Data Re-population for Development

Currently, every time we restart the containers, the sample data, including files and feature extractions data is wiped and had to be rerun which takes a lot of time. This is particularly annoying during development. We should:
- [ ] Move the population command outside of the start script
- [ ] Create a quick way to dump test data back when starting the containers.

**Note:** This should only be for the development process. Data Re-population for Production should be handled differently to take into account the:
- Changes in features
- Users' uploaded files/changes
- etc?