fuse_roots(graph: dask.highlevelgraph.HighLevelGraph, keys: list)
Often Blockwise sections of the graph fill out all of the computation except for the initial data access or data loading layers:
Large Blockwise Layer | | | X Y Z
This can be troublesome because X, Y, and Z tasks may be executed on different machines, and then require communication to move around.
This optimization identifies this situation, lowers all of the graphs to concrete dicts, and then calls fuse
on them, with a width equal to the number of layers like X, Y, and Z.
This is currently used within array and dataframe optimizations.
The full graph of the computation
The output keys of the computation, to be passed on to fuse
Fuse nearby layers if they don't have dependencies
Hover to see nodes names; edges to Self not shown, Caped at 50 nodes.
Using a canvas is more power efficient and can get hundred of nodes ; but does not allow hyperlinks; , arrows or text (beyond on hover)
SVG is more flexible but power hungry; and does not scale well to 50 + nodes.
All aboves nodes referred to, (or are referred from) current nodes; Edges from Self to other have been omitted (or all nodes would be connected to the central node "self" which is not useful). Nodes are colored by the library they belong to, and scaled with the number of references pointing them