Incorporate into Biopython about jgi-query HOT 3 CLOSED

Hey @rchurt, thanks for the idea. I wonder, though, how frequent of a use-case this would be (accessing it programmatically from within Python vs. via CLI)—do you have any examples of how you're thinking that might be of use? The way I've always used it is for pulling JGI files onto remote servers where I don't have any web-based interface options, but I'm curious to hear if you've encountered specific cases where having direct programmatic access would be helpful.

One thing I certainly plan to do at some point is package everything up to be pip-installable in order to make installation as painless as possible (though, since it's currently just a single Python script, "installation" is really just putting the script somewhere convenient).

from jgi-query.

So I was imagining using it exactly like I use Biopython's Bio.Entrez package for batch retrieving records that happen to only be available on the JGI genome portal (otherwise I would just get them from NCBI). I'd assumed that was what everyone was doing with this script, but you'd know better than me!

from jgi-query.

Gotcha, right—I don't use Entrez very much, but I can see how there are possible applications there as they both involve querying sequence databases. EDIT: Just to add: my sense is that a lot of folks use Entrez to do more nuanced sequence work, e.g. extracting sub-sequences, pulling data from a wide array of datasets matching certain criteria, etc., which isn't really the application I envisioned for jgi-query—I wrote it primarily to get streamlined access to files that are otherwise annoying to pull to the command-line manually.

I'm not too informed about how Bio.Entrez pulls its data, or whether NCBI has a more advanced API to handle things like file streaming (e.g. .read()), etc. JGI's API is fairly anemic in that they really only provide a way to download an XML. Most of what jgi-query does is simply parse that XML and present it more (arguably) accessibly.

Regardless, I think it's a great idea but suspect it would involve essentially a complete re-write of the code. I'd certainly welcome anyone with the Python chops to use as much of my codebase as they see fit, but I doubt I will have time to dedicate to doing it myself in the near/medium future.

I'm closing this for now but if you've got further ideas/input feel free to re-open.

from jgi-query.