pygraphistry

pygraphistry copied to clipboard

[FEA] RDF/sparql support - neptune, ...

3

**Is your feature request related to a problem? Please describe.** Neptune and other DBs support rdf/sparql, it'd be great to enable one-liners fo them! **Describe the solution you'd like** Integration...

lmeyerov

**Is your feature request related to a problem? Please describe.** Neo4j, via APOC, is getting arrow reads/writes, which should be faster & safer to use than pandasneo4j! See discussion: https://github.com/neo4j-contrib/neo4j-apoc-procedures/pull/1859...

lmeyerov

**Is your feature request related to a problem? Please describe.** Sometimes hypergraph creation is slow, or cause a large upload, so it'd be better to just do it remotely **Describe...

lmeyerov

**Describe the bug** In a windows dataset, column `EventID` already exists, so event ids use the windows event code instead of the row id **To Reproduce** ```python import pandas as...

lmeyerov

- [ ] enable dask in cpu tests - likely docker/setup.py dev deps - [ ] reuse dask/dask_cudf scheduler across tests, either by open event loop or (probably more representative?)...

lmeyerov

Hypergraph turns `None` / `np.nan` values into nodes despite `drop_na=True`, see unit test xfail's in https://github.com/graphistry/pygraphistry/pull/225 Blocked by https://github.com/rapidsai/cudf/issues/7735 for rapids side Current workarounds are confusing the code, so we...

lmeyerov

See https://github.com/graphistry/graph-app-kit/issues/56 - [ ] download nb - [ ] single gpu nb - [ ] multi-gpu nb - [ ] parallel io nb - [ ] tutorial

lmeyerov

The label property is overwritten by T.label within vertex_to_dict(), a unique property name should be used to avoid collision. https://github.com/graphistry/graph-app-kit/blob/master/src/python/neptune_helper/df_helper.py#L9

xocliwtb

```python df = pd.DataFrame({'s': ['a', 'b', 'c'], 'd': ['b', 'c', 'a']}) graphistry.edges(df, 'WRONG', 'd').plot(as_files=True, render=False) ``` Will not hint at the binding error, while `as_files=False` will. Both should -- unclear...

lmeyerov

[BUG] memoization fails on df's with lists

1

neo4j movie memoization fails on df with: ``` {'bindings': {'edges': roles _bolt_relationship_id type _bolt_start_node_id_key \ 0 [Emil] 7 ACTED_IN 8 1 NaN 6 PRODUCED 7 2 NaN 5 DIRECTED 6...

lmeyerov