feat(front): Add mask annotations on video frames

[timothee-LJN]

Issue

Fixes #214

Description

• This branch is pulled from feat(front): Create video player with annotations #185
• To add mask to video, the data model had to evolve compared to feat(front): Create video player with annotations #185 since tracklets need to accept both bboxor mask
• A tracklet is now an array of KeyVideoFrame, each can have a bbox or a mask
• I would advise to check first datasetTypes.ts file to understand the data structure
• The mask linear interpolation does not feel super snappy since we need to convert mask.count to svg and then to coordinates and then apply the simplifyJs API on each calculation. We also refer on array index to retrieve points, which is not working perfectly (causing sometimes the mask to turn around). Maybe this data structure should be given a second though?