NASA ADS fix

[keflavich]

NASA ADS querying was broken because of incorrectly used modules in its core.
There are still no local tests, but at least now it uses BaseQuery rather than
commons.

[bsipocz]

@keflavich - speaking of nasa_ads, is #20 still relevant? There is a test in there as a starter, etc.

[keflavich]

Wow, yes, it looks like that has been sitting there stagnant for a long time. It should probably replace the current module, but it needs review & some testing before that can happen.

[keflavich]

#20 should override this at some point, but I'm going to merge it because it fixes a failing remote test. Other than this, the only failures I have gotten today are timeout errors.

[bsipocz]

👍

[bsipocz]

btw have you seen @andycasey's relatively new ads tool? I wonder whether he is interested in making it part of astroquery, or how compatible it is with our query tools.

[keflavich]

yeah, I use it - it's been around for awhile in some other forms. I asked him about this in September 2013, and he expressed interest/excitement, but I think nothing ever came of it.

[andycasey]

Yeah, I failed hard on that.

The code is much more mature now, and half of me was reluctant to move things into astroquery before ADS had updated their API (which was in the works for some time).

I still think it would be advantageous to have ADS part of astroquery, but without having used astroquery, would the "customer usage" be similar? Right now nasa_ads parses the response as an AstroPy Table, but in the ads library we instantiate each article as an object so that we can easily build reference and citation trees to other articles. Would it go against the astroquery mantra if the response returned from a query was some list (generator, actually) of Article objects, each with their own attributes?

[keflavich]

@andycasey No, I think it would be useful to have that interface. But it would probably also be useful to parse into a table, because sometimes tables are nice, no?

[andycasey]

That's good to know! Yeah, tables are great, some of my best friends are tables.

I'll work on getting this into a PR-able form before Christmas.

[bsipocz]

Sounds good. Thanks @andycasey

[keflavich]

@andycasey Christmas is soon, any progress? Would still be great to have your module as part of astroquery.

[bsipocz]

@keflavich I've offered Andy to help out if needed, but I will be back in the UK only after Christmas.