added bar option to timeseries_plot, fixed tests for climatic_summary and added further tests #56
Conversation
…ggplot2 functions
| x_daily <- niger %>% | ||
| group_by(date) %>% | ||
| summarise(mean_rain = mean(rain), st_dv_rain = sd(rain), n_na_rain = naflex::na_n(rain)) | ||
|
|
||
| x_monthly <- niger %>% | ||
| group_by(station_name, year, month) %>% | ||
| summarise(mean_rain = mean(rain), st_dv_rain = sd(rain), n_na_rain = naflex::na_n(rain)) | ||
|
|
||
| x_annual <- niger %>% | ||
| group_by(station_name, year) %>% | ||
| summarise(mean_rain = mean(rain), st_dv_rain = sd(rain), n_na_rain = naflex::na_n(rain)) | ||
|
|
||
| x_station <- niger %>% | ||
| group_by(station_name) %>% | ||
| summarise(mean_rain = mean(rain), st_dv_rain = sd(rain), n_na_rain = naflex::na_n(rain)) |
There was a problem hiding this comment.
@HawardKetoyoMsatsi See here how I fixed the tests. For many examples you weren't grouping by station too.
| niger <- daily_niger %>% | ||
| filter(year %in% 1950:1951) |
There was a problem hiding this comment.
@HawardKetoyoMsatsi I also restricted the data to two years to reduce the time it takes to run tests. This should still give enough data to check correctness.
| t1 <- timeseries_plot(data = niger50, date_time = "date", elements = "tmax", | ||
| station = "station_name", facets = "stations") | ||
| r1 <- ggplot(niger50, aes(x = date, y = tmax)) + | ||
| geom_line() + | ||
| facet_wrap(vars(station_name)) | ||
| expect_equal(gg_data(t1), gg_data(r1)) |
There was a problem hiding this comment.
@lilyclements @HawardKetoyoMsatsi See here I have created a function gg_data which extracts the data that's plotted in a gg graph. So it's a good way to roughly check they are equal - on variables, layers, aesthetics and facets. It won't check theme aspects, we could think of adding functions to do this later.
| niger50 <- daily_niger %>% | ||
| dplyr::filter(year == 1950) | ||
|
|
||
| agades <- niger50 %>% | ||
| dplyr::filter(station_name == "Agades") |
There was a problem hiding this comment.
@lilyclements @HawardKetoyoMsatsi Here I've made two smaller data sets so that testing is efficient.
the bar option allows for doing graphs in #12