ParaView 3 Telecon 03/13/2007
Item | People | Description |
---|---|---|
1 | All | Status update |
2 | Geveci | Icons |
3 | Richardson | Volume Rendering Status |
4 | DeMarle | Selection |
5 | Moreland | Data Analysis functionality |
6 | Karelitz | Cell volumes (again) |
7 | Karelitz | Python proxy property documentation |
8 | Moreland | Resolved bugs |
Status update
Clint
Been fixing bugs assigned to him.
Mark
Same thing. Going over stuff from last months goal. Working on saving and loading of color map.
Dave DeMarle
Was working on clearing bugs last week. Still working on that. Updated selection Wiki page for plan to get PV 2 selection working.
Berk
No bug fixes, but he found several. Working on getting plotting over time.
Utkarsh
Changing the architecture so that the GUI does not have to manage undo/redo. Managed for you. GUI only has to manage "huge" changes such as creating views or sources.
Amy
Fixed LOD bug 2.6. Working on transfer function editor. Mostly working. Going to check in today. Also workong on PV 3 documentation.
Eric
Checked in lookmark changes. Has been fixing bugs. Fixed exodus reader so that it loads state correctly.
Icons
For cut, clip, threshold, asked elevation to do something closer to our orignal icons. For glyph, suggested adding more little circles. For group extract, remove shading.
Eric mentioned the need for the lookmark icon. Berk is going to ask elevation to do that. We also need to ask them to do a splash screen.
Volume rendering status
The transfer function editor should be ready to go. Mark turned on volume rendering this morning and it crashed. So far there is no separate panel for volume rendering, which there needs to be. As a first pass, we should duplicate the PV 2 panel. Berk and Utkarsh will investigate why volume rendering is crashing.
Selection
Dave D. suggests we scale down our expectations to just get back the PV 2 functionality. So, we are getting rid of the frustum selection and threshold selection. The problem with both is that you can easily grab all of the data and send it to the client.
Want to disallow select multiple data sets at the same time. After discussion, as first pass, if multiple objects are selected, PV should make some determination of which is more pertanent, and just select that one.
Making selections usable. Start by making a selection source class. Then make a set of filters that will operate on that source. When you add one of these types of filters, both are added to the pipeline.
Also need to make the selection by id work like the pick and probe in PV 2. Also need to be able to select points instead of cells.
Dave is also working on a bunch of cells. Berk is working on extract over time.
Data analysis functionality
Instead of having one big data analyis filter, we were going to break it into smaller components. Berk is planning to implement plotting over time, he will have an extract over time filter. If you apply to probe or extract, it will extract those things. To plot a few points over time, you will apply the extact filter, and then apply the plot filter.
You can also "probe." We need to clean up the name. Dave K. brought up the request to enable the plot over line to work on line and cell data. Dave K. will make a bug request.
Cell Volumes
There is supposed to be a helper class that allows a python script to find the volume of each cell. For the short term, just hexes should work, but it should not crash if you give it a tet. Of course, it all has to be wrapped in Python.
Python proxy property documentation
We (Tim Shead and co) are working with customers on python pipelines. They are at a point where they are trying to get their customers up and running, but the documentation is lousy. They have to go to multiple locations for documentations, and have to mentally convert from C++ docs to Python. Would like a set of Python specific documentation.
The typical use case is to write a script that sets up a pipeline that has a programable filter. They are working at the API at every level.
Tim is already able to pull on-line documentation on proxies. That works pretty well, except that there are several holes in the documentation in the XML. Amy is working on that documentation, and that should be added to the server manager part of the documentation request.
Tim would like to use doxygen to create an XML description of the documents (which you can) and then post-process the XML description to strip out things that are not available in the Python wrappings. Berk suggests reading this XML description into Python and use Python's introspection to find class and method names and use the XML to document them.
Tim will do some experimentation.
Multiple inputs with programmable filter
Tim has a use case for the programmable filter that has multiple inputs. There are two issues. The first is having the GUI set multiple inputs. Berk suggets multiple filter entries for diffent numbers of inputs.
The second issue is getting the multiple inputs in the filter. Berk will look into this.
Dave K. will enter a bug.