Unexpected behavior when finding the labeled-data folder in analysis_deeplabcut

[rojasgabriel]

When calling the label action, the user is expected to provide the name of the camera as an input to the video filter (-f) flag. This successfully finds the proper folder and is meant to ensure proper function when the user has labeled data from multiple camera views. However, the glob pattern matching does now work if the user provides the entire name of the labeled-data folder as an input to the video filter argument. A better solution would be to find the correct labeled-data folder without enforcing a strict structure to the pattern matching process. For now, users should specify only the name of the camera in the video filter argument to ensure proper behavior.