Add verbose in get_dep_no_obs()

[Rafnuss]

When using get_*() and map_dep() with a species filter, it is expected that some of my deployements won't have any observations of this species. I find the list of deployements display unecessary and suggest to have a verbose argument to the function.

> get_rai(pkg, species = "Suni")
There are 103 deployments without observations: 2022-05-13_L5_c10, 2020-11-10_K3_c05, 2021-06-26_I6_c09, 2020-08-29_G8_c09, 2021-04-25_F4_c01, 2020-08-29_H9_c03, 2020-01-14_I7_C10, 2022-02-27_K5_c09, 2021-09-03_C2_c10, 2022-05-13_M5_c09, 2019-11-07_k7_C01, 2019-08-23_j8_c03, 2020-01-13_h8_c07, 2021-06-26_K7_c05, 2019-04-17_i9_C09, 2022-02-26_D5_c01, 2020-01-14_J7_C09, 2020-05-27_G4_c10, 2019-09-21_h7_C07, 2020-05-28_H4_c09 and others..

[damianooldoni]

The get_*() functions, such as the map_dep() are getting a new shape, see my comment see my comment #351 (comment). and #314. Still, your idea is good 👍 I think we should control the verbosity at package level instead of function level, as described in this nice rOpenSci blogpost. Notice also that the pred_*() functions will be not present anymore in camtraptor v1.0 as they are replaced by the reexported camtrapdp's function filter_deployments(). They are returning also some messages.

So, we will get something like this at the moment using version-1.0 branch of camtraptor:

# Load dev version of camtraptor (branch version-1.0)
library(camtraptor)

# Load example dataset
x <- example_dataset()
# Filter deployments
x %>% filter_deployments(latitude > 100)
#> A Camera Trap Data Package with 3 tables:
#> • deployments: 0 rows
#> • media: 0 rows
#> • observations: 0 rows
#> 
#> And 1 additional resource:
#> • individuals
#> Use `unclass()` to print the Data Package as a list.

^{Created on 2025-01-06 with reprex v2.1.1}

Notice that we get now the printed message as returned by camtrapdp R package. I think your idea of setting a verbosity control, is also a valid enhancement for camtrapdp R package as well. I will create an issue for it mentioning this one.

[damianooldoni]

As mentioned in PR #357, I think it's better to implement this at later stage after release of camtraptor 1.0. Something for camtraptor 1.1.

[Rafnuss]

Thanks for sharing the blogpost! Intresting to read and some food for thought on my side!

[peterdesmet]

From inbo/camtrapdp#152 (comment):

My main point was just to say that when filtering by species, it seems logical that some deployements won't have any observations and I don't think that this info deserved to be outputed to the user.

Conclusion: no action needs to be taken on verbosity control at the moment, since camtraptor 1.0 will rely on camtrapdp's filter_ functions. Those functions do not message the user that there are deployments without observations, which is the desired behaviour.

[damianooldoni]

Added a general issue based on the conversation here, see #358. We can close this specific one as the camtrapdp's filter_*() functions do not return info about deployments with no observations. The automatically printing output doesn't seem to be a real issue.