Add time windows to longitudinal queries

[flaneuse]

For longitudinal queries, add a parameter datemin and datemax (feel free to change the names) to filter queries based on dates.

API handlers / routes covered:

• Prevalence by location API call -- PrevalenceByLocationAndTimeHandler API call
• Prevalence by location all lineages -- PrevalenceAllLineagesByLocationHandler

Expected behavior

1. User runs a GET query like https://api.outbreak.info/genomics/prevalence-by-location?pangolin_lineage=BA.5&datemin=2022-06-01&datemax=2022-10-01
2. ElasticSearch query runs the date filter before the aggregation, to restrict the data going into the aggregation step to those records between the dates.

• datemin and datemax should be in ISO 8601 YYYY-MM-DD format
• filtering should be inclusive (as in <= (less than or equal to) and >= (greater than or equal to) of the datemin and datemax.
• if no datemax is specified, it should only filter the datemin
• if no datemin is specified, it should only filter the datemax`
• if neither datemin not datemax is specified, there should be no filtering.
• Should also have basic error checking, if datemin or datemax is not specified as an ISO string. Should return a 400 error (bad request)

3. ElasticSearch passes the filtered data to the existing aggregation code.
4. Output format is unchanged, just with fewer dates.

Expected outcome

• Should decrease API call query times by only aggregating a subset of the data, rather than all the data.
• Should lower the amount of data passed to the front-end, since only the relevant dates are passed.

Related docs

• https://github.com/outbreak-info/outbreak.api/blob/master/web/handlers/genomics/docs.md

[flaneuse]

closed in 0e38feb