Create filter by day of the week

[rafapereirabr]

Here we have a filter_week_days() function in the gtfs2gps package we could use as base.

[dhersz]

Probably gonna do something like filter_by_weekday(gtfs, weekday, keep), where weekday can be any of c("mon", "tue", "wed", "thu", "fri", "sat", "sun").

[kauebraga]

Looking at the function already implemented in the gtfs2gps package, the new function should also be concerned with GTFS with only the calendar_dates file (without calendar).

[dhersz]

filter_by_weekday() has been introduced in 2cab7b0. It includes a combine argument to control whether you want to use OR or AND when filtering by multiple days of the week. From the function examples:

# read gtfs
data_path <- system.file("extdata/spo_gtfs.zip", package = "gtfstools")
gtfs <- read_gtfs(data_path)

object.size(gtfs)
#> 811304 bytes

# keeps entries related to services than run EITHER on monday OR on sunday
smaller_gtfs <- filter_by_weekday(gtfs, weekday = c("monday", "sunday"))
smaller_gtfs$calendar[, c("service_id", "monday", "sunday")]
#>     service_id monday sunday
#>  1:        USD      1      1
#>  2:        U__      1      0
#>  3:        US_      1      0
#>  4:        _SD      0      1
#>  5:        __D      0      1
#>  6:        USD      1      1
#>  7:        U__      1      0
#>  8:        US_      1      0
#>  9:        _SD      0      1
#> 10:        __D      0      1
object.size(smaller_gtfs)
#> 811248 bytes

# keeps entries related to services than run on monday AND on sunday
smaller_gtfs <- filter_by_weekday(
  gtfs,
  weekday = c("monday", "sunday"),
  combine = "and"
)
smaller_gtfs$calendar[, c("service_id", "monday", "sunday")]
#>    service_id monday sunday
#> 1:        USD      1      1
#> 2:        USD      1      1
object.size(smaller_gtfs)
#> 762152 bytes

# drops entries related to services than run EITHER on monday OR on sunday
# the resulting gtfs shouldn't include any trips running on these days
smaller_gtfs <- filter_by_weekday(
  gtfs,
  weekday = c("monday", "sunday"),
  keep = FALSE
)
smaller_gtfs$calendar[, c("service_id", "monday", "sunday")]
#>    service_id monday sunday
#> 1:        _S_      0      0
#> 2:        _S_      0      0
object.size(smaller_gtfs)
#> 19912 bytes

# drops entries related to services than run on monday AND on sunday
# the resulting gtfs may include trips that run on these days, but no trips
# that run on both these days
smaller_gtfs <- filter_by_weekday(
  gtfs,
  weekday = c("monday", "sunday"),
  combine = "and",
  keep = FALSE
)
smaller_gtfs$calendar[, c("service_id", "monday", "sunday")]
#>     service_id monday sunday
#>  1:        U__      1      0
#>  2:        US_      1      0
#>  3:        _SD      0      1
#>  4:        __D      0      1
#>  5:        _S_      0      0
#>  6:        U__      1      0
#>  7:        US_      1      0
#>  8:        _SD      0      1
#>  9:        __D      0      1
#> 10:        _S_      0      0
object.size(smaller_gtfs)
#> 69880 bytes

For now it only uses the calendar table to filter, as using the calendar_tables adds a lot of complexity to the function. I'm closing this issue for now, but if this becomes a problem in the future we can tackle it later.

[rafapereirabr]

Brilliant !