284 add support for camtrap dp v1.0

[PietrH]

STATUS: Ready for review

• Check that all mentions of the release candidate are removed and replaced by v1.0

Failing tests:

test-read_camptrap_dp.R

• deployments.baitUse is expected to be NA but is actually none, it should be NA since it's false in the source, these values have changed between the examples in the rc and v1.0
• deployments.session is expected to be NA but has values from deploymentGroups, these values have changed between the examples in the rc and v1.0

Missing tests

I feel additional coverage was needed for the changes below, since we were passing tests for a new v1.0 camtrap-dp even though they were not implemented yet.

Add tests for:

• Downconversion of media.captureMethod values activityDetection to motionDetection
• Downconversion of project.catureMethod (in metadata) values activityDetection to motionDetection

[PietrH]

I seem to have pulled some code from #282 into test-read_camtrap_dp.R by accident, I'll try to remove anything that might result in merge conflicts with #282

[PietrH]

I would just use the package as is and remove some of the tests. They are only used to test the downconversion, which will disappear at some point.

I'd rather skip the test for now, so it's easier to debug if it does happen to break before we switch to the different data model in camtraptor. Skipping implemented in fb477ef

[PietrH]

@peterdesmet Should deployments.deploymentGroups be mapped to deployments.session in v0.1.6 ?

This is how it works at the moment;

  if ("deploymentGroups" %in% names(deployments)) {
    # map to session and then remove
    deployments <- deployments %>%
      dplyr::mutate(session = dplyr::case_when(
        is.na(.data$session) ~.data$deploymentGroups,
        is.na(.data$deploymentGroups) ~ .data$session,
        !is.na(.data$deploymentGroups) & !is.na(.data$session) ~ 
          stringr::str_c(.data$session, 
                         .data$deploymentGroups, 
                         sep = " | "))) %>%
      dplyr::select(-"deploymentGroups")
  }
  

If there is no value for session, it becomes deploymentGroups, if there is no value for deploymentGroups, it become session, if there is a value for both, they are concatenated with a pipe.

[peterdesmet]

Regarding deploymentGroups: functionality is good, but would simplify and reverse order (so that deploymentGroup info is always kept):

1. Both present: concatenate with pipe
2. deploymentGroups present (and not session): populate session
3. session present: keep in session

Remove deploymentGroups afterwards

[PietrH]

Ready for review, something is going wrong building the vignettes, any ideas?

[damianooldoni]

I couldn't reproduce the error while building the vignette with windows-latest+ R 3.6.0.
It's a problem while generating the "black multiply" icon in leaflet.
I am afraid that leaflet is not really concerned about supporting older R versions on all type of machines. See https://github.com/rstudio/leaflet/actions/runs/6776222357

They support R 3.6.3 on Ubuntu only. Everything runs on Ubuntu except a machine running R 4.3.2 on windows-latest.
My proposal is to remove such check as well and so leaving only the check with Ubuntu + R 3.6.3.

[PietrH]

Ok for me

[peterdesmet]

Fine by me. Can we merge this request? Note that it contains a version bump to v0.21.0 which you may want to include only after you fixed other issues.

[damianooldoni]

To avoid too much confusion with versioning, I think we should just merge this branch to main as version 0.21.0 asap.
I would then leave the version number as it is for the next imminent PRs and then a release will follow.