Kyle Ellrott

Use case: workflow engine expects files to be mounted on local file system, but worker nodes will be receiving files from object storage. Possible solution: 1. Mount object store on...

Indexing method: https://arxiv.org/abs/1603.09320 Example code: https://github.com/Bithack/go-hnsw One idea would be to refactor existing indexing code used by the KV graph code, so that it can be layered onto other drivers...

The `JobStorage` interface is defined in https://github.com/bmeg/grip/blob/main/jobstorage/storage.go. Currently there is only a file system based implementation. There should also be an ElasticSearch backed interface to store traversal jobs can be...

- [ ] Enable mongo (partial mongo pipeline execution ?) - [ ] Add security options - [ ] Additional unit tests

Allow user to do comparisons against previously stored elements. Example: ``` G.query().V().hasLabel("Character").as_("1").out("homeworld").out("residents").has(gripql.neq("$1._gid", "$._gid")) ```

Currently grip CLI only utilizes the GRPC client, so can't talk to a server that only exposes the HTTP port.

A date type, to enable selection logic based on date ranges, would enable new queries.

The query ``` q = O.query().V(["ENSG00000012048", "ENSG00000012048"]) print(list(q.count())) ``` returns 1. It seems like the return value should be 2