halo-db / storymap

Story mapping
Creative Commons Zero v1.0 Universal
3 stars 0 forks source link

Search across campaigns #13

Open sabarth opened 3 years ago

sabarth commented 3 years ago

As a guest & all users, I would like to be able to search across all campaigns to identify matching records. Possible search parameters:

d70-t commented 3 years ago

This is a great idea! I'd like to see this implemented and would also like to see an addition:

joerg-halo commented 3 years ago

specific altitude/lat/lon/time ranges

In this case, we would require a primary database, right? So I set the "big" label. Please correct me, if I have a wrong understanding here.

d70-t commented 3 years ago

Hmm, probably it is kind of a big thing to the HALO-Database, but I'd expect that any reasonable database which covers field campaign data with defined spatiotemporal extent is able to decode at least the location and time within a dataset.

kdgo commented 3 years ago

Metadata should suffice for finding data; no primary database is needed here. Please see the search mask at https://halo-db.pa.op.dlr.de/search: Amongst others, the time period, the area and variable names may be queried. Such searches are not only possible accross campaigns, but may be extended to external databases (IAGOS). However, this functionality hasn't been maintained for a while and may not work properly ... but it is rather a technical detail to fix than a big issue. Also, connecting other related databases (e.g. NASA's) should not be a structural problem. The devil is in a different detail: Metadata conventions are not enforced strictly. CF conventions are recommended, but violating them just triggers a warning and the file is ingested anyway. In particular, variable names are a mess. For instance, ozone mixing ratio may in principle have a different name in each file. Furthermore, other databases (IAGOS, NASA, ...) may use different conventions at all. Lookup tables might be an option to deal with different conventions or names, but are probably hard to maintain and ambiguities need to be resolved manually. The stricter the parser, the easier to maintain. Compared to variable naming, the spatio-temporal context is rather well curated currently. Most data sets are assigned to flights. Flights are maintained by the DB admin only. They always contain detailed coordinates, start and stop time.

sabarth commented 3 years ago

Again, #32 (or #33) play a key role... Discussion on Search across Databases can be found here: #53