Paraview催化剂用户指南：实用教程与开源授权

需积分: 0 55 浏览量更新于2024-06-28 收藏 4.25MB PDF 举报

Paraview是一款强大的、开源的数据可视化和并行处理软件，它广泛应用于科学计算、工程模拟以及医学成像等领域。ParaView Catalyst是ParaView的一个扩展模块，专为高性能计算环境设计，旨在简化大规模数据集的处理和可视化流程。这份用户指南，由Andrew Bauer、Berk Geveci和Will Schroeder编写，于2019年9月12日发布，遵循Creative Commons Attribution 3.0许可协议。 Catalyst的核心价值在于其能够将并行计算和图形渲染任务分离，允许用户在后端进行复杂的数值模拟，同时在前端快速交互式地查看和分析结果。它支持多种数据源，包括HDF5、VTK等格式，并且与许多高级计算代码（如Sandia National Laboratory的项目）无缝集成，如通过LANL团队的工作，将Catalyst与自家的模拟代码相结合，提升了数据处理效率。在本书中，你会发现关于以下几个关键知识点的详细介绍： 1. **安装和配置**：指南会指导用户如何在各自的系统上安装ParaView Catalyst，包括硬件和软件需求，以及配置步骤。 2. **数据导入与预处理**：涵盖如何导入各种类型的数据到Catalyst环境中，以及如何利用Paraview的预处理功能清洗、转换和组织数据。 3. **工作流程管理**：介绍如何创建、管理和优化Catalyst工作流程，包括数据流图的构建，以及如何利用 ParaView的pipeline系统处理大量数据。 4. **可视化技术**：深入解析Catalyst中的高级可视化技巧，包括体积渲染、切片、剖面图和动画制作等，帮助用户更有效地展示复杂数据。 5. **并行和分布式计算**：讲解如何利用ParaView Catalyst的优势，在多核处理器、集群或GPU环境中实现高效的数据处理和渲染。 6. **案例研究和最佳实践**：书中提供了实际应用案例，展示了Catalyst在不同领域的应用效果，以及如何根据具体场景调整策略。 7. **社区和支持**：指南还强调了ParaView社区的重要性，包括获取帮助、贡献代码和分享经验的途径。 8. **版权和致谢**：最后部分列举了支持ParaView Catalyst项目的重要人物和机构，对他们的贡献表示感谢。这份用户指南为Paraview Catalyst的使用者提供了一个全面的参考手册，无论你是数据科学家、工程师还是研究人员，都可以从中找到适合自己的工具和技术，提升数据处理和可视化的效率和质量。

ParaView Catalyst User’s Guide

Figure 2.2: The Catalyst panel can be toggled in

View Catatalyst Export Inspector

2.2.1 Creating a Representative Dataset

A question that often arises is how to create a representative dataset. There are two ways to do this. The ﬁrst way is to

run the simulation with Catalyst with a script that outputs the full grid with all attribute information. Appendix 6.1.2 has a

script that can be used for this purpose. The second way is by using the sources and ﬁlters in ParaView. The easiest grids

to create within the GUI are image data grids (i.e. uniform rectilinear grids), polydata and unstructured grids. For those

knowledgeable enough about VTK, the programmable source can also be used to create all grid types. If a multi-block

grid is needed, the Group Datasets ﬁlter can be used to group together multiple datasets into a single output. The next step

is to create the attribute information (i.e. point and/or cell data). This can be easily done with the Calculator ﬁlter as it can

create data with one or three components, name the array to match the name of the array provided by the adaptor, and set

an appropriate range of values for the data. Once this is done, the user should save this out and then read the ﬁle back in to

have the reader act as the source for the pipeline.

2.2.2 Manipulating Python Scripts

For users that are comfortable programming in Python, we encourage them to modify the given scripts as desired. The

following information can be helpful for doing this:

• Sphinx generated ParaView Python API documentation at

www.paraview.org/ParaView/Doc/Nightly/www/py-doc/index.html.

• Using the ParaView GUI trace functionality to determine how to create desired ﬁlters and set their parameters. This

is done with

Start Trace

and

Stop Trace

under the

Tools

menu.

• Using the ParaView GUI Python shell with tab completion. This is done with

Python Shell

under the

View

menu.

ParaView Catalyst User’s Guide

Figure 2.4: ParaView GUI pipeline for live connection.

simulation run state to signify the status of the simulation. The icons for this are:

• indicates the simulation is running with no breakpoint set.

• indicates that the simulation has reached a breakpoint.

• indicates that a breakpoint has been set but not yet reached.

A demonstration of this functionality is at www.kitware.com/blog/home/post/722.

