Weathercan Versions Save

R package for downloading weather data from Environment and Climate Change Canada

v0.7.1

9 months ago
  • stations() now uses the most recent version of the data even if it hasn't changed (prevent message regarding age of stations data frame).
  • Remove startup message about deprecated stations data frame
  • Move Mapping Article from docs to website (https://steffilazerte.ca/posts/weathercan-mapping/)

v0.7.0

10 months ago

weathercan 0.7.0

  • Internal updates to tests (testthat 3rd edition)
  • Small changes to messages
  • Switch completely to sf (remove sp dependency)
  • Remove "Use with tidyverse" vignette - better to go to the source: https://r4ds.hadley.nz/
  • Remove "Meteoland" vignette as functions are defunct

Bug fixes

  • Fix bug with Interpolate where silently transforms non-matching timezones. This can produce incorrect matching when using "local-UTC" timezones (as weathercan does as of v0.3.0). Now timezone mismatch results in an error so users can decide how it should be handled.

v0.6.3

1 year ago
  • Internal re-arranging and clean up
  • Stations without lat/lon now have NA timezone
  • Fixed vignette images

v0.6.2

2 years ago
  • Create cache dir for stations data recursively
  • Fix choice of local vs. package version of stations data frame
  • Update to readr v2
  • Add flexibility for csv/tsv stations files (fixes #126)
  • Update stations url
  • Make examples and tests robust to internet issues

v0.6.1

3 years ago

Small changes

  • Save stations() data to local cache

weathercan 0.6.0

(needed to be fixed before could even make release!)

Big changes

  • Move from data frame stations to function stations(). Returns same data but is updateable with stations_dl() and you can check download dates version with stations_meta() (fixes #10)
  • Download climate normals from climate.weather.gc.ca (fixes #88)
    • More stations available (more than 2x as many!)
    • More year ranges available (1981-2010 and 1971-2000; Note that while climate normals from 1961-1990 are available from ECCC, they don't have climate ids making it tricky to download reliably)

Small changes

  • Remove old deprecated function arguments
  • Better test coverage (#81)
  • Better handling of http errors (#101, #119; Thanks @KevCaz!)

Bug fixes

  • Download stations data frame from ECCC Google drive rather than ECCC FTP site
  • Update dependency versions (#111, #112, #118)

v0.5.0

3 years ago

Small changes

  • Internal changes to address change in formatting of historical weather data provided by ECCC (includes new parameters for the amount of precipitation in mm: precip_amt, precip_amt_flag; fixes #107)
  • Updated stations data frame

Bug fixes

  • Updated normals column values (fixes #106)

v0.4.0

3 years ago

Bug fixes

  • Fixed odd bug where some Linux systems failed to download stations data

Features and potentially breaking changes

  • Added caching in memory with memoise (caches for 24hrs, can change this by restarting the R session)
    • Caches individual downloaded files, so you may see a speed up even if you change the parameters of the download.
  • Some missing values in meta data previously were "" but are now explicitly NAs

Internal changes

  • Use readr for reading data
  • Use vcr for tests

v0.3.1

4 years ago

Small changes

  • Internal changes to address change in formatting of historical weather data provided by ECCC (fixes #83)

v0.3.0

4 years ago

Big changes

  • New function: normals_dl() function downloads climate normals. Addresses issue #38.
  • New argument: stations_search() has normals_only to return only stations with climate normals
  • Deprecated url argument in favour of weathercan.urls.stations, weathercan.urls.weather and weathercan.urls.normals options.
  • Deprecated tz_disp in favour of time_disp. Now all timezones are UTC, but the displayed time is either local time or UTC. When time_disp = "none", the time displayed is local time without daylight savings, similar to how ECCC presents the data. This means that data from different time zones will have similar ecological times (i.e. midnights will be comparable), but the actual times are not UTC. When time_disp = "UTC', the time displayed is in UTC timezone, meaning that stations from different times zones will have true times (i.e. midnight observation in Toronto will be three hours before midnight observation in Vancouver). Addresses issue #74.

Small changes

  • Add parameter in station_search() to restrict by start and end dates. Addresses issue #35.
  • Internal change, switching to .data and "" for all non-standard evaluations as opposed to listing global variables
  • Tweaks to keep compatibility wit tidyr

Bug fixes

  • Fix bug #69 which resulted in daily downloads missing partial years when the date range spanned two calendar years
  • Fix bug #70 where internal stations data frame references conflicted with local references to stations
  • Fix bug #72 which was a security vulnerability in an article's json

v0.2.8

5 years ago

weathercan 0.2.8 (2018-10-08)

Bug fixes

  • Add timezones to the stations data frame to remove dependency of Google API. Timezones added with the lutz package, so updates the the stations data frame now require lutz and sf packages (added to Suggests).

Changes

  • Sort stations by station_id not by station_name

Other

  • Update all internal data frames