Hi all,

Here is a quick update on the CS-FileSystem work.

I’ve expanded the filesystem beyond the original ATLAS-only path. ATLAS is
still the reference implementation, but the code now has a generic
table-backed dataset layer that can expose datasets from TSV/CSV metadata
files. I used that to add support for mdCATH, MemProtMD, and GPCRmd
datasets.

The exported tree now looks like this:

index.json
atlas/
mdcath/
memprotmd/
gpcrmd/

Each dataset entry appears as a directory with a metadata.json file inside
it. I also added a root-level index.json that lists the loaded datasets and
their entry counts, so it is easier to inspect what the filesystem is
exposing.

I added small sample metadata files under examples/ as well, so the
workflow can be demonstrated without needing the full datasets locally.

So far I’ve run:

cargo fmt
cargo metadata --no-deps

cargo test is currently blocked on my machine because the fuser dependency
needs pkg-config/pkgconf, which is not installed in this environment.

These changes should give us a cleaner demo path for the PEARC conference.
Instead of only showing ATLAS, we can show the same filesystem abstraction
working across several scientific metadata datasets and have a clearer path
for adding more datasets.

Best,
Jayanth Vennamreddy

Reply via email to