lmeyerov

The pygraphistry[ai] files (feature_utils, umap, ..) are doing some surprising & awkward use of `logging` that make dev + usage difficult, it should probably be: - [ ] files initializes...

[FEA] API personal key support

1

Need support for new api service keys: - [x] `register(personal_key=...)`, used for files + datasets endpoints - [x] add to `README.md` + testing uploads

[FEA] secondary index for node size

6

**Is your feature request related to a problem? Please describe.** In categorical settings, default node sizes should often be normalized per category, not globally, as they may be sampled from...

Waiting on dirty_cat to upgrade scikit support: https://github.com/dirty-cat/dirty_cat/issues/259

**Describe the bug** In `g.nodes(df[['time']]).umap()`, a `datetime64[ns]` col causes an exn **Expected behavior** No exn, either: * natively handle datetimes * convert to numeric and treat like that * drop...

**Describe the bug** When igraph conversion is used on a table with no nodes and then plotted, have to write `node_attributes=["_n_implicit", "cluster"]` **To Reproduce** See account incident from 5/4/2022

**Describe the bug** Copied jwt hub dashboard snippet into jupyter and got `must call login()` first ( @webcoderz ) PyGraphistry 0.25.0

**Is your feature request related to a problem? Please describe.** In the Marvel tutorial (.../simple/...), some somewhat tricky code does some nice spatial separation of clusters via edge weights: *...

**Describe the bug** ```python import graphistry dir(graphistry.ast) ``` prints python ast package, not graphistry's this makes it ~impossible to use `.chain()` (AI branch)