[R] Fix function and unit testing bugs

[gsvidaurre]

score_clusters() has hardcoded timezones argument values in different sections that need to be updated from "America/New York" to the value provided to the tz argument, or just "". Fixing the timezone argument so it accepts geographic/location information should be a future focus.

[gsvidaurre]

Originally issue #477: Fix the following in unit tests:

The path for unit testing needs to be generalized, ideally in the main script to run all unit tests
The code to install packages if they're not already installed should updated to:
is_installed <- function(p) is.element(p, installed.packages()[,1])

invisible(lapply(1:length(X), function(x){
if(!is_installed(X[x])){
install.packages(X[x], repos = "http://lib.stat.cmu.edu/R/CRAN")
}
}))

[gsvidaurre]

Also fix a bug in detect_perching_events(), in which the function fails to detect two empirical RFID detections around 1.5 seconds apart, even using thresholds of 1 or 2 seconds.

[gsvidaurre]

I updated the logic in detect_perching_events() to correctly cluster timestamps within a given temporal threshold into perching events. I also added the number of detections and the min, mean, and max gaps between consecutive timestamps for each perching event into the overall results. Next, I need to modify other functions that use similar temporal rules for clustering.

[gsvidaurre]

I checked preprocess_detections(), detect_clusters(), and score_clusters(), and only updated the logic inside of detect_clusters() to match detect_perching_events(). I tested all functions on an empirical dataset. Next, check earlier posts in this issue to make sure those are resolved, and make sure all changes pass unit tests before submitting a pull request.

[Smith-VidaurreLab]

One more bug to fix in score_clusters(): The function currently returns 3 perching_event_type columns, including with .x and .y. Fix the logic around joins