2.4 Cinema

Cinema is an image-based approach to in situ analysis and visualization. The concept is that by saving an organized set

of images into a Cinema database, an analyst can perform post hoc analysis and visualization directly from the generated

images. ParaView Catalyst can be used to create a Cinema database and the speciﬁcation of the Cinema output can be done

through the ParaView GUI’s Catalyst Script Generator. Figure 2.5 shows the expanded options when Output to Cinema is

enabled. These options are:

• Export Type – This option speciﬁes how the view’s camera should be manipulated when generating the images.

Current options include None, Static and Spherical. None indicates that no Cinema output is requested for this view

and Static indicates that the camera need not be moved. Spherical will rotate the camera around the view’s center at

the given Phi and Theta angles.

• Cinema Track Selection – This option allows changing ﬁlter parameters and what ﬁeld data to pseudo-color with.

By selecting the pipeline object in the left pane, users can specify the arrays to pseudo-color by in the right pane’s

Arrays tab or the ﬁlter’s parameters in the right pane’s Filter Values tab. Note that currently only Slice, Contour and

Cut ﬁlters have been enabled for modifying ﬁlter values.

See www.cinemascience.org or Ahrens, et. al.

[1]

for detailed information on Cinema.

2.5 Avoiding Data Explosion

A key point to keep in mind when creating Catalyst pipelines is that the choice and order of ﬁlters can make a dramatic

difference in the performance of Catalyst (this is true with ParaView as well). Often, the source of performance degradation

is when dealing with very large amounts of data. For memory-limited machines like today’s supercomputers, poor decisions

when creating a pipeline can cause the executable to crash due to insufﬁcient memory. The worst case scenario is creating

ParaView Catalyst User’s Guide

an unstructured grid from a topologically regular grid. This is because the ﬁlter will change from using a compact grid data

structure to a more general grid data structure.

We classify the ﬁlters into several categories, ordered from most memory efﬁcient to least memory efﬁcient and list some

commonly used ﬁlters for each category:

1. Total shallow copy or output independent of input – negligible memory used in creating a ﬁlter’s output. The ﬁlters

in this category are:

• Annotate Time

• Append Attributes

• Extract Block

• Extract Datasets

• Extract Level

• Glyph

• Group Datasets

• Histogram

• Integrate Variables

• Normal Glyphs

• Outline

• Outline Corners

• Pass Arrays

• Plot Over Line

• Probe Location

2. Add ﬁeld data – the same grid is used but an extra variable is stored. The ﬁlters in this category are:

• Block Scalars

• Calculator

• Cell Data to Point Data

• Compute Derivatives

• Curvature

• Elevation

• Generate Ids

• Generate Surface Normals

• Gradient

• Gradient of Unstructured DataSet

• Level Scalars

• Median

• Mesh Quality

• Octree Depth Limit

• Octree Depth Scalars

• Point Data to Cell Data

• Process Id Scalars

• Random Vectors

• Resample with Dataset

• Surface Flow

• Surface Vectors

• Transform

• Warp (scalar)

• Warp (vector)

3. Topology changing, dimension reduction – the output is a polygonal dataset but the output cells are one or more

dimensions less than the input cell dimensions. The ﬁlters in this category are:

• Cell Centers

• Contour

• Extract CTH Fragments

• Extract CTH Parts

• Extract Surface

• Feature Edges

• Mask Points

• Outline (curvilinear)

• Slice

• Stream Tracer

4. Topology changing, moderate reduction – reduces the total number of cells in the dataset but outputs in either a

polygonal or unstructured grid format. The ﬁlters in this category are:

• Clip

• Decimate

• Extract Cells by Region

• Extract Selection

• Quadric Clustering

• Threshold

5. Topology changing, no reduction – does not reduce the number of cells in the dataset while changing the topology

of the dataset and outputs in either a polygonal or unstructured grid format. The ﬁlters in this category are:

剩余76页未读，继续阅读

sunlightwyww

粉丝: 0

Paraview催化剂用户指南：实用教程与开源授权

paraview英文说明书

paraview3.14说明书

paraview手册

ParaView1.4中文说明书

CAST-DESIGNER 7.5用户手册：ParaView使用指南

paraview中文说明书

Matlab 3D 阵列到 ParaView（通过 HDF5 + XDMF）：我们举例说明如何使用 HPC 的经典方法将大型 3D 阵列导出到 ParaView 进行可视化：使用 HDF5 和 XDMF 文件。-matlab开发

paraview4.3.1安装包for windows

ParaView中文使用手册：从入门到进阶

深入解析ParaView-Glance源码

最新资源