server serialization

This page captures ongoing discussion on how plot data should be delivered to the client for rendering. Some general requirements are as foll

  • In the extraction phase, plot data will be extracted from the pipeline using filters written for each type of plot data. Extraction filters will produce VTK data sets, with data set type depending on the type of data to be plotted:
    • Histogram Extraction Filter - vtkRectilinearGrid
    • Curve Probe - vtkPolyData
    • Scatter Plot Probe - vtkTable?
  • In the parallel combination phase, plot data collected in parallel is combined into a single data set for delivery to the client. The data set type remains the same.
  • Plot data is delivered to the client. The data set type remains the same.

Plot Representation

In the short-term, plots will be rendered on the client-side using Qt components including the current chart widgets. It is desirable that the chart widgets remain free of any VTK dependencies, so adapters will be required to convert plot data for display by the Qt chart widgets.

  • The client uses adapter classes to convert the VTK data set delivered from the server as-needed.
    • One adapter class could convert VTK data into a Qt-model that could be displayed in tabular form.
    • Another adapter could combine VTK data with chart-representation data (colors, fonts, etc) and either provide a Qt-model usable by the chart widgets, or call the chart widget API directly.

See Also

Plotting Parallel Data (PGraph)