This feels entirely arbitrary. Many pathogens won't even have subtype. Would be much better to be able to filter to arbitrary fields when downloading and not have the download method continually grow. Reasonable to ask for all Zika sequences with country:brazil for example. Currently, add_filter_to_query is suitably generic, but subtype, segment, etc.. is getting specifically passed in here: https://github.com/nextstrain/flora/pull/2/files#diff-92d1176a15501016fcbcca39dd9099e4R110
Separate discussion is how segment vs locus should behave. For a full genome, like Ebola, should it be segment:genome and locus:genome. Seems odd, but I don't have an obvious suggestion. But would be also good to think about how bacteria will be behave. Does segment work for chromosome vs plasmid?
However, okay to merge this and then separately fix things to be more generic.
I'm not sure I like the direction this is headed. Currently
download
takessubtype
,segment
andlocus
: https://github.com/nextstrain/flora/pull/2/files#diff-92d1176a15501016fcbcca39dd9099e4R16This feels entirely arbitrary. Many pathogens won't even have subtype. Would be much better to be able to filter to arbitrary fields when downloading and not have the
download
method continually grow. Reasonable to ask for all Zika sequences withcountry:brazil
for example. Currently,add_filter_to_query
is suitably generic, butsubtype
,segment
, etc.. is getting specifically passed in here: https://github.com/nextstrain/flora/pull/2/files#diff-92d1176a15501016fcbcca39dd9099e4R110Separate discussion is how
segment
vslocus
should behave. For a full genome, like Ebola, should it besegment:genome
andlocus:genome
. Seems odd, but I don't have an obvious suggestion. But would be also good to think about how bacteria will be behave. Doessegment
work for chromosome vs plasmid?However, okay to merge this and then separately fix things to be more generic.