add the capability to do bulk downloads from /observation endpoint

[elidana]

At the moment, users can only download a data type and a single site per time.
To download data for the siteID=GISB, they will need to run the same command 3 times:

curl -o tmp http://fits.geonet.org.nz/observation?typeID=n&siteID=GISB
curl -o tmp http://fits.geonet.org.nz/observation?typeID=e&siteID=GISB
curl -o tmp http://fits.geonet.org.nz/observation?typeID=u&siteID=GISB

Would be good to provide them the capability of doing bulk downloads.
Two different scenarios I can think of:

• all available typeIDs related to a siteID
  something like: http://fits.geonet.org.nz/observation?typeID=all&siteID=GISB

and

• all available siteIDs related to a typeID
  something like: http://fits.geonet.org.nz/observation?typeID=e&siteID=all

[elidana]

end users comments:
GeoNet/help#33 (comment)
and

would be easier to download n, e and u in one package instead of doing them separately

[rumachan]

I do 'all available siteIDs related to a typeID' with python but its a bit fiddly. Being able to do this with a single query would be nice, and woud enable non-pythoners to do it too.

Both is and the other possible query then result in the obvious question - how do I visualise the downloaded data? You realise that if we have these queries someone will then want to return plots. So we should think about this a bit, sometime.

[elidana]

uuuh, that's tricky. I haven't thought about the possibility of people wanting ALSO to visualize allthethings. But...I think the first example (i.e. all typeID related to a specific siteID), is already covered by #65 , right?

[rumachan]

I just thought it was human nature to ask for the next thing! Yes, right. Before we get much further we really have to think about display of these data