Ben Steer

For the pandas and Parquet loader we want to do the following: - [ ] Bring the order of required arguments in line with add_node/add_edge - this means instead of...

- [x] Move free queries, LANL and graph500 out of repo - [x] Remove examples from the members in cargo.toml - [ ] Add a new step to test and...

Currently the graphql server calculates the schema on the fly - this is fine for small graphs, but is not for anything meaningfully large.

Would be helpful to be able to directly get the id of an edge instead of having to do {src{name} dst{name}}

Having non-graph files in the working_dir at start causes the server to fail to start, these should just be ignored, logging a warning.

Now that we have g.latest() this brings up the question of if applying a view such as subgraph/layer - should this reduce the latest/earliest time? This is because in the...

Currently can select a set of node_types we want, but not those we want to filter out

The following elements are currently missing from read the docs and need to be added in: ## Ingestion - [ ] node_types - [ ] node_types via df loaders -...

Similar function to subgraph, but instead filtering down to a set of provided edges.

```python g = PersistentGraph() g.add_node(1,1) g.node(1).add_updates(1,{"weight":10}) g.node(1).add_updates(3,{"weight":20}) g.window(5,7).node(1).properties.temporal.get("weight") #Returns (5,20) ``` ```python g = PersistentGraph() g.add_property(1,{"weight":10}) g.add_property(3,{"weight":20}) g.window(5,7).properties.temporal.get("weight").values() #Returns nothing ```