Fetch vpts data from aloft

[PietrH]

• Needs to support fetching multiple radars as a character vector
• Needs to support a single date as input
• Needs to support date interval as input
• Fetch from https://aloftdata.s3-eu-west-1.amazonaws.com in parallel,
• should we provide an option to set a different aws region?
• add user agent to requests to s3
• compare to coverage to figure out what data is available
• Allow users to provide intervals with a resolution smaller than a single day
• Add a column to the returned data.frame with the radar source
• Add source argument
• return object compatible with {bioRad}
• also allow returning tibble with source column: as_vpts = TRUE argument?
• set readr spec/cols for read_csv to be more defensive and efficient!
• R CMD CHECK
• update NEWS
• up dev version number
• update DESCRIPTION
• split tests to do with vpts output object testing
• rename as_vpts to as_tibble and set behaviour as inverse
• move helpers to utils.R
• abstract fetching csv from web in parallel with user agent to helper
• ❗ ❗ fix bug in bioRad:::vpts_validate() call ❗ ❗
• rename arguments in get_vpts_aloft() to be in line with get_vpts() as much as possible
• pass R CMD CHECK on CI
• rename get_aloft_coverage() to aloft_coverage()
• set aloft_coverage() to public function
• create general file getter helper instead of read_vpts_from_url(): use for all functions that get multiple files from web

If a radar is present in both UVA and BALTRAD, both are fetched.

[bart1]

How about using col_types for read_csv and specifying some columns as factors (saves memory). That will directly also so some data type checking

[bart1]

I would also consider using glue over reprex magick :), its a dependency anyway and much easier to understand

[PietrH]

There is now a progress bar with an ETA for larger intervals, eg:

get_vpts("bejab", lubridate::interval("20210101",20210301"))

[PietrH]

vpts objects in bioRad can only contain data of a single radar, how do we want to handle that? via bioRad::as.vpts()

[PietrH]

vpts objects in bioRad can only contain data of a single radar, how do we want to handle that? via bioRad::as.vpts()

I've decided to return a list of vpts objects

[peterdesmet]

I discussed the get_vpts() function with @CeciliaNilsson709 and here's what we find the most straightforward:

• source is required parameter
• source only accepts one value. This is in line with get_pvol(). Downloading from multiple sources in one function call adds unnecessary complexity.
• source can have baltrad as default
• Return the data as a list of vpts objects
• If only one radar was chosen, return as a single vpts object
• When a list, use the provided radar code (not the one in the data) as identifier