When the amount of data exceeds 2000, fcose layout consumes a lot of performance, accounting for 65% of the total rendering time

[duuliy]

My configuration is as follows "cytoscape-fcose": "^1.2.3" ，Is there a better optimization plan

name: 'fcose', fit: true, padding: 30, animate: false, nodeSeparation: 225, piTol: 0.0000001, nodeRepulsion: 18000, idealEdgeLength: 100, edgeElasticity: 0.45, nestingFactor: 0.1, tile: false, tilingPaddingVertical: 30, tilingPaddingHorizontal: 30,

[hasanbalci]

@duuliy Thanks for using fCoSE. fCoSE is designed for small to medium-sized graphs (up to a few thousand of nodes and edges), so depending on the graph topology closing to or exceeding these numbers may cause a decrease in the performance of the algorithm. Here are some ways we can suggest for a better performance:

• If your node dimensions are uniform or close to uniform, you can set uniformNodeDimensions option to true. This saves important amount of force-related calculations.
• If your graph consists of disconnected components, then you can activate packComponents option. This option enables fCoSE to lay out each component separately and then pack them in a pleasing way. Laying out smaller portions instead of the whole graph also increases the performance. Please note that this option requires cytoscape-layout-utilities extension to be registered in your app.
• We suggest using complexity management techniques if the graph is large. These techniques include:
  • Clustering the graph into compound nodes (if currently not) and then collapsing/expanding when necessary. Please see cytoscape-expand-collapse extension.
  • Hiding currently unnecessary details and gradually revealing on demand. Please see cytoscape-view-utilities extension.

I also suggest using the latest version of fCoSE, since it includes new features and miscellaneous bug fixes.

[duuliy]

@duuliy感谢您使用 fCoSE。fCoSE 是为中小型图（最多几千个节点和边）设计的，因此根据接近或超过这些数字的图拓扑，可能会导致算法性能下降。我们可以建议以下一些方法以获得更好的性能：

• 如果您的节点尺寸统一或接近统一，您可以将uniformNodeDimensions选项设置为true。这节省了大量与力相关的计算。
• 如果您的图形由断开连接的组件组成，那么您可以激活packComponents选项。此选项使 fCoSE 能够单独布置每个组件，然后以令人满意的方式将它们打包。布置较小的部分而不是整个图形也可以提高性能。请注意，此选项需要在您的应用程序中注册cytoscape-layout-utilities扩展。

• 如果图很大，我们建议使用复杂性管理技术。这些技术包括：

  • 将图形聚类为复合节点（如果当前没有），然后在必要时折叠/展开。请参阅cytoscape-expand-collapse扩展。
  • 隐藏当前不必要的细节并逐渐按需显示。请参阅cytoscape-view-utilities扩展。

我还建议使用最新版本的 fCoSE，因为它包含新功能和各种错误修复。

OK, thank you for your reply