bug: read_parquet mangles huggingface prefix

[cboettig]

What happened?

duckdb supports reading data directly from huggingface! https://duckdb.org/2024/05/29/access-150k-plus-datasets-from-hugging-face-with-duckdb.html

But read_parquet() in ibis doesn't recognize the hf:// prefix, and decides that it must be some kind of local file path.

What version of ibis are you using?

9.10

What backend(s) are you using, if any?

DuckDB!

Relevant log output

parquet = "hf://datasets/boettiger-lab/gbif/usa_h3/*/*.parquet"
df = con.read_parquet(parquet)

FileNotFoundError: /home/rstudio/huggingface/spaces/gbif/hf:/datasets/boettiger-lab/gbif/usa_h3/*.parquet

Code of Conduct

• I agree to follow this project's Code of Conduct

[cpcloud]

I suspect that we should probably avoid hard-coding the list of ignored prefixes, which we are using so that we can turn files into absolute paths, but I don't think that's necessary either.

I'll remove the hardcoded list, which allows hf and any other prefix someone comes up with to work, and see what breaks.