Usability and security improvements
This version improves several usability and security issues in the v1.0 release. The changes are summarized below:
- Update README.md to include the documentation of setting up https.
- Tweak the text to also ask people to label low opacity smoke.
- Fix a google sign-in problem that users cannot select a new account after signing in with one account.
- Add a script and several files for adding new videos.
- Fix the shell script for removing a detached screen on the background.
- Add a new file that contains latest camera findings.
- Remove gold standard videos from get_pos_labels and get_neg_labels API calls.
- Change get_pos_labels_by_researcher and get_neg_labels_by_researcher API calls to only admin (researchers and experts) access.
- Exclude gold standards from get_pos_labels_by_researcher and get_neg_labels_by_researcher API calls.
- Add a function to download all video labels for researchers on the gallery page.
- Allow expert-type users to access the advanced gallery dashboard (without the function for setting label states).
- Update the front-end related to the above API permission changes.
- Prevent blacklisted users from getting valid user_token from the server.
- Add a view table in the database to track records of viewed videos.
- Add current user scores to the connection and batch table.
- Reduce the duration for video playback testing.
- Prevent the "My Collection" button on the label page from opening a new tab.
- Sign each video batch with the user id and verify this when sending labels back to the server.
- Request a new video batch after google sign in or out (because the signature does not match).
- Fix a bug that the google client id will be wrong after google sign out.