lmeyerov

I recall 10M graphs being slow with other methods, if you're finding fast, then we can pass. Just given we generally support cudf + cugraph out-of-the-box, and they have hierarchical...

Typical case: - multiple labels for a node/edge

Optional enrichment from 1 path to another: ```cypher match p1=(m1)(b) with p1 match p2=(m2)

uh sure maybe starting point is what already works :) next week can do team umap st (thomas) + query (you) + this (me)? or maybe your/daniel/thomas/tanmoy stuff, and this...

as part of the fix, maybe we add a unit test? at least '1 out of 3 cold starts should be subsecond'

@tanmoyio @dcolinmorgan can we look as part of landing cu_cat?

I documented how I tracked it down last time: https://www.graphistry.com/blog/import-pygraphistry-from-10s-to-1s-by-python-import-lazy-loading-and-tuna

i mean why is `import graphistry` slow -- is it pulling in some unexpected deps during import time that need to be lazy loaded?

w/ lazy loading, `cu_cat`, `cudf`, etc shouldn't load at `import graphistry`, i'd expect at most just `pandas`

Option 3. Lazy load Delay any cudf imports until the code using it (including "auto") is run. 'import graphistry' shouldn't need to import cudf. Can you find the path triggering...