Interactive plotly backend

[ejolly]

@ljchang Here's another fresh PR that specifically isolates out the new interactive plot_detections method you added. It does not include the live demo plotting code. We can prepare a separate PR for that functionality.

As discussed I've renamed your method it to iplot_detections and kept the old plot_detections for backwards compatibility. Few things:

1. There looks like there's an issue with plot_multipleframes_detections if images are different sizes. Any quick thoughts to fix this? You can reproduce it using the following code:

from feat.utils.io import get_test_data_path
from feat import Detector

detector = Detector(verbose=False)
multi_face = os.path.join(get_test_data_path(), "multi_face.jpg")
single_face = os.path.join(get_test_data_path(), "single_face.jpg")

out_combined = detector.detect_image([single_face, multi_face])

out_combined.iplot_detections()

2. Here are the additional features I'm going to try to get working based on your other PR https://github.com/cosanlab/py-feat/pull/178:

• [ ] Explore tooltip functionality
• [ ] add adjustment buttons for playback speed in plot_multipleframes_detections
• [ ] Add smoke tests

3. Here are feature that seem like they're for the live demo plotting feature which isn't part of this PR so I won't work on here. But correct me if I'm wrong:

• [ ] streamline muscle au dictionary
• [ ] add examples for saving plots and videos
• [ ] allow multiple detectors to be toggled in plot_singleframe_detections()
• [ ] ideally, both types of interactivity would be combined into a single plot
• [ ] add new function to plot faces from AUs (the other main way to generate plots in py-feat)