Concerning undelying graph data structure / model API : `ngraph` vs `graphology`
cliffordfajardo opened this issue · comments
ngraph vs graphology
Not urgent, just wanted to share graphology and get your thoughts on it compared to ngraph for reagraph's underlying graph data structure 😄
grapology highlights
Compared to ngraph,graphology supports a more rich out of the box API for the graph:
- comprehensive standard library of algorithms, graph generators, layouts traversals , event listeners & utility functions
- layout algorithms offer sync and async versions (using web workers)
The library is used by other more recent graph libraries like sigmajs and pixi-graph for the graph backend.
Besides the rich API, another strong aspect for the library is that its a community driven effort to help graph libraries share layouts, algorithms etc. In practice this means someone using sigmajs for example could move to pixi-graphor even reagraph (if it used graphology) since they share the same underlying graph structure
Jan Zak, talks about graphology in this YouTube video here
Centralized documentation
- Ngraph's packages are all scattered across different individual repos. Whereas's graphology has everything in one mono-repo
- For those wanting to customize reagraph, the centralized documentation with
graphologyis a big DX & productivity win
Thanks for sharing this - never seen graphology before. I'm open to exploring it. Would you be interested in taking a first stab at it?
Thanks for sharing this - never seen graphology before. I'm open to exploring it. Would you be interested in taking a first stab at it?
I'm interested - busy at work & life transitions, but can't really give an ETA 😅
Created the issue mostly so I don't forget to revisit / get feedback for now
@cliffordfajardo I dug into this and I'm on board. Still open to helping w/ it ?
Haven't forgot about the github ^^ comment
Been super busy these days but its definetely something I want to explore! I'll get back to you later in the week
I don't think I'll get around this as I was originally planning to opportunistically do this work concurrently with tasks I had at work, but the dataviz project I was working on got scrap'd
just wanted to provide an update in case anyone was wondering or passed this thread in the future
@cliffordfajardo - this is done!